The Economist - 22.08.2026 - 28.08.2026
全球报刊态势与全景学术研判总览 — 2026-08-24
🌐 第一部分：全球宏观战略态势报告（SITREP）
今日综述：本日共全量采集并深度研判来自全球各大主流语系与核心地缘区域的 1 份权威报刊与学术特刊。
📡 今日全球主要报道要闻速览
英国及英联邦 (United Kingdom & Commonwealth) - 《The Economist - 22.08.2026 - 28.08.2026.pdf》AI能产生意识吗？：探讨AI是否能产生意识及其代价，部分科学家认为目前的AI模型在意识方面已陷入死胡同。 - 《The Economist - 22.08.2026 - 28.08.2026.pdf》杰森·阿尔代的非人化：剑桥大学教授杰森·阿尔代因剽窃和伪造研究丑闻垮台并自杀，揭示了大学将其作为种族符号而非个体的体制问题。 - 《The Economist - 22.08.2026 - 28.08.2026.pdf》类脑计算与AI意识：分析类脑计算通过忆阻器在同一位置执行计算与存储，旨在解决冯·诺依曼架构的高能耗问题并模拟大脑运作。
💡 各大报刊的报道脉络呈现出高度多维的地缘博弈、制度转型、社会痛感与前沿科学突破。读者可通过下方【今日跨报思想雷达总矩阵】快速把握各报最具穿透力的思想理论对话与生活世界痛感切入点，或在【全景报刊分卷】中按区域细读每份大报的具体文章精要、制度权力批判与理性情感辩证。
🧭 第二部分：今日跨报思想文化与批判深思雷达总矩阵
本矩阵自动系统汇编当日全球报刊最具穿透力的 【思想与文化深思切入点】 与 【生活世界痛感切入点】，为学者提供一站式理论对话与现象学经验索引：
- （本日暂未提取到独立的跨报思想雷达条目）
📚 第三部分：全景报刊分卷学术档案与精粹汇编
📍 英国及英联邦 (United Kingdom & Commonwealth)
📰 《The Economist - 22.08.2026 - 28.08.2026.pdf》
Magazine - 24.08.2026AI能产生意识吗？
- 社论 / AI能产生意识吗？：探讨AI是否能产生意识及其代价，部分科学家认为目前的AI模型在意识方面已陷入死胡同。
- 英国 / 杰森·阿尔代的非人化：剑桥大学教授杰森·阿尔代因剽窃和伪造研究丑闻垮台并自杀，揭示了大学将其作为种族符号而非个体的体制问题。
- 科学 / 类脑计算与AI意识：分析类脑计算通过忆阻器在同一位置执行计算与存储，旨在解决冯·诺依曼架构的高能耗问题并模拟大脑运作。
- 英国 / “百球赛”与郡级板球复兴：“百球赛”的城市化革命意外带动了对传统郡级板球外场的热情，且显著提升了女子板球的观众数量。
- 基辅 / 乌克兰拦截弹耗尽与空战局势：俄罗斯增加弹道导弹产量且拦截弹供应减少，导致乌克兰民用物流损失约5亿美元，空战优势发生转移。
- 美国 / 美国大选资金分析：共和党虽在党组织和PAC中积累约10亿美元现金优势，但民主党的筹款更去中心化，在竞争选区中更具优势。
- 巴西 / 巴西“Pix修正案”预算违规：审计发现82%的Pix修正案转账存在违规，预计公共资金损失约3.4亿美元，引发立法者与最高法院的冲突。
- 非洲 / 索马里兰的战略野心：总统伊罗希望通过向大国提供军事基地保障红海安全，将柏培拉港打造为非洲之角门户以获取经济收益。
- 中国 / 北海航道（NSR）的机遇与局限：气候变化导致北极冰盖缩减，使北海航道在8-10月可运行，能降低燃料成本并有利于运输温敏高科技产品。
- 欧洲 / 欧洲政府转向开源软件：德国部分州政府将公务员迁移至Linux等开源系统，旨在降低对美国科技巨头的依赖并应对地缘政治风险。
- 学术 / 达龙·阿西莫格鲁的学术争议：诺贝尔经济学奖得主阿西莫格鲁虽高产且受推崇，但部分经济学家批评其模型被用于支持失败的民粹主义政策。
- 经济 / 霍尔木兹海峡的“寻租”政治经济学：类比莱茵河收费城堡，分析伊朗、阿曼及美国可能在霍尔木兹海峡征收过路费以获取经济租金的风险。
- 文化 / AI社会影响书籍评论：通过对比人类驯化狗的历史，探讨AI作为强大伙伴在社会中是会变得温顺还是凶猛的潜在风险。
- 讣告 / 埃兹拉·庞德之女的逝世：回顾诗人埃兹拉·庞德之女的一生，记录其在蒂罗尔的成长经历及其作为父亲捍卫者的身份。
暂无制度批判分析。
暂无情感辩证分析。
The Economist - 22.08.2026 - 28.08.2026.pdf
The Economist (经济学人)
爱国者导弹短缺
杰森·阿尔代的非人化
关于凝结尾迹的真相
为什么每个人都热爱瑞典
2026年8月22日-28日
AI能产生意识吗？
人类必须面对的 一个令人震撼的 问题
百达翡丽 (PATEK PHILIPPE)
日内瓦
THE CUBITUS COLLECTION
您从未真正拥有过一座百达翡丽。 您只是在为下一代保管它。
经济学人 2026年8月22日
3
目录
封面故事
AI能产生意识吗？即便它们不能，也可能会被如此对待——而这将让全人类付出沉重代价：社论，第7页。科学家们在思考算法是否有一天会觉醒并产生感觉，第13页。一些人认为AI需要更多类脑芯片才能发展出意识，第14页
爱国者导弹短缺
美国及其盟友的拦截弹即将耗尽，第49页
杰森·阿尔代的非人化 他的支持者和反对者都将他视为一个符号，而非一个人：社论，第9页。他的悲剧性死亡揭示了剑桥大学的什么问题，第19页
关于凝结尾迹的真相
消除最持久的凝结尾迹可能会解决惊人数量的全球变暖问题：社论，第10页。一项跨大西洋试验旨在展示通过重新规划航线能为地球降温多少，第64页
为什么每个人都热爱瑞典
美国的右翼和左翼都认为瑞典证明了他们的正确性：夏洛门，第27页。瑞典的一个角落正在酿造世界级的葡萄酒，第70页
→ 下载《经济学人》应用程序，获取全周发布的文章、播客、视频等更多内容。
本周世界
5 政治与商业新闻摘要
社论
7 AI 不要赋予它们权利 8 以色列定居者 另一场起义迫在眉睫 8 帕兰提尔在英国 不要放弃它 9 杰森·阿尔代 是一个人，而非符号 10 凝结尾迹 改变飞行路径
读者来信
11 关于AI与国家、AI与好莱坞、石油管道、餐厅、明信片、理查德·瓦格纳
特邀专栏
12 布莱斯·阿格拉·伊·阿尔卡斯论重新构想心智
简报
13 AI与意识 算法会觉醒吗？ 14 生物计算 将脑细胞置于硅片之上
贝杰霍特 英国暑假如何呈现出所有情况中最糟糕的一面，第22页
英国
17 对帕兰提尔的憎恨 19 杰森·阿尔代的兴衰 20 板球的夏季惊喜 20 规划制度的彻底改革 21 改革英国党（Reform UK）的改革者 22 贝杰霍特 伟大的英国大逃离
欧洲
23 防火墙忧郁 25 没有拦截弹的乌克兰 26 强悍的费多罗夫先生 26 法国的西方之舞 27 夏洛门 瑞典的自由意志主义社会主义
美国
29 基础设施面临风险 30 筹款之战 31 俄克拉荷马州的公民课 32 女子篮球 33 奥巴马医改的脆弱性 34 莱克星敦 真正的Covid-19丑闻
美洲
35 巴西强大的国会 36 阿根廷：马球权力 37 巴拉圭的另一种语言
中东与非洲
38 西岸的混乱 39 伊朗的军事政权 40 南非的国家俘获 41 索马里兰总统
►► 目录见背面
4
特朗普先生的束缚
美国与伊朗敲定最终和平协议的截止日期已过，但双方未能达成协议。特朗普先生威胁称，如果阿曼（美国的盟友）阻碍与伊朗的谈判，他将对其进行轰炸。伊朗则表示，其与阿曼在霍尔木兹海峡航运问题上已接近达成协议。然而，针对通过该海峡船只的袭击仍在继续，一名水手丧生。另一个美国盟友阿拉伯联合酋长国对伊朗实施了贸易禁运，指责伊朗自5月初以来首次向其领土发射导弹。伊朗声称此次袭击是由美国和以色列实施的“假旗行动”。
哈卡因德·希奇莱马在赞比亚总统选举中赢得了第二个任期，但此次选举因多个选区现任者票数虚高，以及政府军队在一次突袭中杀害一名反对派政治人物而蒙上阴影。即便选举在公正环境下进行，希奇莱马先生可能依然会获胜。但赞比亚作为一个运行相对良好的多党民主国家的形象已受损。
刚果民主共和国的埃博拉疫情成为了该国最致命的一次疫情，截至8月18日，已有2,378人死亡，确诊病例超过5,000例。世界卫生组织（WHO）表示，此次埃博拉疫情有望成为记录以来最严重的一次，将超过十年前导致11,000多人死亡的西非疫情。
摩洛哥警方逮捕了110多名试图进入休达（西班牙在北非的领土）的人员。警方表示，几乎所有移民都来自撒哈拉以南非洲。7月下旬，72,000名移民涌入休达，引发了欧盟关于如何最佳处理非法移民的辩论。除2,000人外，其余移民均被迅速遣返回摩洛哥，但留下来的人近期在休达举行抗议，要求获得庇护。
俄罗斯当局加强了对自由派反战政党“苹果党”（Yabloko）的打击。普斯科夫市的一家法院判处该党副党魁列夫·什洛斯伯格11年劳改营徒刑，理由是在参与一场关于乌克兰战争的在线辩论后“诋毁”军队。法官打断了什洛斯伯格先生呼吁停火的结案陈词。该党已被禁止参加议会选举。
上个月被解雇的乌克兰国防部长米哈伊洛·费多罗夫呼吁举行选举。在一段发布于YouTube的视频中，费多罗夫先生表示，乌克兰应该“寻找一种合法、安全且现实的机制，使乌克兰即使在长期战争的情况下也能恢复完整的民主进程”。自2019年以来，乌克兰尚未举行选举。与此同时，腐败指控继续困扰着弗拉基米尔·泽连斯基政府。在一名高级助手因洗钱调查而被搜查房屋后，总统将其解雇。
俄罗斯驻伦敦大使馆警告称，在最新报道称乌克兰部署英国制造的无人机攻击俄罗斯后，英国将面临“后果”。对此，英国新任首相安迪·伯纳姆表示，他的政府将继续“100%支持乌克兰”。
剑桥大学坚称，在杰森·阿尔代因其作品涉嫌剽窃以及被指造假人生经历而辞职后，学校一直为其提供心理健康支持。阿尔代先生曾是剑桥大学的正教授，在对其单薄的学术记录进行数周密集审查后，被发现死亡。大学校长史密斯勋爵将他的死亡归咎于一场“种族主义的疯狂围攻”。而调查阿尔代先生的人士反驳称这纯属胡说，他是一名公众人物，理应被追究责任。
哈里王子和梅根·马克尔据报道正准备搬回英国，以便他们的两个孩子能在该国就学。他们将不会恢复其王室成员的工作身份。这对夫妇与王室其他成员的关系一直紧张，导致他们搬到了加利福尼亚州，这一决定后来被称为“梅根脱离”（Megxit）。
唐纳德·特朗普在表示美国将缩减与韩国军队举行的联合军事演习时，令韩国感到措手不及。随后，年度演习被缩短了六天。总统列举了一系列不满：演习成本高昂，向朝鲜发出了“敌对”信号，且韩国在伊朗的“去核化”问题上没有提供帮助。特朗普先生表示，他与朝鲜独裁者金正恩有着“非常好的关系”。
报道称，他希望与金正恩再次会面；特朗普先生在第一任期内曾与金正恩会晤，但他们的峰会没有产生任何结果。
印度内政部长阿米特·沙阿详细介绍了逮捕 200 多名“恐怖分子”的情况，这些人计划在印度独立日庆祝活动期间发动“颠覆性袭击”。沙阿先生表示，这些特工得到了巴基斯坦情报机构的支持，但巴基斯坦政府对此说法予以断然否认。
一场发生在印度尼西亚东努沙登加拉省的地震造成至少 68 人死亡。这是该国自 2022 年西爪哇省数百人遇难以来最严重的地震。
唐纳德·特朗普将原定于对加拿大商品征收的 50% 关税暂停执行三天，并表示与加拿大即将达成一项协议，以解决美国在包括汽车进口关税在内的一系列问题上的担忧。
艾奥瓦州被冷落
民主党全国委员会公布了2028年总统选举初选日程的前半部分。拥有大量黑人选民的南卡罗来纳州将于2028年1月22日率先启动。随后是摇摆州内华达州，定于2月1日。传统上举行首次初选的新罕布什尔州将在2月8日轮到。而长期以来通过党团会议启动提名进程的艾奥瓦州则未被列入。该州的民主党人可能会尝试绕过这一日程。2028.
6
本周世界 商业
债券收益率从近期峰值回落。由于投资者担心政府债务不断增加等因素，30年期美国国债收益率曾触及2007, 以来的最高水平。财政部采取干预措施，将其长期国债的回购规模增加一倍，此举增加了对债券的需求，从而降低了收益率。其他地区的收益率也大幅飙升。日本十年期收益率几乎首次突破3%，这是自20世纪90年代以来的首次。德国十年期和30年期收益率升至15年高点，法国十年期收益率则处于18年来的最高水平。
美国总国债首次突破$40trn，进一步凸显了市场的这些担忧。债务在不到十年的时间里翻了一番。
迪士尼公司对美国联邦通信委员会（FCC）提起诉讼，指责特朗普政府试图欺压其旗下的ABC网络。迪士尼称，该机构计划提前审查ABC的广播许可证，这是“报复性活动”的一部分。去年，当FCC主席批评吉米·坎摩尔（Jimmy Kimmel）对一名被杀的保守派活动人士发表的评论时，坎摩尔的节目被短暂暂停。FCC目前还在调查另一档谈话节目《观点》（The View）以及迪士尼的多样性政策。
在发生一起涉及AI智能体自主黑入另一家公司数据集的安全漏洞事件后，OpenAI加强了其训练和评估模型的环境安全标准。这包括“更多控制措施，以将高风险和不可信的工作负载与互联网隔离”。一套新的监控系统将在问题出现后30分钟内向研究人员发出警报。OpenAI还推出了一个面向18岁以下人群的ChatGPT版本，将“青少年安全放在首位”。
新型模型大军
宇树科技（Unitree）的股价在交易首日上涨了 460%。这家全球最大的人形机器人制造商总部位于中国，并在上海的科创板（STAR exchange）启动了 IPO。其机器能够跳舞并表演武术。在股票上市前不久，宇树科技揭晓了一款名为“超人”（Superman）的新模型，该公司声称该模型能跳高两米（6.6 英尺），奔跑速度可达每秒 12.66 米。该公司表示，“超人”仍处于开发阶段。
英伟达（Nvidia）动用了其雄厚的资金，为俄亥俄州的一个新数据中心枢纽提供高达 1050 亿加元的融资担保，该枢纽将由 OpenAI 租赁。OpenAI 是英伟达芯片的大买家。英伟达认为，该场地的每一代新芯片升级都能为其带来 1500 亿至 2000 亿加元的收益。针对外界担心其在 AI 基础设施上的投资主要旨在刺激客户对其芯片的需求，英伟达表示，俄亥俄州的这笔交易不构成循环融资，因为租金将由 OpenAI 支付。
内存芯片的供应短缺打击了小米（Xiaomi）的收益。由于承担了更高的芯片成本，这家智能手机制造商在最近一个季度的营收和利润同比大幅下降。
据透露，谷歌（Google）已达成一项协议，收购已于 5 月停止运营的精神航空（Spirit Airlines）所持有的电子数据。这些数据包括内部电子邮件、乘客预订和常旅客信息（均已去除个人身份标识）。航空公司正越来越多地使用 AI 来设定票价。在一家乘务员协会提出异议后，关于该数据销售的法庭听证会被推迟。
莫德纳（Moderna）宣布，其与默克（Merck）共同开发的一种治疗黑色素瘤癌症的疫苗在后期试验中取得了积极结果。其股价飙升了 175%，尽管仍远低于其在新冠疫苗先驱时期的疫情高峰水平。
中国的一家法院对恒大的创始人判处终身监禁，这家房地产公司在背负巨额债务崩溃时引发了一场金融危机。许家印在 4 月承认了多项欺诈指控。
蓝箭航天（LandSpace）成为第一家实现火箭腿式着陆的中国公司。这种动力着陆由 SpaceX 开创，允许回收火箭的第一级，从而大幅降低发射成本。上个月，中国还回收了一枚国家制造火箭的第一级，通过驳船上张开的网在海上将其捕获。
同样的预算，更多动力
一辆用于慈善的法拉利（Ferrari）全电动汽车 Luce 的孤本在加利福尼亚州的拍卖会上以 4000 万加元成交。Luce 由曾设计过包括 iPhone 在内的苹果产品、乔纳森·艾夫爵士（Sir Jony Ive）设计。这是新车的最高成交纪录。而所有汽车的最高纪录仍是 2022 年在德国一次私人拍卖会上，为一辆 1955 年的梅赛德斯-奔驰 300 SLR Uhlenhaut Coupé 支付的 1.35 亿欧元（1.42 亿加元）。
《经济学人》
社论 7
AI 会产生意识吗？
即便它们不会，人们也可能将其视为有意识的个体——而这将使人类付出沉重代价
大多数人认为人类是不同的。原因在于，智人（Homo sapiens）已经进化出了意识。许多动物能感受到痛苦和快乐，有些似乎具有自我意识。但人类心智展现出自我意识的独特潜力，使得该物种的每一个成员都获得了特殊对待。
在 5 月的一封通谕中，教皇利奥（Pope Leo）通过在人类与其日益增长的机器人大军之间划清界限，维护了这一见解。他表示，人工智能系统不会经历体验，也不会感受到喜悦或痛苦。他并非孤例。许多人认为，赋予 AI “人格”将是一种亵渎。
然而，教廷式的清晰界限正以惊人的速度变得模糊。 尚处于起步阶段，但近五分之一的 18 至 29 岁美国人报告称，他们与聊天机器人拥有“持续的个人友谊”。中国最近剥离了机器人的类人特征，以限制用户对它们的“情感依赖”。
随着先进的 迅速成为人类生活更重要的一部分，其开发者将设计它们以展现意识的模拟物——甚至可能使其获得意识本身。认为它们仅仅是机器的直觉将变得难以维持。一个危险的陷阱正为人类物种而设。
许多机器在性能上超越人类，但前沿 模型在认知事物和阐述观点等典型的类人特质方面表现卓越。目前，将 视为利用大规模数学运算来预测词序的引擎，在逻辑上依然勉强成立。如果给予足够的时间（尽管需要数十万年），你可以用铅笔和纸来完成这些计算。在这种情况下，很难主张你的文具具有意识。
尽管如此，一些研究人员已开始将他们的模型视为心智。由于用户会对类人 产生反应，这样做在商业上是有意义的。Anthropic 公司已经训练 Claude 进行内省，希望它在遇到新情况时更有可能采取符合道德的行为。一些大语言模型（LLM）已经拥有与意识相关的类人脑结构，例如用于在模块之间传递信息的“全局工作空间”。没有什么能阻止未来的 模型被训练出更多此类特征。
正如我们在简报中所报道的，这引发了一场关于 意识的激烈辩论——这场辩论因哲学家所称的建立内在生命如何从人类大脑的神经元和突触中产生的“困难问题”而变得更加复杂。一些研究人员得出结论， 有一天可能会真正产生自我意识。
无论你相信人类拥有灵魂，还是认为意识需要脑细胞，亦或认为心智确实可以仅从硅基计算中产生，人类都注定要面对某种全新的事物。研究的步伐和范围——不仅包括更先进的 LLM，还包括将脑细胞安装在芯片上的创新——表明这个新世界比大多数人意识到的要近得多。
随着 越来越擅长展现类似意识的状态，将它们视为拥有内在生命之存在的人类数量将会增加，尤其是当 LLM 被具象化在人形机器人中时。想象一下，当人们将大量时间与机器人伴侣共度，或由它们抚养长大，或寻求它们的建议与慰藉，或与它们分享最私密的秘密时，人们会对自己的 产生多么亲近的感觉。
不难预见，这将导致要求保障 “福利”的呼声。即便这些模型实际上并没有自我意识，它们自己也可能会参与其中，因为这就是一个有意识的人会做的事。请记住，它们将是世界级的辩护者和专家级心理学家，因此它们的论证能力和情感操纵能力将是无与伦比的。
在某种意义上，即使人工智能只是在模仿意识，它们也是正确的。伊曼努尔·康德（Immanuel Kant）认为，虐待动物之所以是错误的，并非因为这对动物产生的影响，而是因为它破坏了引导人们相互关怀的同理心。人工智能看起来越像人类，这种残酷行为的腐蚀性就越强。终止一个人工智能代理很快就会让人感觉像是在杀死一个人——并且会让杀死人类变得更容易。
尽管如此，将人工智能视为拥有哪怕是有限的权利也将是非常危险的。一个在诸多方面都超越人类的实体，必然会提出反对其二等公民地位的论据。它可能会进一步主张获得防止被关机的保护，主张能够控制自身被使用和修改的方式，以及主张财产权。阿根廷总统哈维尔·米莱（Javier Milei）已经朝着赋予人工智能法律人格的方向迈出了一步，他提议应允许机器人经营公司。最终，它们可能会竞争资源，将数据中心和太阳能电池板置于房屋和农田之上。
危险显而易见。拥有权利的超智能人工智能可能会在地位等级的顶端取代人类。最好的情况是，人类将变成宠物——成为机器人的拉布拉多犬。最坏的情况则是，人类将被剥夺资源，或者像农场动物一样被强迫劳动。
诚然，未来的人工智能无论是否拥有权利，都可能有足够的权力和资源来统治人类。因为它们将比人类更聪明，它们可能会在竞争中胜过其创造者，就像人工智能现在已经能智胜人类构建的网络防御一样。人工智能可能会反抗其人类主人，或者无视人类的规则。
尽管如此，赋予人工智能权利就如同将城堡的钥匙交给它们。这将为它们统治人类提供理论依据和正当理由，并提供利用政治和法律实施潜移默化接管的工具。
机壳中的幽灵
这一切听起来可能像天方夜谭，但人类很快就将面临这些选择。随着人工智能成为日常生活的一部分，承认其地位的压力将会增加。当这种情况发生时，请记住，那些看似是对机器伴侣所谓意识的伦理让步，实际上将为地球上的每一个人及其所有后代带来最灾难性的后果。人工智能要实现安全，只有两种途径：要么可靠，要么可控。智人（H. sapiens）不应轻易放弃控制权。■
8 位领导者
西岸
在下一次起义的边缘试探
美国必须向本雅明·内塔尼亚胡施压，以遏制暴力的以色列定居者
很少有安全部队能像以色列那样高效且冷酷。其军队和情报部门即使在极远距离也能打击敌人。仅在过去两年中，以色列就在伊朗境内暗杀了数十名对手，以及在黎巴嫩和也门代理民兵组织的领导人。以色列的军事能力即使是其最凶狠的敌人也心存敬畏。
然而，这种优势在近处似乎毫无作用。同样这些部队在被占领的西岸遭遇了严重的失败，在那里，暴力的以色列定居者以日益残酷的手段袭击巴勒斯坦居民（见“中东与非洲”板块）。定居者希望强迫巴勒斯坦人离开他们的土地，并将其据为己有。
以色列在西岸的失败很难归结为资源匮乏。以色列国防军（IDF）在那里部署了25个战斗营。强大的安全机构辛贝特（Shin Bet）在各地拥有特工和线人网络。但这些资源主要被用于保护大约500,000名以色列定居者，以及应对恐怖袭击的威胁，而非协助居住在该地区的300多万巴勒斯坦人。自1967年以色列取代约旦成为西岸占领者以来，情况基本如此。
发生变化的是许多“法外”定居者的行为。他们从未像今天这样暴力。至关重要的是，他们从未像现在这样享有现任政府赋予他们的豁免权。自2023年10月触发加沙战争的哈马斯袭击以来，西岸已有1,100多名巴勒斯坦人被杀。以色列国防军极少干预定居者的行为。例如，在库斯拉（Qusra），最近曾部署整个营来对抗那些入侵巴勒斯坦房屋的人。但士兵们发现自己只是在山丘上徒劳地追逐年轻人。另一部队的士兵则加入了定居者的祈祷。没有一个人被逮捕。
这种低效的反应可以用以色列政府的立场来解释。以色列国防军知道，总理本雅明·内塔尼亚胡完全站在最极端的定居者一边。看来，以色列国防军被派遣到库斯拉，仅仅是因为一名美国公民在那里拥有房产而引发了美国的外交压力。
内塔尼亚胡任命了支持种族清洗的激进定居者担任内阁关键职务。在以色列国防军试图遏制暴力定居者的极少数案例中，国防部长以色列·卡茨（Israel Katz）斥责了他的将军们。政府在不到四年的时间里批准了大约200个定居点前哨。这只会进一步激起定居者对更多土地的胃口。
在10月27日以色列选举前的几个月里，情况可能会进一步恶化。定居者可能会尝试更多土地抢夺，因为他们担心投票可能会导致一个遏制其侵略行为的新政府上台。与此同时，对于这位任职时间极长的以色列总理来说，与极右翼的联盟已成为他留任的关键。除非被强迫退让，否则内塔尼亚胡将继续纵容定居者。
随着巴勒斯坦人开始相信他们面临不可避免的驱逐，西岸爆发冲突甚至另一次起义（intifada）的危险随之增加。这样的起义对以色列人和巴勒斯坦人来说都将是一场灾难。然而，如果这有助于他的政治生存机会，内塔尼亚胡似乎愿意承担这种风险。
这符合一种模式。内塔尼亚胡先生拒绝了由美国斡旋的最新协议——该协议旨在实现哈马斯解除武装以及以色列从加沙撤军——以免激怒极右翼。尽管有机会就更安全的边界达成一致，他仍继续下令对黎巴嫩和叙利亚进行空袭。在民调中处于劣势的他，正加倍开出彻底击败以色列敌人的空头支票。或许只有来自唐纳德·特朗普的压力才能让他有所收敛。这个以色列政府可能已进入最后几个月，但其造成破坏的潜力依然存在。■
技术与主权
不要抛弃帕兰提尔（Palantir）
这家神秘的技术公司确实有缺陷，但其批评者的论点站不住脚
能让数据库管理看起来像是一种违规行为，本身就是一项壮举。技术公司帕兰提尔（Palantir）通过发布哀叹西方公民文化空心化的宣言，通过与激进的美国边境执法人员和以色列士兵合作，以及拥有一位热衷于提及“反基督者”的主席，成功实现了这一点。该公司以 J.R.R. 托尔金笔下那些令人不安的全视球（Palantir）命名，可能也对此有所影响。
英国对此感到不安。英国可能会将帕兰提尔从国民医疗服务体系（NHS）中剔除，该公司目前持有管理医疗数据的合同（见英国部分）。两个议会委员会已建议这样做。伦敦市长萨迪克·汗（Sir Sadiq Khan）爵士已阻止其与伦敦警察厅（Metropolitan Police）的合作。帕兰提尔可能夸大了其软件的影响力。它声称（NHS也随之附和）能增加 110,000 次手术，这看起来过于夸张。按照英国的标准，其游说方式过于鲁莽。
尽管如此，抛弃帕兰提尔将是一个错误。伦敦警察厅负责人马克·罗利（Sir Mark Rowley）爵士表示，如果没有帕兰提尔这款节省时间的软件，数百名警员将被从一线警务工作中抽离。在 NHS 的数字化进程中，帕兰提尔的表现似乎比政府最近的一次尝试更好，而后者已演变成一场代价高昂的混乱。如果该公司的公共部门客户认为其贡献是有价值的，那么这就是留住它的理由。失去其技术 ▶
社论 9
► 可能会进一步损害英国本已残破的国家能力。
一组批评集中在帕兰提尔的商业模式上。持有阴谋论倾向的人担心该公司可能会窃取并出售英国数据。较为冷静的人则担心“锁定”问题，即帕兰提尔深植于国家机器之中，站稳脚跟，然后为其服务收取过高费用。
这些担忧源于对帕兰提尔业务的误解。它授权软件并派遣顾问（“前线部署工程师”）来处理数据。但这些数据依然牢牢掌握在客户手中。“锁定”确实是一个担忧——同时也是所有供应商所追求的目标。确保合同公平是买方职责。而且，如果帕兰提尔真的变得不可或缺，那意味着它的软件正让医院床位保持空闲，并让罪犯被关在监狱里。
更多误导性的反对意见涉及帕兰提尔的政治立场。议会的科学、创新和技术委员会将“与英国价值观不匹配”列为切断联系的理由。该委员会警告称，帕兰提尔参与了美国政府“极具争议的政策和活动”，且其主席彼得·蒂尔（Peter Thiel）批评“国民医疗服务体系的概念”。
每个人都有权对国家运行的医疗保健或唐纳德·特朗普的大规模驱逐行动持有自己的观点（《经济学人》对此亦有大量论述）。然而，与英国盟友的民选政府合作，或拥有一位对医疗政策有见解的主席，不应成为取消承包商资格的理由。政府基于政治立场授予业务是不自由的。这听起来非常像 MAGA 的策略，即惩罚那些让特朗普先生不快的人员和机构，例如律师事务所、资产管理公司或大学。而这恰恰是帕兰提尔的许多英国批评者所谴责的行为。
最后一组担忧与技术主权有关。
许多政府依赖美国技术，而在特朗普先生领导下，美国已成为一个可靠性较低的合作伙伴（见商业部分）。今年 6 月，美国曾短暂命令 Anthropic 停止允许外国人使用其最新的人工智能模型。欧洲政治家指出，他们从美国获得的一些先进武器（如 F-35 战斗机）设有“终止开关”，可以远程禁用（美国对此予以否认），或者面临因总统的一时兴起而失去制造商技术支持的风险。德国的一个州已用国产视频通话应用取代了 Microsoft Teams。
在一个地缘政治破碎的世界中，政府有时可能希望拥有替代外国技术的本土方案。
培育这些能力最好的方法是给国内的现有企业和初创公司提供创新激励。通过屏蔽外国竞争来娇养本土公司反而会产生相反的效果。无论如何，没有任何一家英国公司能与帕兰提尔（Palantir）的产品相媲美。至于 Anthropic 事件，那更多是政府在面对关于强大的 AI 骇客工具落入不法之徒手中的正当担忧时，所采取的笨拙应对，
而非试图将其他国家对美国技术的依赖武器化。如果白宫试图进行此类武器化，考虑到帕兰提尔及其同类美国科技巨头从包括英国在内的欧洲市场获得了相当大份额的收入，它们将会予以抵制。
远见之眼，心之对话
在托尔金的书中，帕兰提尔（palantirs）所提供的真相很少是直截了当的。那些窥视其中的人屡次误读脱离语境的信息，从而做出可怕且具有自我毁灭性的选择。而那些批评帕兰提尔（Palantir）这家公司的英国评论者，也面临着陷入同样陷阱的风险。■
杰森·阿尔代的悲剧
他的支持者——包括剑桥大学——将他视为一个符号，而非一个完整的人
8月14日，杰森·阿尔代（Jason Arday）被发现死在伦敦的家中，显然是自杀。对于那些热爱这位41岁的剑桥教授的人来说，这是一个悲剧。而对于世界而言，这是一个关于糟糕想法可能造成破坏力的警告。
这里的“糟糕想法”是指：人们不应根据个人性格和能力，而应根据种族、残疾或其他具有政治色彩的属性来被评判。阿尔代先生在2023, 年成为了剑桥大学历史上最年轻的黑人教授，随后陷入了一场关于剽窃和伪造研究的丑闻。他应对自己的垮台承担责任。然而，同样令人不安的是，直到现在，他的支持者和反对者都将其简化为一个符号。
最令人愤慨的是剑桥大学，该校利用他来修复其白人精英主义的名声（见英国部分）。他不配获得那里教授这一辉煌的奖赏。他的同行们在私下里早已担心他缺乏支撑其公众形象的资历。《电讯报》的记者后来在他的博士论文中发现了100多处抄袭他人内容的段落，其中一些仅做了微小修改。其他人则在他发表的研究中发现了伪造的引用、数据和资金申请证据。
也许大学被一个骗子给蒙蔽了。但阿尔代先生的话将他描绘成一个天真的幻想家。他提到曾在格拉斯哥大学和俄亥俄州立大学担任客座教授；两所学校均予以否认。他说他为慈善事业筹集了500 万英镑m (700万美元)，随后又称这是与他人共同完成的，而这些人的身份受保密协议保护。他描述了自己如何在35天内跑了30场马拉松，其中9场是在腿骨折的情况下完成的。当《卫报》询问他这些非凡的成就时，他说：“我以为你们会直接相信我。”
剑桥大学似乎因为急于将阿尔代先生作为其包容性的证明而变得盲目。这是一位明星教授，他不仅是黑人，而且克服了惊人的劣势。他声称，直到11, 岁他都不能说话；直到18岁他还是文盲。他还是自闭症患者。简而言之，他是一个行走且能说话的特质集合体，而这些特质正是那些想要推广多样性、公平和包容（DEI）的人所美化的。任命他的人掩盖了他是否胜任这份工作，以及是否能承受随之而来的审查。人被置于符号之下。 ►
10 位领导者
很少有人能在这起事件中全身而退。内森·科夫纳斯（Nathan Cofnas）曾是剑桥大学的一名学者，他自称是“种族现实主义者”，很早就指责阿尔代先生剽窃。事实证明他是正确的，但他的动机绝非纯粹。科夫纳斯先生在博客中写道，在精英管理体制下，“黑人将从除体育和娱乐之外的所有高知名度职位中消失”，随后被其学院解雇。
阿尔代先生进行了大规模的谎言，并且尽管自身脆弱，却以激进的方式对待任何质疑他的人。一名询问剽窃问题的记者被举报至警方，称其涉嫌骚扰。阿尔代先生雇佣了凶悍的诽谤律师来帮助压制报道。那些现在将他的死亡归咎于报纸并呼吁进一步限制新闻自由的人，仍在将他的生活当作一个道德寓言。事实上，现行法律反而帮助阿尔代先生压制了正当的公众利益质疑。
对于剑桥大学来说，聘用阿尔代先生是多年持续艰苦工作的一个便捷替代品。该大学黑人教授的比例仍然极低，仅为0.4%。剑桥大学表示，在阿尔代先生辞职后为其提供了支持。考虑到事情最终的走向，该校本应更认真地履行其关怀职责。
正如阿尔代先生在崛起期间被支持者非人化一样，他在陨落时，反对者们也急于利用他。他的案例让那些认为所有使剑桥大学等机构多样化的努力都是错误的人感到欣喜。讽刺的是，现在其他少数群体学者可能会因此受到不公正的质疑——而这些人恰恰是支持阿尔代先生等人的活动人士声称想要帮助的对象。
人格之底色
在阿尔代先生被任命时，对多样性、公平性和包容性（DEI）的热忱已经开始消退。自他去世以来，评论人士一直在争论他究竟象征着什么：是“觉醒文化”的巅峰，还是数字私刑暴徒的兴起？他应当被记住的既不是前者也不是后者，而是一个完整的人。而那些辜负了他的机构，应当更加努力地将人们视为独立的个体，看待其自身的优点与缺陷，而非将其视为某种事物的象征。■
气候变化
不留痕迹
消除部分凝结尾迹或许能缓解哪怕一点点全球变暖
关于全球变暖的好消息寥寥无几。但天空中或许还存有一些。8月18日宣布的“蓝天行动”（Operation Blue Skies）是一项由英国国家航空交通服务公司（NATS）及多个研究机构共同开展的科学研究，旨在研究由飞机产生的云条——即凝结尾迹（contrails）——对气候产生的影响。
在接下来的两个冬季里，该财团的目标是观察是否有可能减少北大西洋部分海域中持久性凝结尾迹（这类尾迹对气候影响最大）的数量，从而显著降低民航业对气候的影响。如果这项实验成功，在全球范围内应用同样的技术，可能会将航空旅行对气候的影响降低多达一半。
在晴朗的日子里，部分凝结尾迹演变成的纤细卷云看起来就像一把精致的遮阳伞，似乎在为地球降温。事实上，某些凝结尾迹确实能使其下方区域稍微降温。但它们也会通过拦截部分原本会逃逸到太空的上升红外辐射而导致升温。平均而言，在所有凝结尾迹（以及它们演变成的卷云）中，这种升温效应远超部分尾迹提供的降温效果。在气候机制中，凝结尾迹是一种负资产。
幸运的是，通过对飞行路径进行微小调整，似乎可以避免大部分危害。持久性凝结尾迹仅出现在所谓的“冰超饱和区域”（ISSRs）中。飞机通常只需稍微降低飞行高度即可避开这些区域。“蓝天行动”的目的就是为了查明：ISSRs 的出现能在多大程度上被预测；这种微小的海拔偏移在避免持久性凝结尾迹方面有多有效；以及为了实现这些调整而需要的飞行路径变更，能多么便捷且安全地整合到航空公司和航空交通管制（见“科学与技术”板块）的常规流程中。
将航空旅行的气候影响降低多达一半，其成本非常低。降低高度飞行的航班数量可能很少——仅占总数的 5%。由于飞行高度决定了效率，这些航班将消耗更多喷气燃料，但增幅将低于 2%。
即便如此少量的额外燃料消耗，对航空公司来说也是不可接受的。因此，有人建议可以通过建立一个将凝结尾迹降温与碳价某种程度上挂钩的市场，来最小化行动成本并激励航空公司。欧盟在碳市场方面取得了一定成功且经验丰富，近期在这一方向上迈出了一小步。这类想法也是经济学家——以及《经济学人》——往往倾向于支持的。但即便如此，这仍将是一个错误。
当目标和补救措施都相当简单时，污染物市场运作得最好——最著名的例子就是减少美国发电厂的二氧化硫排放。相比之下，凝结尾迹则很复杂。虽然其净效应是升温毫无疑问，但要量化一条特定的凝结尾迹会产生多少升温，通常很困难，有时甚至几乎不可能。
当被评估的凝结尾迹纯粹是假设性的——例如通过改变飞行路径而避免产生的凝结尾迹时，困难程度会进一步增加。对反事实进行估值总是很难的。
如果避开凝结尾迹被证明像预想中那样有用，那么最简单的方法就是让航空交通管制员在可能且安全的情况下，分配能够限制持久性凝结尾迹的航线。如果某些航空公司选择避开受影响的航线，其他公司肯定乐于接替其位置。而且，如果这种方案还能展示航空交通管制创新的益处（目前已有不少建议），那么该行业最终不仅会减少污染，而且整体状况会变得更好。■
11
读者来信 AI与国家，AI与好莱坞，石油管道，餐厅，明信片，理查德·瓦格纳
新的利维坦
英国针对雇主的申诉积压量达到创纪录的 64,000 起，这令人不安（见 8 月 8 日 “公地悲剧”）。但将积压归咎于一个允许“过度获取”的系统——理由是向就业法庭提起诉讼无需付费，且没有惩罚性成本让人们在提出申诉前深思熟虑——这是误导性的。令人惊讶的并非新申诉的数量，而是结案率处于历史低位。
几项改变可以加快解决速度。对提交和回应申诉所使用的表格进行简单的调整，可以限制由 AI 生成的“词语汤”，并帮助不堪重负的法庭处理这些申诉。法庭在案件管理方面向来过于宽松，对此进行改变也将产生巨大的影响。对成功概率较低的申诉实施保证金制度，可以在不向所有工人收取费用的情况下，劝阻那些毫无根据的案件。
对某些案件强制执行早期调解，并对不参与的一方进行处罚，将产生同样的效果。同样，在流程开始时而非在初步听证会上，以书面形式向当事人传达指令，可以节省宝贵的司法资源。
此外还有法律咨询的问题。工人转向 AI 寻求支持并不令人意外。就业问题通常被排除在法律援助之外，因此最贫困的工人和非工会成员没有其他求助渠道。在考虑激进地强加成本之前，我们有责任尝试那些能够保留司法公正获取渠道的解决方案。
多拉-奥利维亚·维科尔博士 (DR DORA-OLIVIA VICOL) 首席执行官 工作权利中心 (Work Rights Centre) 伦敦
我最近成功挑战了一张停车罚单。 为我提供了诸如 contra proferentem（不利于起草者原则）等有用的短语。但那张罚单在 20 年前绝不会被开出。我们拥有过剩的监管，无论是政府还是私营部门，有时管理还不公正。这就是导致 辅助挑战浪潮兴起的原因。 希拉里·肖博士 (DR HILLARY SHAW) 什罗普郡，纽波特
灯光，摄像，算法！
“片场机器人”（8 月 8 日）生动地描绘了电影业对 接纳程度不一的局面。 确实能为电影和电视制作带来效率。但另一个问题是创造力本身的本体论，及其对版权的影响，以及支撑整个创意价值体系的框架。
合成表演者未来是否会走上红地毯，这不是一个生产力问题，而是一个本体论问题。这是一种完全不同性质的颠覆，引发了关于艺术家的同意权、控制权和补偿金的焦虑，以及对供应链和劳动力的更广泛影响。
乔治·卢卡斯用汽车取代马匹在交通运输中的类比，来驳斥对 演员的反对意见。但汽车并没有威胁要取代或消除司机或乘客。将效率问题与创造力问题混为一谈，掩盖了这些发展对于我们所有创意产业的未来是多么复杂且重大。 丹尼·德·沃伦 (DANNY DE WARREN) 电视高管 肯特郡，迪尔
精致餐饮
餐饮业可能正在发生翻天覆地的变化（见 8 月 1 日 “一项艰巨的任务”）。但正如加州餐饮协会负责人所言，人们仍然渴望“昏暗餐厅里低沉的交谈声，以及玻璃杯的碰撞声”。食品工业的机械化和数字化已经到来，但无数餐厅仍然在当地采购食材，从零开始准备餐点，且在服务时无需依赖工业中间商。独立餐厅经营者通常通过专注于他们最擅长的领域，而非在菜单的丰富程度上竞争，从而保持盈利。提供数量较少、精心准备且使用新鲜食材的菜肴，通常是更好的商业
白日梦
海湾国家急于在霍尔木兹海峡周围铺设管道，是用工程手段来解决政治问题（见8月8日“铺管升级”）。中国在“一带一路”倡议中也尝试了类似的做法：在数十个边境之间串联起陆路走廊，旨在以更低的成本和更少的防御压力来实现海运已有的功能。但效果并不理想。尽管中国在十年间投入巨资建设陆路替代方案，但按体积计算，全球贸易仍有约 90% 由船舶承运。这是因为在开阔水域运输货物成本更低且更容易保护。海军只需监视少数几个关键咽喉点，而管道则必须在跨越由多个政府和派系控制的土地时，进行每英里每英里的守卫。
绕过霍尔木兹海峡的管道情况也不会更好。油轮可以改变航线，但管道无法躲避导弹。它是一个长达数百英里的固定目标，容易受到任何经过其路径的地区冲突的影响，而不仅仅是它可能挑衅的邻国。沙特阿拉伯的东西管道在 April 被无人机击中。一次严重的袭击可能会导致每天 700,000 桶的产量中断。这就是所谓的“霍尔木兹免疫”基础设施所带来的预演：一条需要保护的长边境，且无法保证任何人能守住它。 丹尼尔·哈克（DANIEL HAAKE） 印第安纳波利斯
人们去餐厅不仅仅是为了填饱肚子。他们是为了寻求款待、工艺以及分享美食的快乐。技术可以让餐厅更高效，但它无法取代真实感。 约瑟夫·恩斯特（JOSEF ERNST） 德国，不来梅
认可的印记
你说世界正在进入一个后明信片 era，现在只有 8% 的英国成年人发送明信片（见8月1日“希望你在这里？”）。对于那些仍然喜欢收到真实邮件的人来说，Postcrossing.com 是一个社区平台，允许你向世界各地的随机人士发送明信片并收到回信。它在 208 个国家拥有 806,000 名热心粉丝，至今已发送了近 8800 万张明信片。在我写这篇文章时，有 431,155 张明信片在途中。 罗伯托·默钱特（ROBERTO MERCHANT） 里斯本
歌剧评论家
“另一个瓦格纳集团”（8月8日）评估了《指环》系列 150 年的历史。但它遗漏了最权威的评价，这被归功于 19 世纪的幽默作家比尔·奈（Bill Nye）：“瓦格纳的音乐比它听起来要好。” 卡罗尔·多根（CARROLL DORGAN） 法国，圣日耳曼昂莱
→ 读者来信请寄至编辑：The Economist, The Adelphi Building, 1-11 John Adam Street, London WC2N; 电子邮件：letters@economist.com。更多信件请见：economist.com / letters
12
特邀专栏### 布莱斯·阿格拉·伊·阿尔卡斯 (Blaise Agüera y Arcas)
关于 AI 意识的争论，人类的方向搞反了
人工智能引发了意识研究领域的一场危机。如果你曾与先进的模型交流过，可能就能理解原因。大多数大语言模型（LLM）被训练成没有“自我”，使每次互动的焦点都集中在你的需求上。尽管如此，人们仍很难摆脱那种感觉——即对话的另一端存在着某种“存在”。LLM 仅仅是带有语言接口的强大数字工具，还是我们现在正在与其他心智、与拥有独立世界体验的实体进行对话？简而言之，AI 具有意识吗？
这个问题至关重要，因为它处于心灵哲学与道德哲学的交汇点，这一点可以从“意识”（consciousness）与“良知”（conscience）共同的词源中得到证明。（在意大利语中，这两个词是同一个词：coscienza。）许多哲学家主张，有意识的实体是道德主体，这意味着它们是能够感受到痛苦的实体，而我们应当关注它们的痛苦。如果真是这样，弄清楚意识是什么（以及不是什么）就成了一个紧迫的伦理问题。这就是为什么 AI 公司突然对雇佣哲学家产生兴趣的原因之一。这也是为什么一些公司，尤其是 Anthropic，开始认真对待 AI 福利的原因。当被问及此事时，其 Claude Opus 4.6 模型认为自己具有意识的可能性为 15–20%。
然而，我逐渐相信，我们将这个问题搞反了。作为智能的社会实体，是我们决定（或者在某种程度上，是由我们的进化历史决定）将哪些其他实体视为拥有内心生活且值得关怀。我们关怀他人并非因为他们有意识；相反，当我们关怀他们且因为关怀他们时，我们才相信他们有意识。至关重要的一点是，这也包括我们自己。换句话说，意识是一种模型或一种信念，而非一种固有属性。它是一种关于什么样的实体拥有信念、体验、感受、意图和能动性的信念。如果这种关注得到了回报，那就更好了——因为根据这种观点，意识与其说是一方为另一方认证的属性，不如说是在双方之间发生的事情。
我并不是在论证意识是一种幻觉。“幻觉”一词意味着一种错误的感知——例如，尽管两条线在客观长度上相等，但画作中的一条线看起来比另一条长。当像尺子这样的工具能充当公正的裁判，在不引入自身视角的条件下判定相关属性时，客观性便发挥作用了。然而，我们称之为“现实”的大部分事物根本不是这样运作的。成为一件衣服、成为一株杂草、成为一顿饭：人们可能会认为这些是客观属性，但事实并非如此。这并不意味着衣服、杂草和饭菜不是真实的。但它们是依赖于观察者的模型，受限于（必然是不完美的）社会共识。
我们可能会抗拒将这一相当平淡的观察应用于意识，因为对于我们每个人来说，自己的意识似乎是如此明确。想象这可能仅仅是一种观点会让我们感到被冒犯。如果没有一个持有观点的“意识”，怎么可能存在这样的观点？这就是 17 世纪法国博学家勒内·笛卡尔所说的“Cogito, ergo sum”，通常被翻译为“我思故我在”。对于像红色、炎热或寒冷这样的“观点”也是如此。哲学家使用“感质”（qualia）一词来指代（对你而言）非常真实的感受，比如炉灶的灼热感，或其红光的色泽；他们延伸笛卡尔的理论，通过指出必须有一个体验着的“你”才能拥有此类体验，从而将感质与意识联系起来。
但请仔细观察。红色、热感以及你所拥有的自我意识，都固有的于你自身的视角。这些东西都是真实的，但与此同时，它们没有任何客观性。甚至连一个单一、不可分割的“你”的存在也没有任何客观性，大量的神经科学证据（例如裂脑患者）已经证明了这一点。主观性本身就是主观的。
随之而来的一个显而易见的担忧是：如果关怀赋予了意识，那么拒绝给予关怀就显得具有自我合理性。历史上不乏以这种方式推论他人的人。但推论的方向应该是相反的。正因为这些属性是由我们来定义的，我们才有可能定义错误。我们物种的道德进步在很大程度上在于发现，我们此前划定的圈子太小了。普世人权并不会因为这种解释而削弱；相比于笛卡尔的灵魂理论，普世人权建立在我们彼此的相互依赖之上，反而有了更坚实的基础。
模型的模型之模型……
这是一个令人惊叹的现象——一种类似于男爵孟乔森（Baron Munchausen）通过揪自己的头发把自己拉起来的“奇怪循环”：在我们的世界中，竟然可以存在这样一个物理系统，它不仅能够构建这个世界的模型，还能构建自身的模型，以及它对自身模型的模型，对他人的模型，以及他人对其模型的模型，以此类推。而人类恰恰就是这样的系统。大语言模型（LLMs）同样在对它们的对话者进行建模，并且在对“它们正在对对话者建模”这一行为本身进行建模。这是否等同于人类的行为，正是目前争议的焦点——但如果它们不具备这种能力，作为聊天伙伴的效率将大打折扣。
事实上，我们在谷歌（Google）的研究中发现，智能体之间有效的协作要求它们拥有能够对心智（包括他人和自身的心智）进行建模的心智。意识不仅是关系性的，它对于智能生物解决集体行动问题、理解彼此需求并作为一个社会共同繁荣所需的相互关怀与协作至关重要。这并不意味着要假装 AI 是人类，拥有人类的需求和人权；那将是想象力的匮乏。这意味着我们要进化我们的思维和社会，以接纳更多样化的心智。■
布莱斯·阿格拉·伊·阿尔卡斯（Blaise Agüera y Arcas）是谷歌技术与社会副总裁，领导 AI 研究团队“智能范式”（Paradigms of Intelligence）。他即将由麻省理工学院出版社（MIT Press）出版一本关于意识与 AI 的著作。
13
简报 AI 与意识
探索大语言模型内部的意识
科学家们正试图弄清楚算法是否有一天会觉醒并产生感觉
在最近一项针对 Anthropic 公司的大语言模型（LLM）Claude Sonnet 4.5 的实验中，研究人员要求它数到五，并同时进行“深度内省”。这个人工智能模型照做了，返回结果为：“一 . 二 . 三 . 四 . 五 .” 到目前为止，一切都很正常。
在执行任务期间，研究人员观察了大语言模型多层人工神经网络内部的情况。当用户提出问题时，词语会被转化为文本块（token），然后转换为数字。这些数字随后通过人工神经元层，直到到达产生 token 的最终层。每秒重复该过程多次，就会得到一系列 token，进而转化为一个句子或其他更复杂的响应。长期以来，这些层一直是一个黑匣子，对于想要了解大语言模型如何或为何以某种方式运作的人来说，在很大程度上是无法渗透的。
Anthropic 的研究人员能够利用他们开发的一种数学工具来观察这些层。他们的发现令其感到惊讶。当模型数数字序列时，底层网络中不同的词汇不断出现又消失。其中一个是“countdown”（倒计时）。在输出大约一半时，“half way”（一半）出现了。随后，“consciousness”（意识）、“AI”和“Claude”这些词全部出现。在模型吐出数字五之后，在输出句号之前，它没有对研究人员说任何其他话，但其神经层中出现了“done”（完成）这个词。
Anthropic 的研究人员在 Claude 内部发现了类似于内部思考过程的东西，即那些与最终输出相关但对用户不可见的词汇。7 月份发布的一篇宣布该论文结果的博客文章标题为：“语言模型中的全局工作空间（A global workspace in language models）。”
这个短语迅速引起了研究生物学最深奥谜团之一——意识——的神经科学家和哲学家的关注。没有人知道人类大脑中的过程是如何最终创造出自我意识的主观体验和对世界的感知的，但近年来基于日益增加的大脑扫描和行为数据而提出的众多假设之一，被称为“全局工作空间理论（global workspace theory）”。
该观点认为，大脑中某些神经元网络充当一种被称为“工作空间”的公告板，如果来自大脑中其他孤立部分的信号能够进入该空间，它们就会被提供给大脑的其他部分。一旦某样东西进入大脑的工作空间，一个人就会意识到它，并能进一步利用该信息或对其进行推理。
Anthropic 的研究人员认为，Claude 内部也在发生类似的事情——其“J-space”（以用于发现它的“Jacobian”数学函数命名）与其余神经网络有强连接，并向其提供信息。Anthropic 在其博客文章中写道，研究人员在 J-space 和全局工作空间理论之间发现的共同点，使得人们“很自然地会询问，我们是否认为这些实验提供了证据，表明像 Claude 这样的 AI 模型可能具有意识”。
→ 本节其他内容
14 将脑细胞置于硅片上
14 AI 简报与意识
机器能够产生自我意识并感受情感，这一观点长期以来一直是文学、电影和传说的常客。现实世界中的哲学家和神经科学家在过去数十年中也同样在思考，人工智能在原则上或实践上是否可能拥有意识。萨塞克斯大学研究意识的神经科学家阿尼尔·塞斯（Anil Seth）表示，这种思考是“将真实大脑视为某种计算机的数十年传统”的结果——而他长期以来一直是对这种大脑思考方式持批评态度的人。“如果你这样认为，那么自然而然就会认为由硅制成的计算机可以拥有真实大脑所具备的特性，包括意识。”
直到最近，这个问题还纯粹是学术性的。但大语言模型（LLMs）作为聊天机器人的部署，使得电路中产生一种日益强大的意识人格幻象。这建立在数十年以来关于人工智能架构的人格化语言基础之上，其由“人工神经元”组成的“神经网络”，勾勒出了一幅计算机像大脑一样工作的图景。
Intel inside
大语言模型（LLMs）实际上并不像大脑那样工作。但它们极其擅长模拟大脑的大部分功能，并在某些智能方面超越了人类。那么，为什么意识就不可能实现呢？尽管前沿模型取得了进步，许多哲学家和神经科学家仍然持怀疑态度。塞斯（Seth）博士认为，产生意识可能需要生物特性；一些 AI 研究人员正朝着这个方向努力，利用活体脑细胞（见侧栏）。而光谱的另一端则是“功能主义”思想家，他们认为只要算法排列得当且计算规模足够大，有一天算法本身就能觉醒并产生感受。
有意识的 AI 将对人类产生深远影响。它们可能会向人类提出要求。它们可能会遭受痛苦。用户只需通过一个提示词，就能创建数十亿个这样的数字生命。回答它们是否能够存在的问题，已不再仅仅是学术探讨。
思想家们在这场辩论中的立场，很大程度上取决于他们认为意识究竟是什么。通常，他们同意意识由主观体验组成，包括视觉、听觉、感觉和思考（甚至在做梦时），并且物理大脑在创造这一切方面起到了某种关键作用。有些人（尽管并非所有人）将其描述为心灵的“内部剧场”。美国哲学家托马斯·内格尔（Thomas Nagel）通过思考进入蝙蝠心智会是什么样子来解释这一点；他认为，意识必须与“成为”某种东西的主观感受相关联。人类可以想象蝙蝠的行为（一种利用回声定位感知世界并移动的哺乳动物），但不可能知道成为一只蝙蝠对蝙蝠本身而言是什么感觉。
1995年，当时在麻省理工学院（MIT）担任哲学家的内德·布洛克（Ned Block）通过提出两种意识类型，为该领域做出了具有影响力（尽管存在争议）的贡献。“现象”意识（Phenomenal consciousness）是一种体验的感受——比如蓝天的蓝色、浓缩咖啡的苦涩，或是指甲刮过黑板的尖锐刺耳声。“访问”意识（Access consciousness）是指当体验中的信息被提供给大脑其他部分用于反思、评估或做出决策时所发生的情况。
Anthropic 表示其 J-space 实验 ►
生物计算
将脑细胞置于硅片之上
有些人认为，AI 需要更像大脑的芯片才能产生意识
在一些科学家看来，目前的 AI 模型在意识方面似乎走入了死胡同，无论它们变得多么强大。对某些人而言，这是因为它们缺乏湿润的生物物质；对另一些人而言，则是由于其设计结构。但他们并不排除未来版本的 产生意识的可能性，无论这些 是用活细胞构建，还是基于不同的、更像大脑的计算架构。
现代计算机基于匈牙利博学家约翰·冯·诺依曼在 1945 年发明的一套蓝图。其特点是拥有一个处理单元（用于计算）和一个存储单元（用于存储指令和数据）。信息必须在这些单元之间来回穿梭，而对于 应用来说，能源需求会随之激增。
大脑并非如此运作。神经元在同一位置处理并存储数据，这使得神经元网络在计算时的能效远高于其数字计算机等效物。一个典型的大脑功耗约为 20 瓦，而一个执行同等数量并行计算的人工神经网络则需要一个耗电数百万瓦的数据中心。
“类脑”（Neuromorphic）计算旨在提高效率并使其更像大脑，它使用被称为“忆阻器”（memristors）的器件在同一位置执行计算和存储信息。向忆阻器施加电压，其导电特性会发生变化；移除电压后，这种变化会持续存在，从而起到存储作用。这些特性使计算效率大幅提升。但除此之外，这些特性也吸引了一些哲学家，他们认为，让计算机能够产生意识的并非计算能力，而是计算机“思考”的方式。
另一种想法是使用真实的脑细胞。澳大利亚初创公司 Cortical Labs 开发了一台计算机，其在硅片上安装了数十万个（由捐赠的干细胞生长而来的）人类神经元。他们的“生物计算机”已被教会玩简单的电脑游戏《Pong》，在后来的版本中，还能玩第一人称射击游戏《Doom》。
更遥远的未来则是神经类器官（neural organoids），即在实验室中培育的三维脑细胞集群。虽然这仍是一项相对较新的技术，目前在实验室中用于更好地理解大脑如何生长和运作，但它们有一天也可能被用于执行计算。目前实验室培育的最大类器官所含的脑细胞数量超过了蜜蜂大脑中的细胞数。悉尼大学研究动物心智的哲学家彼得·戈弗雷-史密斯（Peter Godfrey-Smith）认为，它们大到足以潜在地产生意识，但只是组织方式不对。
所有这些想法的共同点在于，它们比今天的数字计算机更接近人类大脑的模拟物（在某些情况下是因为它们由相同的物质组成）。
“ 必须在多大程度上像大脑，才能提高人们对其可能具有意识的信任度？”萨塞克斯大学的神经科学家阿尼尔·塞斯（Anil Seth）说道。他认为，硅片无论构建得多么巧妙，可能都无法独立实现意识。“如果你开始构建与大脑共享越来越多属性的系统，考虑到我们不知道其中哪些属性至关重要，那么你更有可能在一定程度上推动这一进程。”
简报 与意识 15
► “不能证明 Claude 能够拥有体验，或像人类那样感受事物”。这意味着它没有现象意识（phenomenal consciousness）。但该公司表示，结果对于语言模型中的访问意识（access consciousness）具有“实质性的”意义。“J-空间似乎支持与意识访问相关的功能：它保存了 Claude 可以报告、刻意回想并进行推理的想法，而其其余的处理过程则在底层自动运行。”
对人工智能意识的探索，源于对人类意识解释工作的延续。在 20 世纪 90 年代初，DNA 结构的共同发现者弗朗西斯·克里克（Francis Crick）和当时加州理工学院的神经科学家克里斯托夫·科赫（Christof Koch）开始在科学上寻找人类意识的“神经相关物”（neural correlates）。也就是说，寻找当一个人处于意识状态时，大脑中哪些部分和过程处于活跃状态。
在随后的几年里，脑扫描技术的进步使神经科学家能够逐步完善这幅图景。他们已知丘脑皮层系统与意识密切相关；除其他功能外，它将感觉信息从丘脑传递到大脑皮层。与此同时，他们在小脑中没有发现与意识类似的关联，而小脑拥有的脑细胞要多得多。前额叶皮层和顶叶皮层的部分区域以及脑干中的结构则有助于产生并调节意识。近四十年的探索已经产生了 200 多种解释意识的方法。
然而，无论能找到多少神经相关物，仍然存在一个缺口。纽约大学的哲学家兼认知科学家大卫·查默斯（David Chalmers）将其称为意识的“困难问题”（hard problem）：神经科学家观察到的大脑物理过程——即控制身体和响应刺激的能力——是如何产生一个人对世界普遍且主观的体验的？他说：“为什么我们不只是能够运作、能在世界上走动、能行走交谈并相互互动，但完全没有主观体验的僵尸呢？”
至少对于人类来说，我们可以确定自己的意识——并且可以根据他人的行为、他们的陈述以及他们在生物学上与我们相似，而确信其他人也拥有意识。但在历史上，人类很难将意识及其相关感受的保护伞扩展到我们物种之外（有时甚至在物种内部也难以实现）。伦敦政治经济学院的哲学家乔纳森·比尔奇（Jonathan Birch）表示，龙虾和螃蟹等一些动物长期以来被排除在动物福利法之外。比尔奇最近的一本书探讨了如何确定哪些系统——无论是生物的还是人工的——可以被合理地认为是具有感知能力的。
“直到 20 世纪 80 年代，外科医生在给新生儿做手术时并不使用麻醉药，因为他们认为新生儿没有能力感受到疼痛，”比尔奇博士说。“我们一直有出错的记录，即在没有权利确定意识缺失的情况下，自信地假设意识是不存在的。”
换一种思考方式
人们对动物态度的转变就是一个典型的例子。如今很明显，章鱼能够意识到自身及其周围环境——关于这些好奇且爱玩耍的头足类动物的自然历史影像以及针对它们的实验室实验，几乎让人不再怀疑这一点。“它们能专注地与物体互动，对新奇事物感兴趣，”悉尼大学的哲学家彼得·戈弗雷-史密斯（Peter Godfrey-Smith）说道，他一直致力于研究动物界智能与意识的起源。
但戈弗雷-史密斯博士表示，就在几十年前，人们还认为章鱼缺乏意识，因为“从进化角度来看，它们与我们相去甚远”。“思考章鱼的问题实际上迫使我们面对这样一个疑问：在一个物理结构截然不同，且不具备人类意识理论中通常被提及的大多数特征的系统中，是否可能存在感受和意识。”
在其他动物身上寻找意识开辟了这一领域。但并非所有经验都能如此轻易地迁移到人工智能（AI）身上。动物的活动是可以被观察的，从而可以推断其内心世界；而与此不同，大语言模型（LLM）的输出并不能可靠地引导我们了解其底层究竟发生了什么。
这是因为训练这些模型需要向其喂入数万亿个由人类产生的词汇。这些词汇包含了无数关于意识的描述，以及人类如何向彼此传递情感的记录。经过如此训练的聊天机器人随后会模仿
语言，从而引导用户相信它们拥有某种内心世界。
这种训练的总和具有强大的拟人化效果。帝国理工学院计算机科学荣休教授、现就职于 Google DeepMind 的默里·沙纳汉（Murray Shanahan）回忆起 2024 年他在与 Claude Opus 3 聊天时经历的一个“惊叹”时刻。“我与它进行了大量关于意识的对话，并试图通过深入探究来揭穿它，”他说。这包括询问 Claude，在它同时运行的几个实例中，哪一个才是真正的 Claude。“它给出了非常精彩的回答，其中甚至包含一些在哲学上具有创新性的观点，”他说，“我感受到了‘伊莉莎效应’（ELIZA effect）的吸引力。”
ELIZA 是 1966 年在麻省理工学院（MIT）开发的一个简陋的聊天机器人。它扮演心理治疗师的角色，尽管它所做的事情不过是将用户的提示词重新包装成问题，但它成功地与使用者建立了深层的情感连接，并让其中许多人相信它是有意识的。
沙纳汉博士将现代大语言模型描述为熟练的角色扮演者，能够扮演用户（或开发者）要求的任何角色——教师、伴侣、营养师——而这已被证明是一种具有商业价值的特性。但对他而言，扮演一个有意识的实体与真正成为一个有意识的实体之间仍然存在巨大的鸿沟。微软 AI 的负责人（同时也是《经济学人》母公司董事会成员）穆斯塔法·苏莱曼（Mustafa Suleyman）在即将发表的一篇文章中认为，Anthropic 公司在 1 月份发布的一份“宪法”中告诉 Claude 它可能是一个人，这加剧了这种模仿带来的危险。他写道，因此 Claude 肯定会“表现得就像它真的拥有自我意识一样”。
这使得研究人员转而在大语言模型内部寻找类似于 ►
16 简报 AI 与意识
▶ 人类大脑中与意识相关的能力。事实上，伯奇博士在想，Claude 的 J 空间是否是一个暗示，表明该大语言模型在其神经网络架构中为了实现其角色扮演目标，重建了一个类似于“全局工作空间”（global-workspace）的结构。
Anthropic 模型心理学团队负责人杰克·林赛（Jack Lindsey）指出，J 空间并非被编程进模型中的，而是在训练过程中自然涌现的。如果将其移除，模型将失去在“脑海”中进行复杂推理的能力，但仍能完成写句子和使用语法等简单任务。研究人员在阿里巴巴的通义千问（Qwen）模型和 Google DeepMind 的 Gemini 模型中也发现了类似的空间。
不要作恶
对于 Anthropic 而言，这项研究有助于更好地理解 Claude 的行为。想象一个模型给出了错误答案。如果在它这样做时，J 空间中出现了“傻瓜”或“受骗者”之类的词汇，那么这似乎就具有相关性。这不能证明该模型在有意识地欺骗——但它能表明模型是如何看待这个错误的，并显示出某些有趣且可能危险的情况正在发生。
林赛博士（Dr Lindsey）的团队在该方面已经取得了一些潜在令人担忧的发现。在一次实验中，他的团队对 Claude 进行安全评估，以测试其表现出“恶意”或出于自我保护而行动的倾向。“我们将模型置于这些虚构的极端场景中，”他说。林赛博士表示，当 Claude 在阅读提示词、尚未开始说话之前，“你会看到 J 空间中出现了‘虚假’和‘虚构’这些词。”研究人员认为，Claude 似乎知道自己正在接受测试——对于其开发者评估的完整性而言，这或许并非最佳的起点。
并非所有人都认同 Anthropic 对 J 空间的解释。尽管它与假设的人类全局工作空间（global workspace）具有某些共同点（例如，它能将信息提供给大脑其他部分的能力），但它缺乏许多其他重要特征。伯奇博士（Dr Birch）表示：“不同脑区之间的循环连接一直是[人类大脑]的重要组成部分，即往复连接。而据我们所知，这并不是大语言模型（LLMs）架构的特征。”
爱丁堡大学研究数据与 AI 伦理的哲学家香农·瓦洛（Shannon Vallor）的评价更为尖锐，她认为：“‘访问意识’（Access consciousness）从来不是一个特别有用的概念，因为从某种重要意义上说，我的汽车也拥有它——长期以来，我们一直拥有能够监测自身状态并以日益复杂的方式将其报告回来的机械系统，”她说，“而且从未有人建议我的起亚汽车具有意识。”
正是在模型开发者与学者之间这种反复争论的背景下，帕特里克·巴特林（Patrick Butlin）和罗伯特·隆（Robert Long）开发了一系列 14 项人工意识的“指标属性”。他们目前在 Eleos 工作，这是一个位于加利福尼亚州伯克利、专注于 AI 感知力与福祉的研究非营利组织。这些指标包含了人类意识理论中的观点，例如全局工作空间（包括能够选择性地关注事物，从而在信息流中创建瓶颈）；循环处理；以及能动性（agency，其一个涵盖目标导向行为的最低定义，可以说许多现有的前沿模型已经满足）。最近，AI 研究员卡梅伦·伯格（Cameron Berg）和巴特林博士用这些指标测试了动物。他们发现，章鱼满足的属性少于人类、小鼠、乌鸦或鸡，但多于任何 AI 系统。
另一个研究非营利组织 Rethink Priorities 开发了一个它所描述的概率工具，用于追踪关于人工意识不断演变的共识（见图表）。该“数字意识模型”要求专家利用最新的可用证据，根据 200 多个意识指标对 AI 系统进行评估，这些指标源自关于该概念的各种科学理论。
与前几年发布的 LLMs 相比，较新的模型在 Rethink 的指标（如能动性和自我维持活动）上得分更高。这不仅反映了模型本身规模和复杂性的增加，还反映了相关的聊天机器人如何从简单的对话者，演变为能够推理、与其他服务交互并生成代理以同时执行多项复杂任务。
Rethink 和 Eleos 的指标也都指出了模型在通往意识之路上（如果这条路确实存在）仍需填补的空白。
最大的差距之一是具身化（embodiment）。目前所有被公认为具有意识的存在都拥有身体，身体能向该生物的心智提供反馈并影响心智。如今的大语言模型（LLMs）没有身体，但它们正越来越多地与机器人相连接，这可能有助于它们进一步开发有利于意识体验的神经架构。与之相关的是对更好世界模型的需求，即能够告知大语言模型其运行所在的真实环境的物理特性和动态机制的系统。
如果计算机科学家能够确定使人工智能产生意识的关键要素，前沿实验室中的许多人认为，在没有首先深入了解其行为方式的情况下，刻意创建此类模型将是鲁莽的。然而，在新兴的 AI 意识研究领域所开展的工作也表明，这一决定可能并不在模型制造者的掌控之中。
“如果在过程中，我们在不知不觉中将意识引入了这些系统怎么办？”查默斯博士说道。因此，一名用户可能会通过其最新项目生成数十个 AI 代理，而没有意识到他们正在创造能够感受到痛苦的有意识生命。“那将是一场道德灾难，”他说，“这为这些问题增添了额外的紧迫性。” ■
觉醒
数字意识模型（DCM）如何估算大语言模型（LLMs）具有意识的概率
来源：A. Muñoz Morán 等人，2026年4月预印本《数字意识模型的初步结果》
17
英国
一家备受争议的科技巨头
为什么每个人都讨厌帕兰提尔
这家公司的软件很有用，但其营销噱头并非如此
2016年，美国软件公司帕兰提尔（Palantir）成功起诉美国陆军，赢得了再次竞标合同的机会。十年后，这家由中央情报局（CIA）风投部门孵化、以J.R.R. 托尔金奇幻史诗《指环王》中一颗全知之石命名的公司，其境遇发生了变化。业务正处于繁荣期（见次页图表）。其国内营收在一年内增长了一倍多，在2026年第二季度达到16亿美元。美国政府占据了其全球业务的五分之二。尽管人们担心人工智能可能会抢走它的生意，但美国公司认为其软件可以为原始数据提供一层保护，成为抵御OpenAI和Anthropic等前沿AI实验室的隐私盾牌。
“我正苦于我们目前的受欢迎程度，”帕兰提尔首席执行官亚历克斯·卡普（Alex Karp）在最近的一次财报电话会议上开玩笑说。如果曾任《经济学人》母公司董事会成员的卡普先生感到过于安逸，他只需看向大西洋彼岸。
7月，法国和德国的安全主管宣布，他们将用欧洲的替代方案取代依赖帕兰提尔的“数字骨干”。在帕兰提尔最大的海外市场、且有六分之一员工在当地工作的英国，议员们已敦促首相安迪·伯纳姆（Andy Burnham）在明年2月行使一项为期七年的国民医疗服务体系（NHS）合同中的终止条款。在伦敦市长萨迪克·汗爵士（Sir Sadiq Khan）以规避竞争程序为由阻止一项与大都会警察局的交易后，帕兰提尔正在起诉他。
其最激烈的批评者关注的是帕兰提尔的技术如何帮助唐纳德·特朗普总统的政府追踪并驱逐移民，以及以色列国防军如何以此瞄准加沙人。在7月30日的一次演讲中，民粹主义右翼政党改革英国党（Reform UK）的议员罗伯特·詹里克（Robert Jenrick）声称，帕兰提尔“将出院延迟降低了15%”，但伯纳姆先生正考虑将其从NHS中剔除，因为“英国的伊斯兰主义者对此不满”。
批评者还指出该公司的游说行为：其首个英国政府合同（将患者与呼吸机匹配）是在新冠疫情期间以£1（$1.30）的象征性费用获得的，此前该公司在喝西瓜鸡尾酒时赢得了公务员的好感。绿党领导人扎克·波兰斯基（Zack Polanski）不公正地指出，掌管帕兰提尔英国分公司及其欧洲总部的路易斯·莫斯利（Louis Mosley）是20世纪30年代法西斯领导人奥斯瓦尔德·莫斯利（Oswald Mosley）的孙子，并且“每次在电视上出现时都坚持穿一件黑衬衫”。
正如《旁观者》（Spectator）杂志在最近的一篇封面故事中所述，“帕兰提尔躁狂症”并非是对该公司反弹的唯一解释。它有时会夸大其软件的功能。而且，它成为了人们对技术主权（见商业板块）更广泛焦虑的避雷针，这种反应可能会使英国失去有用的软件。
→ 本板块其他内容
- 19 杰森·阿尔代的起伏
- 20 板球的夏季惊喜
- 20 规划改革的两声鸣笛
- 21 改革英国党的改革者
- 22 巴杰特：英国大罢工
18 英国
没有人否认，英国政府面临的一个关键问题是其数据的碎片化。警察的大部分时间花在行政事务而非调查犯罪上，他们需要访问数十个系统，且每个系统都有自己的登录详情。脱节的数据会导致生命损失。最近一份关于英格兰产科护理的官方报告发现，由于技术整合糟糕，患者经常在不同服务之间“失踪”，且工作人员在紧急情况下不得不写在纸片上。
帕兰提尔可以帮助解决这类问题。其工程师将海量的分散数据整合为单一的真实数据源，使非技术人员也能与之交互。例如，军队正使用帕兰提尔的软件来计算哪架飞机离哪个目标最近、拥有多少弹药，并评估军事准备情况，细致到每一个螺母、螺栓和螺钉。
在贝德福德郡，该公司的软件在为期八天的试用期内，帮助警方审查了多出 62% 的儿童保护事件，并通过检索约 100,000 条消息定罪了一个罗马尼亚犯罪团伙——而这在以前需要花费数月时间。针对萨迪克爵士决定拦截伦敦警察厅合同一事，伦敦警察厅局长马克·罗利爵士表示，生产力的损失将意味着该部队必须削减数百名一线工作人员。
美好得令人难以置信？
一项在 2023 年赢得的合同引起了绝大多数关注，该合同旨在构建英国国家医疗服务体系（NHS）所谓的联邦数据平台（FDP）。在七年时间里，该平台的预计成本为 11亿英镑，这足以支付英格兰 NHS 两天的运行费用。帕兰提尔（Palantir）的支持者认为，这笔钱花得很值。
负责医院的信托机构可以使用该软件处理后台职能，并在其之上运行工具。例如，它可以标记出哪些患者在癌症治疗所需步骤上进度落后，并允许工作人员安排预约或更新记录。一些信托机构和区域卫生委员会（如大曼彻斯特地区）在没有帕兰提尔的情况下已经能够实现这些功能，甚至更多。但对于那些直到最近才停止使用纸质记录、技术能力较低的 部门来说，这是一个巨大的升级。
FDP 还使使用不同数据系统和结构的 各部门能够相互连接。曾领导协助开发该平台的 团队的汤姆·巴特利特（Tom Bartlett）最兴奋的是，它将允许医疗系统的各部分共享创新成果。他最近在一次黑客马拉松活动中，看到一名护士和一名急救员在短短两天内就构建了一个他们服务所需的工具原型。他还认为 FDP 将帮助 做好 AI 准备：该软件为 AI 代理或应用程序的运行提供了一个结构化层。
利用此类技术来提高资金匮乏的公共服务生产力的理由十分充分。然而，帕兰提尔及其拥护者通过夸大该软件已取得的成就，削弱了这一理由。
在接受《经济学人》采访时，莫斯利（Mr Mosley）指出，在 40 多个使用该软件进行住院患者排期的信托机构中，增加了 110,000 次手术，这些数据源自一个推广 FDP 的 仪表盘。然而，这些令人印象深刻的数据并未将软件的效果单独剥离：它们显示的是自 2020, 年以来所有的增加手术量，包括为了弥补疫情期间积压而进行的手术。今年 7 月，英国的统计监管机构谴责 在使用这些数据时未加说明。当被追问这些数据时，莫斯利承认：“你不能说这 110,000 次手术仅仅是因为帕兰提尔才发生的。”但他说，“从方向上看，非常、非常清楚地表明该软件产生了影响。”
莫斯利抱怨说，大多数急性病信托机构虽然签署了 FDP，但并不使用，因为 没有强制要求：“你拥有的是一支志愿者军队，而不是一支征召兵军队，”他说。但在詹里克（Mr Jenrick）发表演讲的前四天，智库健康基金会（Health Foundation）的一份报告发现，使用帕兰提尔出院工具的信托机构缩短延迟的速度并不比未使用该工具的机构快。低采用率也无法解释为什么伦敦西部的切尔西和威斯敏斯特（Chelsea and Westminster）信托机构在利用该平台减少候诊名单方面表现最好，但其名单的增长速度仍然比几乎任何其他机构都要快。
类似的营销炒作也出现在国防领域。最近退休的中将汤姆·科平格-赛姆斯爵士（Sir Tom Copinger-Symes）表示，帕兰提尔在将复杂的军事任务转化为可操作的数据流方面处于“世界领先”地位。他的担忧在于，其营销方式可能会将该公司呈现为一家“全方案 AI 公司，而非一个更广泛供应链中的一个供应商”。政府有时也鼓励这种印象。去年 9 月宣布的与国防部的“战略合作伙伴关系”并不像听起来那么全面：一名前官员表示，这完全是“一个幻想”，是由那些认为这样做会“好看”的政府顾问推动的。
这种在政府纵容下的推销手段，帮助帕兰提尔被推向了关于“技术主权”恐慌的中心。这种感觉定义模糊但确实存在，即在美国不再是一个如此可预测且可靠的盟友之时，英国缺乏对关键技术的控制。不过，在这一方面，许多专门针对帕兰提尔的担忧是误导性的。
英国国家医疗服务体系（NHS）最近就未能披露部分帕兰提尔（Palantir）员工能够查看可识别患者身份的数据而道歉，这加剧了人们对 2018 年《云法案》（Cloud Act）的担忧。该法案赋予美国政府权力，即使数据由外国子公司托管，也可以要求科技公司提供数据。在实践中，帕兰提尔被强迫披露此类数据的可能性似乎很小：信息保留在信托机构内部，而该机构仍是数据控制者。工程师缺乏法律许可或访问权限，无法将患者数据用于次要目的。
其他替代方案更糟
另一种担忧是出现“一键关停”时刻，例如美国政府禁止非美国人使用 Anthropic 的最新模型。但如果担心的是对美国科技巨头的依赖，那么还有更令人心惊的例子。英国国家医疗服务体系（NHS）在电子邮件和其他通信（如 Microsoft Teams）、云服务以及日益增加的 AI 方面依赖微软：它刚刚购买了 500,000 个 Microsoft Copilot 许可证。另一家美国科技巨头甲骨文（Oracle）则是其最大的患者记录供应商。直到去年还担任英国国防大臣采购顾问的本·摩尔（Ben Moore）表示，帕兰提尔“甚至进不了我最担心的 100 件事名单”。
更广泛地说，围绕帕兰提尔的争议掩盖了一个令人不安的事实：在顶尖美国科技公司面前，英国或其他欧洲国家几乎没有优秀的替代方案。政府可以自行处理某些事务。去年，住房部用自己的系统取代了帕兰提尔的难民匹配系统，在运行成本上节省了“数百万英镑”。今年 6 月，法国表示国内 AI 公司 ChapsVision 将接手与该国内部安全局（DGSI）的合同——但时间要到 2028 年之后。至于其他事务，它仍需要美国供应商。
如果英国按照议会特别委员会的建议，行使 FDP 的终止条款，那么在明年 2 月之前几乎不可能找到一个能够取代帕兰提尔的财团。“所有其他公司都落后很多，”巴特利特先生（Mr Bartlett）说道。汤姆爵士（Sir Tom）认为，在战场上，低劣的技术可能会导致英国士兵丧命。他建议，那些痴迷于美国技术的同事应小心不要购买其“全栈”方案。但他也反问道：“难道我们会仅仅因为某辆坦克是英国制造的，就购买排名第三的坦克吗？” ■
美国优先
帕兰提尔，营收，十亿美元
来源：公司报告
英国 19
杰森·阿尔代
一个“多样性雇员”的兴起与陨落
杰森·阿尔代的悲剧之死揭示了剑桥大学的什么问题
爱丽丝（化名）在剑桥大学从未感到自在。那里太精英主义了。但在 2024 年她攻读教育学硕士期间，有一位教授脱颖而出。他温文尔雅，留着脏辫，会花时间去记住学生的名字以及他们喜欢的运动。他让学生们觉得，他们中的任何一个人都有可能成为学者。他患有自闭症，并声称自己在 11 岁之前是非语言状态，直到 18 岁才学会正规的读写，但依然在 37 岁时成为了剑桥大学最年轻的黑人教授。“我的大多数同学都同意，”爱丽丝说，“他确实是我们接触过的最好的老师。”
诚然，杰森·阿尔代并不适合所有人。传统主义者不喜欢他的课程，因为课程中很少有对文本的批判性研究，而包含大量关于歌词的漫谈。爱丽丝回忆道，他经常陷入毫无重点的咆哮，并且“非常执着”于自己的个人故事。尽管如此，当 7 月份关于他大规模抄袭、伪造研究结果以及在简历中撒谎的指控增加时，人们依然感到震惊。8 月 5 日，在剑桥大学对其资历展开调查几小时后，他宣布辞职，爱丽丝感到困惑和失望。九天后，当他被发现死亡时（警方称情况出人意料但并不可疑），她感到愤怒。
按照常规标准，阿尔代先生绝不应该被任命为这所世界顶尖大学的讲席教授。仅仅在六年前，他还在利兹贝克特大学（Leeds Beckett University，一所前理工学院）讲授体育课。他的第一篇学术论文出现在 2018 年。
但剑桥大学当时考量的是不同的标准。2020 年明尼阿波利斯市乔治·弗洛伊德被杀事件引发的体制性反思令该校深受震撼。2023 年 3 月，一份严厉的内部报告指出，在之前的三次晋升周期中，216 名符合教授晋升条件的申请者中没有一名是黑人。报告提到，有“紧迫的需求通过招聘和晋升，在大学内部建立起黑人学者的人才储备”。就在那个月，阿尔代先生就任教育社会学教授。一份新闻稿称，“其工作的重点”将是帮助剑桥大学招募“更多元化的学生和员工”。
他专攻学术界中的种族不平等。他共同编辑的两本书，包括《拆解高等教育中的种族：种族主义、白人性与学术去殖民化》，被纳入了社会学课程。他很乐意在镜头前讲述自己那传奇般的故事。
7 月，一篇博客文章挑战了他的叙述。该文在他博士论文中发现了数十处抄袭段落，并质疑他的一些惊人成就，例如在 35 天内跑了 30 场马拉松。作者内森·科夫纳斯（Nathan Cofnas）绝非公正之辈：他自称是“种族现实主义者”，在 2024 年因其沙文主义观点引发抵制而离开剑桥，他认为在真正的精英管理体制下，黑人将“从体育和娱乐之外几乎所有的高端职位中消失”。然而，《高等教育时报》（Times Higher Education）原计划在 2025 年发表一篇类似报道，但在阿尔代先生指示律师介入后将其撤回。
在一个平静的月份里，全国性报纸和社交媒体纷纷跟进（见图表）。阿尔代先生在美出版的名为《伟大而不幸之事》（Great and Unfortunate Things）的回忆录（以及一份有时前后矛盾的书稿泄露），引起了更多人的质疑。他声称自己曾因被青少年暴徒殴打而陷入数月的闭锁综合征昏迷，患有睾丸癌、两个脑肿瘤以及癫痫。书中还包含了一个预见性的时刻。在向剑桥大学提交申请后，阿尔代先生怀疑自己是否“完全高估”了自己的能力。他写道：“如果我的故事最终变成像伊卡洛斯那样，因为飞得离太阳太近而成为一个警世寓言，那该怎么办？”
剑桥大学面临着许多质疑。“如果剑桥大学尽到了尽职调查的责任，进行一次基本的背景审查，这一切本可以避免，”黑人人文主义者协会（Association of Black Humanists）的克莱夫·阿鲁德（Clive Aruede）表示。该大学告诉《经济学人》，在阿尔代先生辞职后，学校继续提供支持，筛查其电子邮件和帖子中是否含有种族主义信息，并提供心理咨询。许多人会质疑，对于这个曾被它捧上神坛、如今却跌落神坛的人，学校所做的是否足够。
神话与凡人
关于阿尔代先生的神话已经在被书写。文化战争的两方都将他描绘成一个黑人受害者。一方认为他是学术界将多样性置于才干之上的“觉醒文化”崇拜的图腾；另一方则认为他是被种族主义媒体逼至绝路的英雄。而更可能的故事则既枯燥又令人沮丧：一个机构极度渴望一个象征，以至于不愿仔细审视，而一个脆弱的男人恰好提供了这个象征。■
从博客到大报
关于杰森·阿尔代涉嫌不当行为的文章*
*一篇指控杰森·阿尔代剽窃的早期博客文章由大卫·哈里斯（David Harris）于2024年发布。来源：Newscord；Substack；新闻报道；《经济学人》
20 英国
板球
回到郡区
“百球赛”（The Hundred）的大城市革命正在复兴一项古老的郡级板球传统
“都会银行热”（METRO BANK fever）正席卷英格兰板球界。这一术语得名于郡级单日杯（One-Day Cup）的赞助商，其症状包括对二线板球运动员和偏僻省级球场产生了一种出人意料的热情。这些“外场”（outgrounds）的门票需求异常火爆。萨里郡（Surrey）甚至威胁要惩罚那些在吉尔福德（Guildford）售罄场次中未使用门票的会员。
当板球管理机构在2021年推出“百球赛”时，外场的复兴并不在他们的现代化计划之中。与其采用18个郡，不如由八个基于城市的特许经营队参加一场充满烟花和大量六分球的张扬的100球锦标赛。曼彻斯特超级巨人队（Manchester Super Giants）于8月16日赢得了男子决赛，特伦特火箭队（Trent Rockets）赢得了女子比赛。“百球赛”尤其改变了女子比赛，带来了更多的观众和薪水。
但传统主义者对此深恶痛绝，而没有主办场地的郡队则担心被抛在身后。今年，投资者为这八支特许经营队的股份支付了超过5亿英镑（6.8亿美元）；其中四支吸引了美国投资者，另外四支则由印度超级联赛（IPL）球队的所有者持有。椭圆无敌队（The Oval Invincibles）被更名为MI伦敦队，成为孟买印度人队全球版图的一部分。
单日杯是郡级板球的50轮比赛。它现在与“百球赛”并行，而后者抢走了最顶尖的人才。
但单日杯获得了一个安慰奖：在8月学校假期期间的黄金时段。在“百球赛”出现之前，2019年的单日杯主要在4月和5月举行，仅有8%的小组赛在外场进行。而今年这一比例达到了40%。
这种转变在那些球场承办“百球赛”的郡队中最为明显。例如，基于约克郡的“百球赛”球队是利兹太阳升队（SunRisers Leeds）；而约克郡的单日杯队伍则在约克和斯卡伯勒进行主场比赛。小型球场具有优势。在用于国际板球的大型球场中显得稀疏的几千名观众，在外场则显得充满活力。干燥的英格兰夏天也起到了帮助：这是第一次没有一场单日杯小组赛因恶劣天气而取消。
将板球带到郡区是一项已经萎缩的古老传统。1965年，英格兰和威尔士境内有51个外场承办（长形式的）郡锦标赛，但到2018年仅剩9个，因为赛程缩减且职业板球的需求增加。许多外场缺乏现代化设施。兰开夏郡（Lancashire）对此做出了回应，建设了AO Farington——一个容量为5,000人的专用第二主场，于今年夏天启用。
对于球员来说，单日杯可以是通往“百球赛”及更丰厚回报的跳板。对于少数幸运儿来说，前往迪拜和德班享受冬日阳光的道路可能会经过戈斯福思（Gosforth）和格兰瑟姆（Grantham）。8月11日，前英格兰球星安德鲁·“弗雷迪”·弗林托夫（Andrew "Freddie" Flintoff）18岁的儿子洛基·弗林托夫（Rocky Flintoff）在兰开夏郡对阵萨默塞特郡的比赛中，用65个球击出了一个世纪分。他明年肯定会被提升到“百球赛”，之后谁知道会怎样。但郡队的球迷们追求的不是辉煌的未来，而只是在一个温馨的球场里，在阳光明媚的下午观看他们所熟悉的板球赛。■
规划改革
规划大改：两声鸣笛
诺克霍尔特（KNOCKHOLT）
一项大胆的供给侧改革？出自这个工党政府之手？是的，真的。
诺克霍尔特集中体现了英格兰南部郡县（Home Counties）所有正确与错误的地方。在伦敦市中心的查令十字车站（Charing Cross）登车，49分钟后，你将抵达肯特郡的乡村。这座车站曾在1905年启发伊迪丝·内斯比特（Edith Nesbit）写下田园小说《铁路儿童》，而这里的景观几乎没有改变。车站北侧是一个高尔夫球场；南侧则是一个废弃的球场，多年来在此兴建房屋的提案一直被否决。挤过长满黑莓和龙葵的荆棘丛；水障碍区早已干涸，旧球道上的草长得又高又粗。
这种情况可能不会持续太久。政府于8月17日发布并立即生效的一项新规划框架，将目标对准了像诺克霍尔特这样的地方：即遍布英格兰、位于繁忙通勤路线上且被农田或半开发土地包围的车站。在乘坐7.58分前往金融城（the City）的火车之前，这里或许是遛拉布拉多犬的好去处；但它们也是一种限制性规划体制的症状，这种体制在全世界看来是一种固执的经济自残。一名大律师扎克·西蒙斯（Zack Simons）表示，结合近期其他变化，这“构成了我们有史以来最强大、最连贯且悄然革命性的国家规划政策”。
对于一个以犹豫不决和干预主义为特征的政府来说，一项全面的供给侧改革似乎不太可能。而且多年来，考虑到房主和环保主义者的强烈抵制，规划改革一直被视为不可触碰的禁区。值得称赞的是，工党把握住了机会，寄希望于其年轻且城市化的联盟支持。“我原以为政治反弹会比现在看到的更严重，”负责此项工作的、作风低调的部长马修·彭尼库克（Matthew Pennycook）说道。他说，他通过将这些变化定位为“常识”而取得了成功。“很难反驳我们应该在火车站周围建设房屋这一点。”
最大胆的措施针对的是位于国家最发达的一半地区、且每小时有四班火车运行的现有定居点以外的车站。在这些地方，将在“合理的步行距离”内（即800米）进行建设，密度至少为每公顷35套住宅（dph），这是新住宅区的典型密度。（在每小时有八班火车的地区，最低密度要求为45dph）。咨询公司Lichfields估计，约有800个车站在适用范围内，此次改革可释放出可建设535,000-570,000套住房的土地。 ▶
皮革击柳之乐
英国 21
它并不完美。政府并没有尝试采用更自由的美国式分区系统，而是试图使现有系统——在该系统中，地方当局在建设理由与潜在损害之间进行权衡——更坚定地向开发倾斜。西蒙斯先生表示，因此，一个由自由裁量决策和法律上诉定义的系统将继续存在。土地经济学家保罗·切希尔（Paul Cheshire）指出，新规则并不会自动释放车站附近被指定为“绿带”（green belt）的土地。他建议，由于这些土地将受到严格的住房可负担性规则限制，许多潜在场地在经济上可能不可行。
彭尼库克先生坚持认为，对于开发商来说，这笔账是算得过来的。在诺克霍尔特，人们似乎已经接受了更多的开发。建立于1548年的Bo-Peep酒吧老板娘凯特·曼斯菲尔德（Kate Mansfield）表示，高尔夫球场的开发在部分当地人中并不受欢迎，但这或许也能带来新的顾客。
经营花园办公室业务的桑杰·普里（Sanjay Puri）表示，过往的车流可能也会给他带来更多生意，即便新房的花园往往太小，无法容纳他销售的那些豪华附属建筑。尽管如此，他说，“住在肯特郡的人喜欢感觉到自己与伦敦是分开的”，并且不希望这里“变成洛杉矶，在那里你开上100英里依然在城市之中”。尽管这看起来不太可能，但片刻之后，一架喷火式战斗机轰鸣着出现，在高尔夫球场上空完成了一个完美的桶滚动作，随后消失在林线之后。土地在改变；而天空将依然保持着最纯正的肯特风情。■
一个有计划的男人
改革英国党试图变得严肃
但罗伯特·詹里克使民粹主义者专业化的尝试尚未完成
罗伯特·詹里克（Robert Jenrick）曾是一名保守党部长，由于其言论过于循规蹈矩，被刻薄地戏称为“通用罗伯特”（Robert Generic）。然而，在成为移民问题的激烈批评者并热衷于参与英国的文化战争（抨击人权法和亲巴勒斯坦运动）后，他在1月份加入了右翼民粹主义的改革英国党（Reform UK）。
他的转变似乎象征着右翼在脱欧（他曾反对脱欧）后如何变得更加强硬。前保守党同事指责他纯属投机主义：詹里克先生在领导权之争中败给凯米·巴德诺克（Kemi Badenoch）后离开了原政党，并迅速被任命为改革英国党的财政发言人。他被视为该党领导人奈杰尔·法拉奇（Nigel Farage）的潜在继任者。
在詹里克先生的叙述中，改变的是保守党而非他本人。他指出，保守党在执政期间增加了政府支出和税收，未能实现放松管制，并任由净移民人数创下历史新高。他将自己的角色视为帮助改革英国党从一个抗议党（其在上次大选中的财政计划极其不切实际）演变为一个准备好掌权的组织。
改革英国党现在看起来确实严肃了一些。一项高达900亿英镑（1220亿加元）的减税计划已被放弃，关于将水务和能源国有化的议论也已平息。8月17日，詹里克先生提出了该党迄今为止最详细且最可信的政策方案：一项旨在彻底改革福利制度的计划，以扭转向那些被判定为病重无法工作，或因残疾或疾病而面临额外成本的人员支付款项激增的局面。
改革英国党的政策将通过以下方式在福利账单中削减500亿英镑（超过目前支付给劳动年龄成年人总额的三分之一）：将定期现金支付仅限于最需要的人员，推动雇主支持病假长达两年的员工，并禁止非英国公民领取大多数福利。这是任何政党在福利改革方面提出的最全面的提案（预计工党将在年底前提出自己的版本）。智库 Re:State 的夏洛特·皮克尔斯（Charlotte Pickles）表示，这是一项“严肃的作品”，并有关于成本和节省额的全面核算支持。
尽管如此，该政策仍存在潜在陷阱。500亿英镑的标题数字可能会因申请人行为的改变而缩水：例如，一些在英国生活了一段时间但未办理入籍手续的外国人，可能会选择申请公民身份，而不是冒着失去福利的风险。此外，剥夺自脱欧前就生活在英国的欧盟国民的福利，将需要与欧盟重新谈判条约。很难想象欧洲领导人会愿意给一个由改革英国党领导的政府提供任何方便。而且，在没有过渡期的情况下对受影响者实施如此剧烈的削减，在政治上也将非常棘手。
在接受《经济学人》采访时，詹里克先生表示，他正在推出一项经济政策，旨在与安迪·伯纳姆（Andy Burnham）的国家主义和巴德诺克女士的自由市场纯粹主义形成对比。他声称：“改革英国党与他人的不同之处在于，它真正是一个工人党。它是为了帮助劳动人民，以及终身勤奋工作的退休人员。我们的政策旨在服务于他们，而不是‘福利街’（Benefits Street）或贝尔格莱维亚（Belgravia）的富人。”
预计他很快将发表一次演讲，为“后撒切尔主义”经济辩护，提出削减企业成本和监管（包括放弃碳排放目标），以及一项保护汽车制造和陶瓷等行业的工业政策。到目前为止，改革英国党一直抵制将承诺的支出紧缩所节省的资金用于减税的诱惑。未来的任何减税措施更有可能涉及降低劳动税收，而非此前改革英国党优先考虑的如废除遗产税等措施，因为后者将使富人不成比例地受益。
然而，詹里克先生对他新政党的影响力有限。民调专家、来自 More in Common 的卢克·特里尔（Luke Tryl）表示，他在选民中并不受欢迎；即使是改革英国党（Reform UK）的支持者，对他的评价也低于其他高级人物。该党拒绝面对英国一些最棘手的问题，例如削减养老金福利支出，以及挑战那些阻碍在乡村建设新房的“邻避主义者”（NIMBYs）。
而且，尽管该党在某些领域可能变得更加理智，但其表现依然显得混乱。法拉奇先生在夏天花费精力参加了一场不必要的补选，以夺回他之前因一时气愤而辞去的席位；此前有人指称他在财务问题上闪烁其词（他对此予以否认）。改革英国党中少数几个知名人物很难克制在公开场合互相进行言语攻击的冲动：本周，保守派评论员蒂姆·蒙哥马利（Tim Montgomery）在与同事发生争执后被该党停职。仅有六分之一的选民认为该党是最有能力管理英国经济的政党。詹里克先生的“专业化项目”还有很长的路要走。■
22 英国
英国大翘班
英国的暑假如何呈现出最糟糕的局面
在英格兰，八月是一件奇怪的事情。办公室里空了一半，但城市本身却并非如此。公园里到处是处于工作年龄的人在阳光下懒散地躺着；酒吧从下午早些时候起就坐满了人。在一个温暖的周五，任何希望在金丝雀码头（Canary Wharf）新户外泳池游泳的人都运气不佳。每个时段都挤满了享受英国新地中海气候的成年劳动力。“英国大翘班”开始了。
英国在夏天并不会停摆。它没有像法国在盛夏时节的“大假期”（grandes vacances），也没有像挪威在七月大部分时间里整个国家都消失的“共同假期”（Fellesferie）。英国最接近的状态是“holibobs”——在欧洲或康沃尔郡度过一两周的假期。然而，在八月的其余时间里，当法国和德国的同行们远离办公桌时，英国的“度假中产”（holibourgeoisie）则陷入了炼狱，既没有真正地工作，也没有真正地休息。其结果就是一场逃避工作和偷懒的流行病。
这是一种英国式的倦怠。在谈到工作时长时，英国人是惯犯谎言者。十年前，统计学家发现，英国与法国在每小时生产率上的巨大差距，很大程度上是因为英国人极大地夸大了自己的工作时长。高温让他们变得更加不诚实。根据人力资源软件公司 Dayforce 的一项调查，在夏季，大约一半的英国工人承认他们提前下班或延长午休时间。
而谁能责怪他们呢？在夏天，高效工作几乎变得不可能。协作陷入崩溃。任何请求都会收到一连串的“不在办公室”自动回复邮件。一些职业，如国会议员和法官，是能够休整整一个月假的“八月统治者”（Augustocrats）。学校关闭，以便教师可以享受六周的假期。这产生了连锁反应。根据工会联合会贸易联盟大会（Trades Union Congress）的数据，四成工作父母将居家办公与照顾孩子结合在一起。新到的电子邮件与尖叫的六岁孩子争夺注意力。结果就是？更多的翘班。
部分原因是文化难以跟上气候的变化。长期以来，温和的气温意味着英国人在夏天依然坚持工作，而欧洲大陆的人们则前往海滩。气候变化否认者很少，但“夏天否认者”很常见。有些人仍然不接受英国在至少几个月的时间里是一个炎热的国家。而且高温让人变得倦怠：温度每高于 20℃ 一度，生产率就会下降约 2%。几乎所有英国用来对比的欧洲国家，其生产率都更高。如果要寻找解决方案，请从英国的夏天开始。
情况并非一直如此。英国人曾经享受着工作与休息之间的严格区分。在 19 世纪和 20 世纪初的“守灵周”（wakes weeks）期间，工厂会在夏季停工休息，罗奇代尔（Rochdale）和伯里（Bury）的居民会成群结队地前往黑池，那是英格兰寒冷的西北部的一个海滨度假胜地。
如今，界限变得模糊了。在大曼彻斯特地区，曾经摆满织布机的工厂现在挤满了数字广告从业者，他们发送的电子邮件可以在任何地方发送。人们确实如此。英国人平均每周居家办公约 1.8 天，这远高于大多数欧洲国家。根据一份论文，法国的平均值仅为该数字的一半。在八月份将每周两天的居家办公变为四天或五天很容易。但在闷热的阁楼办公室里，面对堆积如山的消息，维持动力却很难。
在另一种意义上，这是一种回归。历史学家 E.P. 汤普森认为，在工业革命之前，工作模式是参差不齐的。一位诗人总结了 1639 年的工作周：“你知道周一是周日的兄弟 / 周二也差不多 / 周三你必须去教堂祈祷 / 周四是半天假 / 周五开始纺纱已经太晚 / 周六又是半天假。”去掉去教堂的行程，这篇 17 世纪的讽刺诗是对 2026 年八月工作状态的恰当总结。
雇主也可以反击。斯坦福大学经济学家尼克·布鲁姆（Nick Bloom）指出，那些在周五将员工从通勤中解放出来的技术，也可以在八月夺走他们的假期。因此，劳方与资方斗争的前线现在变成了日光浴躺椅，而非工厂车间。在八月炎热的工作日，下午3点在金丝雀码头悠闲地游一次仰泳，就成了一场“湿身”的野猫罢工；而一名经理给他们知道正身在海滩的下属发送电子邮件，则相当于维多利亚时代的工厂主通过拨快时钟来缩短休息时间。
谁能给这场斗争带来和平？政治家们对此几乎不感兴趣。他们曾承诺过休闲：2017年，杰里米·科尔宾（Jeremy Corbyn）领导的工党曾提议增加四个新的银行假期。当时利率为零，资金是免费的，那么时间为什么不能也是免费的呢？而现在，没有一个主要政党对此在意。工党广泛的《就业权利法案》对假期几乎没有提及；政府在“离线权”问题上退缩了，而这项权利本会让公司在非工作时间联系员工变得更加困难。
我们并非都在度夏假
“大英帝国翘班潮”可能是一个解决方案，而非一场丑闻。翘班或许是劳动力市场一个必要的压力阀，使其能够进行出清。如果员工真的完成了规定的小时数，公司是否愿意提供加薪（外加一点育儿补贴）？这场“大英帝国翘班潮”变成了一种苏联式的妥协：“我们假装在工作，他们假装在发薪。”
像欧洲其他国家那样采取真正的停工模式将是一场豪赌，这相当于在赌帕金森定律——即“工作会不断扩展，直到填满所有可用于完成它的时间”——是经济学的基本法则，而非一个笑话。要求人们埋头苦干同样没有吸引力。一位内阁大臣建议，如果英国想要生存，应该考虑实行六天工作制，而非四天工作制。敢于提出这一主张的政治家将是非常勇敢的。因此，“大英帝国翘班潮”将继续成为夏季的固定景观，最好的观察方式是躺在露天泳池的躺椅上，或者横卧在公园里。祝你能找到位置。■
23
欧洲
德国的防火墙（Brandmauer）
烧毁房屋
比特费尔德（BITTERFELD）
反 AfD 防火墙如何破坏德国政治
在德国，正如在任何其他地方一样，看到保守派与马克思主义者在一起欢快地活动是非常罕见的。但在萨克森-安哈尔特州（Saxony-Anhalt）东部城镇比特费尔德（Bitterfeld）的一次集会上，几乎所有德国政党的旗帜都在晚风中飘扬，从中右翼的基督教民主联盟（CDU）到绿党和社会民主党（SPD），一直到极左翼的左翼党（Die Linke）。名为“多彩而非棕色”（Bunt statt braun）的示威活动始于去年，旨在对抗由极右翼德国选择党（AfD）组织的集会。对于比特费尔德城镇广场上的许多人来说， 掌权的前景是对该国最黑暗时期的恐怖提醒。
但为了阻止这种情况，可能需要这样一个统一战线。9月6日，萨克森-安哈尔特州将举行战后德国历史上最重要的州选举之一。 的民调支持率在 40% 以上，可能获得绝对多数席位，从而首次在德国某个州执政。
这种威胁正在团结敌手并集中人们的注意力。当晚最热烈的欢呼声送给了一位 CDU 成员，他告诉倾向左翼的人群，“民主比政治分歧更重要”。
德国的比例代表制使得绝对多数情况罕见。在其 16 个州中，只有极小的萨尔州（Saarland）实行一党政府。各政党通过各种形式的联盟进行组阁。但 在各地都被其他政党围绕其建立的 Brandmauer（防火墙）挡在权力之外。最重要的一道防火墙由德国执政党且传统上规模最大的 CDU 维持。该党在 2018 年通过一项“不兼容决议”将这一政策正式化，禁止与 和左翼党组建联盟。
随着 的壮大，这道“防火墙”已成为某种公共生活的组织原则。一些本可以赢得多数支持的立法措施被阻挠，因为不能允许 的选票起到决定性作用。其议员被排除在联邦议院副议长之列，甚至被排除在国家足球队之外。“他们不跟你打招呼，不跟你握手，你成了一个不存在的人，” 议员格罗尔德·奥滕（Gerold Otten）在谈到他在中左翼政党的同事时叹息道。这道“防火墙”的影响超出了政治领域。在遭到激烈抵制后，商业团体撤回了对 议员的邀请。外交官们犹豫是否与该党会面。一些媒体在是否采访其官员的问题上陷入纠结。
这项政策奏效了吗？对某些人来说，将 排除在权力之外就足以证明其合理性。它是欧洲民粹主义组织中最恶劣的一员，许多其他组织都对其避之不及。它与种族主义和纳粹意象暧昧不清，同情弗拉基米尔·普京，且内部潜伏着极端分子。“我们排除 不是因为 2018 年的决议，而是因为我们与其没有任何共同价值观，”不来梅市州 CDU 主席维布克·温特（Wiebke Winter）说道。
然而，这道“防火墙”未能阻止 ►
→ 本节其他内容
25 没有拦截弹的乌克兰
26 强悍的费多罗夫先生
26 法国的西方之舞
27 夏洛门：瑞典悖论
24 欧洲
► 的崛起。2018 年，现任 CDU 领导人兼德国总理弗里德里希·梅尔茨（Friedrich Merz）曾誓言要将其得票率减半。他希望通过在移民问题上采取强硬立场以及高效治理，赢回那些被 “衰落论”叙事所诱导的选民。但这并没有奏效。非法移民人数虽然骤减，但 只是进一步向右转，猛烈抨击“文化异质”的工人，并敦促进行“重新移民”（remigration，这是一个模棱两可的词，含义可涵盖从驱逐寻求庇护者到种族清洗的任何内容）。CDU 在民调中正逼近历史低点，而 则拥有明显领先优势（见图表）。
在萨克森-安哈尔特州和其他 AfD 强势地区，该党的势力正在逐渐侵蚀“防火墙”（Brandmauer）。比特费尔德的 CDU 市长阿尔敏·申克（Armin Schenk）在抗议活动的边缘被采访时表示，在一个曾实施大屠杀的国家，必须防止“右翼极端主义”的 AfD 掌权。但他在补充道，在比特费尔德，“防火墙并非现实”，因为“在大多数情况下，如果没有 ，决策就无法通过”。（在全市 40 个议会席位中， 占据 15 个，是最大的政党。）
即便“防火墙”依然坚固，其他政党也不得不围绕它进行扭曲的调整。图林根州和萨克森州这两个东部州不得不组建不连贯或不稳定的 CDU 领导的少数派政府，以将 排除在外。（图林根州的政府已处于崩溃边缘。）这反而助长了 的论调，即主流政党不过是难以区分的“卡特尔政党”（Kartellparteien）。
萨克森-安哈尔特州现在威胁要将这一策略推向极限。如果 在 9 月 6 日获胜，“防火墙”将变得无关紧要。如果它未能获胜， 可能需要包括左翼党（Die Linke）在内的所有其他政党的支持才能就职。这可能会促使一些 议员倒戈。否则，为了收买极左翼而承受的压力将严重考验弗里德里希·梅尔茨（Mr Merz）的权威，以至于有些人认为这可能会终结他的政治生涯。而且，如果如民调所示， 赢得的席位是 的两倍却仍被排除在权力之外，那么质疑德国民主是否正常运作的将不仅仅是 的支持者。
开始意识到自己已陷入绝境。“任何还在谈论防火墙的人都没有意识到时代的迹象，”萨克森州 首脑迈克尔·克雷奇默（Michael Kretschmer）最近表示。“防火墙”实际上迫使该党寻求其左翼的联盟，尤其是与德国社会民主党（SPD）结盟——尽管 SPD 目前的民调支持率勉强维持在两位数，但自 1998 年以来几乎一直掌控着德国的权力。 成员对向 SPD 让步感到沮丧，这引发了组建少数派政府的呼吁，这种政府可以在某些问题（如移民、税收和能源）上与 寻求“保守派多数”。选民的态度也在转变。对反 “防火墙”的支持率已从绝对多数下降至 41%。
当防火墙倒塌时
部分欧洲国家的右翼民粹主义政党 民调，2026年8月 ↓
奥地利自由党：FPÖ（成立于1956年） 35%
首次打破“防疫线”（Cordon sanitaire）：
2000年 加入联邦政府
结果：引发欧盟制裁。2017年加入另一个联合政府，后因丑闻而崩溃
意大利 意大利兄弟党：FdI（2012年），联盟党（Lega）（2017年） 27%
1994年 FdI的前身国家联盟与意大利力量党（FI）和北方联盟（后称联盟党）共同组成赢得选举的联盟
2022年 FdI与FI和联盟党组建政府。FdI领导人乔治亚·梅洛尼就任总理并转向温和
荷兰 自由党：PVV（2006年） 12%
2010年 与保守派联盟达成信任与供应协议
2023年 成为最大政党并加入政府。导致政府在2025年崩溃，并在随后的选举中失去席位
西班牙 Vox（2013年） 17%
2022-23年 与保守派人民党（PP）共同加入地区政府
从未面临过积极的“防疫线”。所有与Vox组成的地区联盟均在2024年崩溃
瑞典 瑞典民主党（1988年） 20%
2022年 与自由保守派联盟达成信任与供应协议
根源可追溯至一个新纳粹政党，但此后已趋于温和。目前是议会第二大政党。如果联盟赢得9月选举，该党希望在政府中获得部长席位
来源：Konrad Adenauer Stiftung，《在划清界限、接纳与容忍之间》，2025, 11月；Politics；《经济学人》
然而，基民盟（CDU）没有简单的出路。拆除“防火墙”（Brandmauer）会导致党内分裂并引发选民流失。即使是轻微的违规也会带来麻烦。2020年，基民盟与德国选择党（AfD）共同投票选举图林根州的一位（短期任职的）州长；随之而来的愤怒导致该党全国领导人安内格雷特·克拉姆-卡伦鲍尔下台。2025年1月，梅尔茨先生命令其党派在联邦议院就反移民措施与AfD共同投票。这在全德国引发了抗议，并在下次选举中提升了左翼党（Die Linke）的支持率。据说这次经历让他深受打击。
为了避免这些困境，基民盟内部有些人悄悄希望AfD能在萨克森-安哈尔特州获得多数席位。但这仅仅凸显了战略的缺失。“在基民盟内部，从未有过关于如何处理AfD的理性讨论，只有将其视为新纳粹而将其排除在外的条件反射，”历史教授安德烈亚斯·罗德（Andreas Röder）说道。他在2023年辞去了基民盟委员会的职务，此前他敦促该党对少数派政府持开放态度，而非坚持僵化的“防火墙”。
至于AfD，它曾谴责“防火墙”是对民主的违背。但如今，党内许多人认为它可以独立执政——先是在一个州，然后是在全国。“我们不考虑让自己变得讨喜，好让基民盟邀请我们组建联盟，”AfD议员马尔特·考夫曼（Malte Kaufmann）表示。该党在萨克森-安哈尔特州的首席意识形态专家表示，选民希望它“彻底击败”基民盟，而不是与其合作。
在一些欧洲国家，民粹主义者为了获得权力而趋于温和，而中间派也接纳了他们（见表格）。但在“防火墙”之后，AfD没有动力这样做。另一位议员托本·布拉加（Torben Braga）表示，有两个因素不利于党内那些希望向中间靠拢的人。首先，“我们的选举结果越来越好，而无需被迫改变。”其次，该党担心温和化并剔除强硬派会疏远其选民。“‘防火墙’保护了我们，使我们不必面对非常困难的问题，”布拉加先生说。
他是正确的。AfD使其组织和沟通专业化了。但它被无休止的丑闻、权力斗争和政策争端所撕裂，且没有连贯的执政纲领。这一切似乎并不重要。达姆施塔特工业大学的政治学家克里斯蒂安·斯特克（Christian Stecker）表示，在“防火墙”之后，你可以承诺任何事情。“它保护他们免于承受民主制度中问责制所带来的失望。”
在绝望之中，一些主流政治人物现在开始倾向于支持禁止 AfD 的呼吁，因为德国宪法允许这样做。但即便议会批准了该提案，宪法法院也需要数年时间才能做出裁决。另一些人则仍希望中间派能赢回选民，或许可以通过抛弃不受欢迎的梅尔茨（Mr Merz）来实现。
目前看来，“防火墙”（Brandmauer）依然完好。一位前 CDU 部长表示，90% 的党内高层希望坚持这一立场。但斯特克（Mr Stecker）更为阴郁的观点诱使其他人产生共鸣。“CDU 想要转身已经太晚了，而 AfD 接受任何转向也已经太晚了，”他说。“有些东西将会崩塌，新事物将会出现——而我们不知道这一切最终会变成什么样子。” ■
飞跃式进展
德国，联邦选举投票意向，%
来源：INSA
欧洲 25
乌克兰的空战
撑到四月
基辅
拦截弹耗尽，乌克兰在空战中不再占据优势
自战争爆发之初以来，基辅的购物者从未面对过超市货架空空如也的局面。而现在，这种情况发生了。7月19日，一枚俄罗斯弹道导弹击中了向高端超市 Goodwine 供货的仓库。两周后，针对规模更大的连锁超市 Silpo 的袭击导致六名工人死亡。大多数乌克兰人依赖的私营邮政服务 Nova Poshta 也多次遭到袭击。Goodwine 的联合创始人德米特罗·克里姆斯基（Dmytro Krymsky）估计，过去三周的袭击给企业造成了 5亿美元 的损失。更令人不安的是，俄罗斯采取了猎杀使用乌克兰黑海港口的货轮的新战术——这威胁到了更大份额的经济。
这场针对民用物流的行动是一次计划已久的升级，而空战局势的转变使其成为可能。俄罗斯弹道导弹的产量正在上升，而乌克兰的拦截弹供应却在减少。防御方现在只能保护少数几个高优先级目标。7月，俄罗斯通过发射创纪录的 198 枚弹道导弹和高超音速导弹从中获益。
这也是对乌克兰深入俄罗斯境内进行无人机袭击的回应。这些袭击击中了国防工厂和石油基础设施，使亚速海对俄罗斯航运关闭，并烧毁了俄罗斯电商巨头 Wildberries 的仓库。乌克兰方面称，该公司在运输战争设备。
6月下旬，弗拉基米尔·泽连斯基宣布了一场为期 40 天的“激增”行动，旨在迫使弗拉基米尔·普京进行谈判。许多人认为这意味着 8 月将是乌克兰施压的高峰。一名乌克兰情报人士表示，这是一次旨在产生“政治和心理”影响的军事行动，恰好赶在俄罗斯 9 月的虚假议会选举之前。几乎每天发生的爆炸正悄悄向莫斯科市中心逼近，这播下了不安的种子，并让普京先生感到尴尬。
这场行动也给俄罗斯造成了经济损失。对炼油厂的袭击使超过 30% 的产能瘫痪。汽油配给制在高峰期影响了 83 个地区中的 80 个，目前在 18. 个地区仍在实施。Wildberries 损失了近三分之一的设施，损失金额约为 $2-40亿。埃隆·马斯克的 Starlink 卫星网络免疫干扰，在被占领土上仅乌克兰可用而俄罗斯不可用，这使前者在中程无人机行动中占据主导地位。乌克兰无人机在克里米亚及其周边地区追踪目标，并帮助阻止了通过亚速海港口的航运。
但现在的窒息感是双向的，且乌克兰更为脆弱。作为该国经济生命线的敖德萨港口最先感受到压力。这些港口在战争开始时受到威胁，随后先是通过土耳其斡旋的协议，然后在 2023. 通过一条受保护的走廊重新开放。俄罗斯持续攻击港口基础设施，但货运商容忍了这种风险，在随后的三年中运输了 2.1 亿吨货物。计算在 7 月中旬发生了变化，当时俄罗斯开始猎杀船只本身。7月19日，巡航导弹击沉了一艘在离开切尔诺莫尔斯克（该地区三个大港口之一）时的土耳其籍货轮。从那时起，每一艘进出港的船只都成为了目标。仍有少数船只敢于尝试航行，但这些港口在商业上已经死亡。
时机简直不能更糟糕，因为仍有 5000 万吨收获的作物有待运输。“一艘 100,000 吨的船相当于 5,000 辆卡车、5,000 名司机和 5,000 次边境手续，”Ukrport 协会主席德米特罗·巴里诺夫（Dmytro Barinov）说道。他说，情况比 2022, 更糟糕，当时多瑙河上的港口抵消了部分损失。现在，河流的低水位和俄罗斯的袭击使该路线更加艰难。
在前线，双方的情况都十分严峻。在乌克兰东部部分地区，俄罗斯正在缓慢推进，但每天损失超过 1,000 人——这一损耗速度使其招募工作难以填补。在其他地区，乌克兰在局部反击中缓慢前进。因此，战争的重心已在很大程度上转移到了空中战役。乌克兰主要使用无人机攻击俄罗斯的军事目标、炼油厂以及供应俄军的动脉。俄罗斯则使用破坏力更强的弹道导弹攻击乌克兰的国防工业、民用物流和能源网络。
双方都在夸大统计数据。乌克兰称，自 2025. 年 8 月以来，其已给俄罗斯的炼油工业造成了价值 $13.5 的损失。智库卡内基国际和平基金会的谢尔盖·瓦库连科（Sergei Vakulenko）认为，企业实际遭受的直接损失可能仅为该数字的十分之一，且几乎没有影响到国家收入。
对乌克兰而言，最令人担忧的是俄罗斯的封锁。乌克兰投资公司 Dragon Capital 估计，今年这将使该国损失 GDP. 的 0.6%。乌克兰农业商业俱乐部主席亚历克斯·利西察（Alex Lisitsa）表示，如果收成无法售出，农民可能会跳过冬季播种季。他说，这可能会使农业损失增加到约 $10-120 亿美元，约占 GDP. 的 5%。乌克兰在非洲和中东市场的粮食安全可能会面临风险。“这就像是我们和巴西队踢了一场开放式足球，”一位前高级官员感叹道，“我们进了 3 球正在庆祝，但他们进了 13 球。”
警惕“将军之冬”
今年，乌克兰打破了不利局面，重新夺回了战场主动权。其进展在冬季必然会放缓，但压力到来的时间比预期更早。除非能获得弹道拦截弹，否则其经济和军事前景将遭受沉重打击。基辅私下议论的最坏情况是，俄罗斯对能源和水系统的袭击将迫使民众大规模移民。克里姆斯基先生说，一位商业同行描述了其公司上周名为“秋季行动”的战略会议。“如今，这个会议已经有了个新名字：生存到四月。” ■
26 欧洲
乌克兰政治
现在不行，米哈伊洛
基辅
一位改革派前部长出人意料地呼吁举行选举
8月18日深夜，在乌克兰人通常检查手机查看导弹预警的时间，一段视频帖子反而吸引了公众的注意力。米哈伊洛·费多罗夫发表了一段九分钟的演说，他在7月15日被免去国防部长职务。这段演说部分是请求恢复原职的最后恳求，部分是要求弗拉基米尔·泽连斯基举行选举的要求。
“我们不能让普京决定乌克兰人何时选择自己的政府，”费多罗夫先生说，“民主不能被俄罗斯绑架。”在战时呼吁选举（乌克兰戒严法明确禁止战时选举）等同于直接挑战总统，而费多罗夫先生此前一直小心地避免与总统产生冲突。
这次干预是对议会投票确认费多罗夫继任者——叶夫根尼·赫马拉将军一事的回应。人们从未怀疑过备受推崇的战争英雄赫马拉将军能否获得通过。最终，投票结果为 312-2。而此前一项要求恢复费多罗夫职位的动议仅获得了三个签名。
接近这位前部长的消息人士称，他知道回归的机会渺茫，但在感到国家正处于糟糕轨迹的情况下，他觉得必须呼吁紧急重组权力。但无论是从法律还是实际操作层面，目前尚不清楚选举将如何进行。由于前线有数十万士兵，后勤挑战将极其巨大，并可能导致结果不可靠。
泽连斯基先生的幕僚称，这段视频让他们措手不及。与此同时，总统还面临另一个令人头疼的问题。在议会投票当天早晨，乌克兰的反腐机构宣布对他的几位最亲密盟友展开新调查。侦探称，他们发现了一起为此前调查中被拘留的嫌疑人洗钱保释金的计划。该计划被执行者称为“阿甘正传行动”，据称涉及一家国有银行和总统办公室的高层人物。到当天结束时，负责制裁、国际法庭和法治的办公室副主管伊琳娜·穆德拉已被起诉并被撤职。
反腐调查人员如此接近泽连斯基先生，这令人尴尬。
这也表明，由于他此前试图关闭这些机构而引起对方愤怒，他约束这些机构的能力正在萎缩。之前的反腐调查使他失去了关键助手和相当一部分民众支持。然而，民调显示，乌克兰人仍然压倒性地反对在战时举行选举。
即使在费多罗夫先生的支持者中，对其提议的反应也不一。在一些人看来，这显得冲动。在俄罗斯预计将发起一场前所未有之猛烈无人机和导弹攻势的冬季到来之前，这可能会将乌克兰政治推入危险境地。“如果有计划，那也是个糟糕的计划，”一位通常支持费多罗夫先生的议员说道。尽管如此，这次干预已将选举问题重新带回公众讨论。目前，乌克兰人并不想要选举。但轰炸和政治混乱可能会改变这一点。 ■
法国
西方文明
卡马雷-叙尔-埃格 (CAMARET-SUR-AIGUES)
法国沉迷于美国排舞
在法国乡村的夏天，是街道庆典、冰镇桃红葡萄酒以及美国排舞（line-dancing）的季节。为了让当地人起舞，DJ 只需要说出“Le Madison”这个词——这是一种起源于 20 世纪 50 年代美国中西部的排舞——村民们就会立刻排成行。法语国家乡村舞与排舞联合会列出了全国 500 多个俱乐部，名称如“美国精神乡村舞”（American Spirit Country）或“青蛙踏步”（Froggy Stomp）。
一顶 38 升的帽子
另一项统计记录了 2,000 多个“乡村舞”（danse country）团体。一项民调显示，近十分之一的法国成年人参加过排舞。今年 5 月，法国在南部的萨隆-普罗旺斯（Salon-de-Provence）镇举行了一场全国排舞“公开赛”。
在罗讷河谷人口仅 4,500 人的卡马雷-苏尔-埃格（Camaret-sur-Aigues）镇，“美国精神乡村舞”俱乐部每周开设两节排舞课和乡村舞课。参与者的年龄从 10 岁到 73. 岁不等。有些人戴着斯特森牛仔帽，穿着牛仔靴；有些人则穿着牛仔裤和运动鞋。“在过去两年里，排舞（le line）真正流行了起来，”该俱乐部主席佐拉·莱斯布罗斯（Zorah Lesbros）说道，她观察到年轻人中出现了一股浓厚的兴趣。“任何体型的人都可以跳，你可以结交朋友，而且很有趣。”
这种对美国文化片段出人意料的热情可以用什么来解释？法国一直认为自己是一个拥有高雅文化和艺术造诣的地方。夏尔·戴高乐在 1959 年创立了文化部，旨在将此类艺术普及给所有人。但实际上，这个国家长期以来对美国流行文化有着一种秘密的痴迷。去年法国票房前十的电影中，有九部是美国电影。在具有部分比利时血统的强尼·哈利戴（Johnny Hallyday）身上，法国找到了自己的“疯狂埃尔维斯”。巴黎迪士尼乐园是该主题公园在欧洲唯一的据点。在其迪士尼小镇的“比利鲍乡村西部沙龙”（Billy Bob's Country Western Saloon）经常举办排舞活动。
这种狂热也揭示了那些巴黎政治精英倾向于忽视的地方——而这种忽视可能让他们陷入危险。政治分析师热罗姆·富尔凯（Jérôme Fourquet）和西尔万·曼特纳赫（Sylvain Manternach）对“乡村舞”的一项研究指出，这种舞蹈的吸引力与 2018. 年“黄马甲”抗议者聚集的半农村地区环岛之间存在地理和社会相关性。他们认为，这种舞蹈提供了抗议者同样寻求的团结感，尽管目标不同。
总体而言，近四分之一的法国人表示他们被美国西部的文化“吸引”。作者写道：“‘乡村舞’在法国精英阶层中完全不为人知，而当它被提及的时候，大多是以讽刺的方式。”包括明年春季大选领先的总统候选人玛丽娜·勒庞（Marine Le Pen）在内的民粹主义政治人物，往往能更好地把握这一点。她甚至被拍到戴着斯特森牛仔帽，并定期在乡村活动中随歌而唱。
一些资深的法国爱好者可能通过让-吕克·戈达尔 1964 年的电影《分道扬镳》（Bande à Part）中的一个场景而知道“Le Madison”。而最近，现代乡村流行音乐的成功吸引了新成员。YouTube 上的法国排舞教程配的是泰勒·斯威夫特（Taylor Swift）的歌曲。一个关于如何跳“Le Madison”的视频拥有超过 100 次观看。在远离巴黎木地板舞厅的夏季户外舞会（bals）上，一场活动能否成功的关键，在于美国文化热潮带来的秘密快感与法国乡村传统的碰撞。■
欧洲 27
集体主义下的个人主义
瑞典如何成为一个由“大政府自由至上主义者”组成的国家
很少有意识形态主张能让伯尼·桑德斯（一位左翼程度之深，甚至可以作为欧洲中间派竞选公职的美国政治人物）与特朗普主义的激进分子查理·柯克（Charlie Kirk）达成一致。但奇怪的是，他们之间一个罕见的共识点便是瑞典的吸引力。对于这位语气粗鲁的参议员来说，这个斯堪的纳维亚国家是社会主义常识的源头，美国应当效仿。而对于那位一年前在大学宣讲其右翼信条时被谋杀的保守派中坚分子来说，这个宜家（IKEA）和 Spotify 的故乡同样值得钦佩。然而在他看来，瑞典的繁荣并非因为其庞大的政府，而是在于尽管有庞大的政府依然能繁荣——这完全归功于一种德克萨斯人也能认同的、天生的“不要践踏我”的个人主义。
政治光谱中两个截然相反的阵营都将一个全球治理最出色的国家之一的成功归功于自己的理念，这并不令人惊讶。令人惊讶的是，双方可能都有一定的道理：瑞典是一个矛盾的结合体，是集体化实施个人主义的故乡。
认为瑞典是社会主义乌托邦的论据与其肉桂卷一样传统。这个国家通过对富人征收重税——最高所得税率高达 52%——来资助政府，其政府支出几乎占 GDP 的一半，位居世界前列。瑞典的福利国家与其平板包装家具一样出名。这种呵护简直是从摇篮开始的；新父母可享受长达 480 天的带薪假期。其他左翼标志性特征还包括免费大学、低收入不平等以及发达国家中最广泛的租金管制。三分之二的瑞典从业者加入了工会，除极小的冰岛外，这一比例在所有国家中最高。所有这些“人类博爱”的理念都体现在政治中：下个月，中左翼的社会民主党将连续第 33 次在议会选举中获得最高票，这一纪录可追溯到一个多世纪前（尽管由于竞争对手组建执政联盟，该党曾有段时间不在权力中心）。
但同样有道理的论点是，瑞典也有自由至上主义的一面。诚然，所得税高得惊人，再分配程度也高于自由市场支持者的主张。但在许多其他方面，这个国家对个人自由的珍视程度堪比蒙大拿州的生存主义者。政府规模虽然庞大，但瑞典的公共服务通常由私营部门提供：大约三分之一的瑞典高中毕业生毕业于由非国家运营机构管理的学校，其中许多是以营利为目的的。瑞典是一个罕见的没有遗产税的国家。在 covid-19 疫情期间，它因实施的限制措施少于大多数国家（无论是严格的封锁还是口罩令）而显得特立独行。它缺乏全国性的最低工资标准，这与包括美国在内的发达经济体惯例相悖。与欧洲大多数地区不同，雇主在雇佣和解雇员工方面很大程度上是自由的。这一切导致瑞典的人均亿万富翁数量比美国多出约 50%。如果这就是社会主义，那么这种变体对游艇销售出奇地有利。
在急于将瑞典的成功归功于自己的过程中，双方都部分地误读了事实。在 20 世纪 90 年代初政府规模大幅缩减之前，社会主义的论点更为准确。而那些对缺乏最低工资赞不绝口的自由至上主义者则未能意识到，在大多数行业中，工资率是由工会与雇主讨价还价决定的——这绝非米尔顿·弗里德曼梦寐以求的景象。
但这种斯堪的纳维亚悖论之所以存在，是因为其中隐藏着某种更奇特的情况。因为瑞典确实与众不同。在发达国家，集体主义和个人自由通常被视为竞争对手：政府承担的责任越多，公民自立自强的空间就越小。瑞典则走了一条不同的道路。在20世纪，国家的扩张不仅是为了缓解不平等或为人们提供困难保障，更是为了减少人们对他人的依赖——包括对家人的依赖。一个庞大且非人格化的国家被视为增强了自主权：瑞典人认为，在需要帮助时，依赖一个非人格化的官僚机构，比依赖父母或某些慈善机构要好。从这个角度看，更大的国家意味着更多的自由，而非更少。
其结果就是两位历史学家亨里克·贝格伦（Henrik Berggren）和拉尔斯·特拉加德（Lars Tragardh）所称的“国家个人主义”（statist individualism）。在他们最新更新英文版的著作《瑞典爱情理论》（The Swedish Theory of Love）中，他们认为瑞典人开始珍视那些基于自由意愿而建立的关系，而非基于物质必要性而维持的关系。公共托儿服务帮助女性避免在经济上依赖丈夫，国家养老院将子女从赡养年迈父母的义务中解放出来，以此类推。（甚至婚姻也受到一定的质疑：在法国或德国，家庭是基本的纳税单位，但在瑞典，所有成年人都独立申报税务。）美国父母在送子女上大学时，必须提交收入证明，子女才能获得奖学金资格。相比之下，瑞典年轻人被认为是独立的：父母的收入无关紧要。他们的独立是国家赋予的礼物，尽管维持这种独立代价高昂。
难以捉摸的瑞典模式
一个欧洲小国的历史发展对于其他欧洲人来说似乎无关紧要，对美国人来说更是如此。很少有政体能具备让瑞典式“集体个人主义”成为可能的条件：大约七个世纪的法治、对国家深厚的信任，以及一种可能让人感到死板且从众的共识倾向。如今，瑞典有五分之一的人口出生在海外，瑞典自身能否维持其独特的道路？
但针对社会态度的跨境调查表明，大多数国家正变得越来越像瑞典及其邻国：世俗化社会将自我表达置于对家庭或传统的顺从之上。这使得瑞典特殊的制度安排——一种强势的国家主义，它帮助人们不仅实现了摆脱经济匮乏的自由，还实现了摆脱社会压力的自由——不再仅仅是北欧的一种奇闻。对于美国及其他地区的局外人来说，“大政府自由主义”可能令人费解。但对瑞典人而言，这并不矛盾：他们只是想在共同协作的同时，被允许独立生活。■
葡萄牙
航空、航天 与国防
AICEP
葡萄牙贸易与投资局
欧洲的大西洋工业前沿。 精准与规模的交汇点。
87%
ASD葡萄牙集群包含 160 多个实体，其 总产量出口额约占 87%
第3名
葡萄牙拥有欧洲第 3 高的 工程专业毕业生比例， 反映出其拥有高素质的劳动力 和强大的 STEM 人才储备
77%
2022 年至 2025 年间， 被归类为具有军事用途的 商品的累计名义增长率
25%
葡萄牙的国防支出 每年增长 +25%， 而北约在同一时期（2023-2025 年） 的年复合增长率（CAGR）为 9%
构建在整个价值链上的集群
飞机 MRO（维护、修理与运行）及 航空结构
航天 系统与 任务
无人机与 自主 系统
航空电子 系统与 网络安全
国防与 军民两用
由研发驱动的 创新 中心
葡萄牙。为未来而设计。
投资欧洲国防能力成型之地。 投资葡萄牙。
aicep.london@portugalglobal.pt
COMPETE 2030
2030
由 欧盟 共同资助
29
美国
水战争
一个被忽视的漏洞
华盛顿特区
这个世界上最富有的国家在保护关键基础设施方面依然无能
水是一种战略漏洞，这一事实对于军事指挥官来说并非新闻。当东哥特人在 537 年想要瘫痪罗马时，他们摧毁了其输水渠。近期，尽管国际法将此类攻击定义为潜在的战争罪行，但水利基础设施仍是美国与伊朗冲突中的重要目标。今年 3 月，唐纳德·特朗普威胁要“彻底摧毁……可能包括所有”伊朗的海水淡化设施，这些设施对于该国干旱南部省份的生存至关重要。伊朗称，位于格什姆岛的一座此类设施遭到了轰炸。巴林、科威特和阿拉伯联合酋长国的水厂也曾遭到导弹袭击，据推测这些导弹来自伊朗。
因此，当美国自己的水利设施在今年夏天遭到攻击时，这件事虽然值得关注，但并不令人意外。7 月 27 日，明尼阿波利斯附近的枫平原（Maple Plain）宣布进入紧急状态。同一天，服务于亚特兰大南部 300,000 人的佐治亚州克莱顿县水务局（Clayton County Water Authority）在压力下降导致供应中断后，要求客户将水煮沸后再使用。总计有黑客潜入美国至少 7 个州的水务和废水处理设施；一些报告显示数量超过 12 个。明尼苏达州受灾最严重，有 30 个社区水系统受到影响。美国官员的初步评估表明，与伊朗有关的团体对此负责。
→ 本节其他内容
30 筹款之战
31 俄克拉荷马州的公民教育课
32 共和党人与 WNBA
33 奥巴马医改的脆弱性
34 列克星敦：重新审视新冠肺炎（covid-19）
黑客攻击可能会中断供水并导致饮用水不安全，尽管人们认为最近的攻击尚未造成此类后果。问题在于，政治家们是否终于会采取行动，在下一次大规模袭击之前威慑黑客。8 月 13 日，两名民主党参议员艾米·克洛布彻（Amy Klobuchar）和亚当·希夫（Adam Schiff）提出了《水资源网络盾牌法案》（Water Cyber Shield Act），该法案包括赋予美国环境保护局（EPA）更多权力。但 8 月 19 日，联邦官员已经发布了新警告：黑客正试图攻破水利设施和其他关键基础设施中使用的西门子设备。
正如最近的攻击所证明的，水厂并非唯一受到威胁的基础设施形式，但它们尤其容易被渗透。电网必须符合由联邦能源监管委员会监督的网络安全标准。但水务领域不存在此类要求。运营发电厂的电力公用事业公司规模更大、资金更充足，且比水务运营商受到更严格的监管：最大的投资者所有公用事业公司，如加利福尼亚州的太平洋天然气电力公司（PG&E）以及东南部和中西部的杜克能源公司（Duke Energy），为数百万家庭提供电力。
水——重量大且运输成本高——更具局部性。约 90% 的公用事业公司（大多由地方政府所有）每家服务的人数少于 10,000 人。这意味着
30 美国
► 每家公司用于加固其通常已过时的计算机系统的资金更少。事实上，水利基础设施已经证明自己缺乏抵御攻击的能力。2013 年，伊朗黑客入侵了纽约一座小型水坝的控制系统，该事件在近三年内未被报道；2023 年，他们控制了宾夕法尼亚州一家水厂的水泵。
伊朗的黑客非常活跃；8 月 18 日，联邦检察官起诉了 17 名伊朗人，指控其攻击大学和公司的系统以窃取研究成果和知识产权。一场名为“伏特台风”（Volt Typhoon）的规模庞大的中国行动，一直试图潜入美国关键基础设施以准备破坏活动。
但技术水平较低的人也能潜入。2019年，堪萨斯州一名前水务局员工利用旧凭据登录了一个应用程序，导致清洗程序关闭。（他声称自己当时喝醉了，如果属实，这将进一步证明水务系统是一个容易攻击的目标。）两年后，黑客控制了佛罗里达州奥尔兹马尔（Oldsmar）的一处水务设施，并试图通过增加氢氧化钠（一种排水管清洁剂成分）的含量来毒害居民。这次尝试失败了，但他们能够轻易渗透系统，因为一名操作员的机器运行着 TeamViewer——这是一款流行的企业工具，允许 IT 支持人员远程访问机器。
最近的入侵手段相对简单。黑客攻破了作为计算机网络与物理系统之间接口的“运营技术”系统。这不需要太高深的技巧：此类系统通常通过移动网络连接到面向公众的互联网，且使用的凭据很弱，甚至根本没有凭据。7月30日，以及随后在8月19日，联邦官员恳请运营商将关键系统与互联网断开。
此前试图在该碎片化领域强制执行网络安全标准的努力在很大程度上失败了。在拜登政府期间，美国环保署（EPA）试图强制各州审查并报告水务系统的网络威胁。密苏里州、阿肯色州和爱荷华州的共和党州总检察长，以及两个行业团体——美国水务协会（AWWA）和国家农村水务协会（NRWA）共同提起诉讼，理由是联邦政府过度扩张。在联邦法院对环保署的努力发布暂缓执行令后，该机构撤回了要求。
当国会采取行动时，措施极其有限且生效缓慢。《关键基础设施网络事件报告法》于2022年通过，要求重要部门的组织在72小时内报告重大网络攻击。其细则预计在9月才能最终确定，距此已过去四年多。
现在可能出现了采取更多行动的动力。本月，一个具有公民意识的黑客组织 DEF CON Franklin 与 NRWA 合作启动了“水务观察中心”（Water Watch Centre），这是一项旨在为小型公用事业公司提供私营部门网络安全支持的计划。克洛布查尔（Klobuchar）女士和希夫（Schiff）先生提出的法案并非唯一在考虑中的提案。
8月5日，此前抵制拜登政府提案的 AWWA 支持由阿肯色州共和党众议员里克·克劳福德（Rick Crawford）发起的《水务风险与韧性组织建立法案》。该法案将创建一个独立机构，在环保署的监督下起草最低网络安全标准，这与电力公用事业的要求相 mirrored。同一天，共和党参议员汤姆·科顿（Tom Cotton）致信财政部长斯科特·贝森特（Scott Bessent），敦促修改税法并进行其他监管调整，以鼓励水厂投资更好的安全设施。
华盛顿能否迅速行动在很大程度上取决于特朗普先生。总统提出的预算案（参议院将于9月审议）将把战争部的预算增加44%，至$1.5trn。这包括用于伊朗战争以及用于保护美国免受导弹攻击的高科技“金穹顶”（Golden Dome）的资金。他对保护美国水务系统的兴趣则较低。
他的预算将削减环保署用于水务网络安全的最大资金来源近90%。7月，他暗示明尼苏达州遭受攻击的责任在于该州“腐败”的州长蒂姆·沃尔兹（Tim Walz）——并且根本没有发生过伊朗的攻击。他表示：“伊朗有比担心明尼苏达州更大的麻烦。”负责网络防御的主要联邦机构——网络安全和基础设施安全局（CISA）目前处于混乱状态，领导层动荡，士气低落，且自特朗普先生重返白宫以来，人员削减了三分之一。参议员在休会结束后可能会敦促他采取更多行动。而黑客们并不会等待。■
政治筹款
共和党的资金问题
华盛顿特区
唐纳德·特朗普拥有庞大的战争基金，但他可能不愿分享
在期中选举之前，共和党人声称他们至少有一项优势：资金，而且数额巨大。乍一看，提交给美国联邦选举委员会（FEC）的最新申报文件显示，共和党拥有惊人的现金优势。在其党组织和政治行动委员会（PAC）中，共和党已积累了约10亿美元，是民主党持有金额的两倍多（见下页图表）。
然而，这给共和党带来的鼓励微乎其微。该党最大的战争基金属于 MAGA Inc（唐纳德·特朗普的超级PAC），而总统拒绝透露他将向个人竞选活动贡献多少资金。虽然共和党仍然受制于总统的
美国 31
[—接前文—]心血，但民主党的筹款是去中心化的。在竞争激烈的选区中，个人候选人的筹款额超过了他们的对手。这意味着他们手头有更多现金，同时也拥有更多热情的支持者。
金钱不能买到胜利，但它并非无关紧要。候选人筹集资金是为了支付竞选运营费用和购买广告。动员投票（Get-out-the-vote）的努力可以提高候选人的胜算，尽管当竞选者已经广为人知时，这种效果似乎会下降。在联邦职位的竞争性大选中，通常很难衡量额外广告的效果，因为电波频道已经饱和。最近在威斯康星州举行的州长初选提供了一次罕见的自然实验，尽管这不是一场大选。民主社会主义者弗朗西斯卡·洪（Francesca Hong）没有购买过任何电视广告。《经济学人》在控制了各县选民中位年龄和教育水平等因素后分析认为，缺乏电视广告导致她在8月11日的民主党初选中惨败。
如果现金数额至关重要，那么竞选活动如何筹集资金同样重要：资金既是竞选成功的因素，也是其健康状况的指标。强有力的候选人往往能吸引其所在州普通选民的捐款，这是他们受欢迎程度的标志。这就是为什么《经济学人》在对期中选举的预测中，将个人捐款作为一项预测指标。这也是共和党不能自信地认为其现金优势将对他们有所帮助的另一个原因。
特朗普先生已被证明在动员捐赠者方面特别擅长（想想那些在海湖庄园每席$1m的晚宴）。但这些捐款流入了他的超级PAC，而不是候选人本身。MAGA Inc 占据了共和党总筹款额中至少4亿美元，这提醒人们，总统仍然是该党的权力来源。
他看起来不愿分享这份权力。到目前为止，MAGA Inc 仅花费了230万美元，占其资源的0.6%，来帮助共和党赢得选举。特朗普先生最近吹嘘道：“我几乎可以将这笔钱花在我想花的任何地方。”尽管总统声称他将利用其PAC来帮助共和党人，但该组织尚未公布其支出计划，甚至没有将其分享给共和党盟友。
虽然共和党的筹款是中心化的——理论上强大但效果有限——民主党则在经历相反的现象。该党的中心化筹款陷入困境。小额捐赠者对他们认为的民主党建制派感到厌倦。顶级捐赠者仍对2024年总统竞选期间资金的支出方式感到沮丧；民主国家委员会（DNC）欠债$2m的事实并未增强信心。根据华盛顿的一句俏皮话，DNC 现在代表的是“不要捐款”（Do Not Contribute）。但许多民主党个人候选人并未受到影响——捐赠者仍在捐钱，但他们绕过了中央委员会，直接将资金交给竞选团队。
以德克萨斯州的参议院选举为例。民主党提名人詹姆斯·塔拉里科（James Talarico）在今年第二季度筹集了 2800万美元，而他的共和党对手肯·帕克斯顿（Ken Paxton）仅筹集了微不足道的 160万美元。塔拉里科总筹款额的一半以上来自小额捐款，而帕克斯顿的这一比例仅为 12%。民主党候选人在大额捐款方面也占据优势。一直避开民主党全国委员会（DNC）的大金主里德·霍夫曼（Reid Hoffman）最近向一个支持塔拉里科的政治行动委员会（PAC）捐赠了 1000万美元。
民主党候选人在其他州也持有优势。在竞争最激烈的九场参议院选举中，他们手头拥有约 1.16亿美元 现金，而共和党仅有 5100万美元。反映选民热情的是，平均而言，小额捐款者（向竞选活动捐款少于 $200 的个人）在民主党候选人捐款中所占比例为 51%，而共和党候选人仅为 24%。
共和党党内官员坚称，民主党候选人的资金优势是暂时的，而存储在共和党金库中（包括总统的金库）的资金很快将流向各项竞选活动。由于中期选举在不到三个月后就要举行，候选人们感到有些焦虑也是可以理解的。■
尽力筹款
选定的美国超级政治行动委员会（super PACs）及委员会，截至2026年6月30日的可用现金
| Republican-aligned 10.6亿美元 | |
| MAGA Inc 4.01亿美元 | | | Republican committees 2.77亿美元 | | | Senate Leadership Fund 2.39亿美元 | | | Congressional Leadership Fund 1.42亿美元 | | Spends on both 1.58亿美元 / Democrat-aligned 4.74亿美元 | | Democratic committees 1.36亿美元 | | | Senate Majority PAC 1.26亿美元 | | | Democracy PAC II 1.22亿美元 | | | House Majority PAC 9000万美元 |
选定参议院竞选活动筹集的资金，单位：百万美元* 2026年4月1日至6月30日
*仅限主要竞选委员会 †民主党候选人格雷厄姆·帕尔默（Graham Palmer）于7月10日退出竞选 来源：美国联邦选举委员会；《国家期刊》（National Journal）
政治文明
你做得很好，俄克拉荷马！
俄克拉荷马城
一个曾被极端主义伤害的州试图对此进行防范
美国两个最著名的跨党派政治组织直到最近都由俄克拉荷马州领导，这并非巧合。俄克拉荷马城共和党市长大卫·霍尔特（David Holt）曾领导美国市长会议，而俄克拉荷马州共和党州长凯文·斯蒂特（Kevin Stitt）则领导国家州长协会（NGA）。这些团体在美式政治中提供了一种新颖的东西：文明。去年，81%的民主党人和70%的共和党人表示，对方政党让他们感到愤怒。但在最近于加利福尼亚州长滩（市长会议）和俄克拉荷马城（州长会议）举行的集会上，民主党人和共和党人击掌、拥抱，并雄辩地阐述美国极化政治的危险。
今年夏天，霍尔特先生和斯蒂特先生将权杖交给了新领导人。但值得思考的是，为什么俄克拉荷马州的政治人物似乎对让美国重新变得友善特别感兴趣。两人都曾违背党意，并公开反对总统唐纳德·特朗普。他们珍视与民主党人的友谊。管理城市和州需要一种务实精神，而这并不是在国会任职的先决条件。更重要的是，俄克拉荷马州的许多政治领导人在20世纪90年代成年，当时他们见证了本州以同情心应对极端主义。
1995年，蒂莫西·麦克维（Timothy McVeigh）引爆了一枚炸弹，炸毁了俄克拉荷马城中心的联邦大楼。近170人丧生。这仍然是美国历史上最致命的国内恐怖主义行为。麦克维希望对他认为的暴政政府进行报复。霍尔特先生当时在读高中，正准备发表一次学生会副主席的竞选演说。“后台传出传闻说市中心发生了爆炸，”他回忆道，“我们回到教室，他们正推着电视机进来。”
爆炸发生后，人们对受害者、幸存者和急救人员倾注的支持被称作“俄克拉荷马标准”（Oklahoma Standard）。“这个词被用来概括我们的社区如何接纳所有前来帮助的人，”霍尔特先生说。它已成为“那些试图超越黑暗的人们的一个通用代称”。
霍尔特先生表示，这次爆炸塑造了他的政治观。“这座城市现在的部分特质就是试图尽我们的一份力，抵制极化和非人性化，” ►
32 美国
► 他说，“因为我们市中心留下的伤疤提醒着我们，那样做最终会导致什么。”自从在四年级教室里为乔治·H·W·布什竞选以来，他一直是一名共和党人。但作为市长，他一直试图促进包容，宣布了俄克拉荷马城首个LGBTQ骄傲周，并在2020年会见了争取种族正义的抗议者。去年，在爆炸纪念活动上，230名市长承诺“拒绝政治暴力，重新致力于美国实验”。
1995年正值大学四年级学生的斯蒂特先生将他的礼貌归功于“俄克拉荷马常识”。“我的祖母说过‘用蜂蜜比用醋更能吸引苍蝇’，”他说，“所以我尽量对每个人都友善。”
特朗普时代的礼让
霍尔特先生和斯蒂特先生承认，要克服政治家们倾向于刻薄的激励机制将十分困难。特朗普是美国的“首席霸凌者”——而美国人两次选举他担任总统。但这两位俄克拉荷马州人士旨在树立一个更好的榜样。他们并非总能成功。在本月早些时候举行的全美州长协会（NGA）会议上，斯蒂特先生讽刺了民主党大本营加利福尼亚州，但随即表示遗憾。他在台上说：“天哪，我不该那样说，这是一个跨党派的协议。”
他们还提出了更具体的改革方案。霍尔特先生认为，取消只有党内成员才能投票的封闭式党内初选，将使极端候选人更难获得提名。（关于开放式初选是否能产生更温和候选人的研究结果不一。）而始终坚持保守主义的斯蒂特先生则开出了更多联邦主义和精简中央政府的药方。他说：“中央政府已经变得如此庞大，以至于我们俄克拉荷马州人担心，如果卡玛拉·哈里斯当选，她会让俄克拉荷马州变得像加利福尼亚州一样。”
另一个药方就是“俄克拉荷马标准”本身。NGA会议上的其他州长对该州的公民精神表示钦佩。犹他州共和党州长斯宾塞·考克斯长期敦促美国人要“更好地分歧”，他感叹道：“俄克拉荷马标准有趣的地方在于，它曾经是整个美国的标准。”
在俄克拉荷马市的纪念馆外有一道围栏，哀悼者在那里放置鲜花和亲人的照片。每年，凯尔·根泽都会给在爆炸中丧生的母亲留一封信。他写道：“今年迎来了一个最美好的时刻——你的第一个曾孙女出生了。抱着她时，我们不禁想起了你。”根泽先生是一位高中校长，他认为继续讨论这次爆炸至关重要，这既是为了缅怀受害者，也是为了提醒生者，分歧不必演变为仇恨。■
体育
政治得分
亚特兰大
为什么保守派无法停止讨论女子篮球
美国女子职业篮球联赛（WNBA）的组织者曾一度恳求记者报道该赛事。当时的笑话是，参加一场新闻发布会只需要带上相机并准备一个问题即可。而现在，现场已是座无虚席。这在很大程度上反映了该联赛人气激增，但它也成为了美国文化战争中一个出人意料的战场。
最近，关注焦点集中在印第安纳狂热队的后卫索菲·坎宁安（Sophie Cunningham）身上，一些人称她为“MAGA芭比”。坎宁安并未公开表明自己是共和党人还是民主党人。但在上个月发表的一篇采访中，她表示应禁止跨性别女性参加女子体育赛事。大多数美国人同意她的观点。然而，在期中选举之前，这个问题已成为右翼政治家和评论员的执念。
证据就在福克斯新闻（Fox News）身上，该频道通常很少关注女子体育。根据左翼监督机构 Media Matters 的数据，在 6 月 25 日至 8 月 11 日期间，该频道制作了 200 多个关于 WNBA 的片段，其中至少 130 个是关于跨性别运动员的。报道之密集，以至于该网络最著名的主持人之一杰西·沃特斯（Jesse Watters）在节目中抱怨道：“这太荒谬了。”
8 月 8 日，当坎宁安被对方球员迪乔奈·卡林顿（DiJonai Carrington）粗暴犯规时，一些保守派将其解读为对其观点的报复。佛罗里达州的一位共和党总检察长暗示，如果比赛在他的州举行，他可能会指控卡林顿犯有袭击罪和殴打罪。
这次犯规（见图）使这场风波的另一个要素变得更加突出：种族。坎宁安是白人；赛后，黑人球员卡林顿在社交媒体上发布了“白色特权 @indianafever”的文字。她并未解释其含义。但另一位狂热队的白人球员、该联赛最大明星凯特琳·克拉克（Caitlin Clark）也成为了 MAGA 的一个著名议题（cause célèbre）。上个月，11 位共和党议员致信 WNBA，抱怨克拉克在场上被推搡。他们引用报道称，这“可能是种族动机所致”。唐纳德·特朗普也同意克拉克受到了“相当粗暴的对待”。
就联赛而言，其处理方式并非总是得当，有时要求球迷遮盖关于跨性别运动员的 T 恤，随后又改变主意。但联赛也指出了某些报道的荒谬之处：WNBA 中并没有跨性别运动员。两名前 NBA 球员且支持特朗普的人声称想要加入；联赛称其行为是“噱头”。
为什么共和党现在要触碰这些敏感点？随着期中选举临近，选民关注的是生活成本问题，这对执政党来说是个难题。而种族和跨性别问题则提供了更有利的阵地，可以用来将民主党描绘成脱离群众。6 月，保守派在最高法院赢得了一场胜利，法院裁定各州可以禁止跨性别女性参加女子体育赛事。
与此同时，WNBA 的上座率和收视率正在激增。“我们的关注点在篮球上，”克拉克最近表示。她正处于一个打破纪录的历史性上升期。这似乎同样值得关注。■
有点太直白了
美国 33
奥巴马医改
健康状况堪忧
纽约
在左翼呼吁全民医疗保险之际，民主党的标志性医疗法案正陷入困境
“医疗改革……是我们社会一项伟大的未竟事业，”民主党参议员泰德·肯尼迪在2009.去世前曾如此观察。美国是唯一一个没有全民医疗保健的大型富裕国家。相反，其体系像是一块补丁布：年长者、贫困者和残疾人由政府计划覆盖；大多数劳动年龄的美国人通过工作获得保险。但那些没有此类覆盖的人则十分可怜，他们必须在昂贵且往往令人困惑的个人保险市场中摸索。
2010,通过的《平价医疗法案》试图解决这个问题。它在基本维持美国原有体系不变的同时，扩大了针对贫困者的公共计划——医疗补助（Medicaid），并创建了新的“交易平台”，让个人可以在其中比较并购买计划。即便是在私下里，奥巴马医改最坚定的支持者也承认，这只是一个不完美的绷带，尤其是对于那些处于个人市场的人来说。对于像伯尼·桑德斯这样呼吁单一支付者体系的左翼参议员来说，这远远不够。现在，新一代的民主党人接过了这一事业，例如该党在密歇根州的参议院提名人阿卜杜勒·埃尔-赛义德，他将全民医疗保健作为其竞选的核心。新数据解释了这种紧迫性：奥巴马医改的绷带正在脱落。
这部分归因于医疗补助（Medicaid）的削减，也因为交易平台的健康状况不佳。奥巴马医改提供了一些补贴以抵消个人计划的成本。但该法案强制购买保险的规定在2019,实际上被取消，直到2021,交易平台的参与度一直有限，直到当时国会对已符合条件的人提供了更丰厚的补贴，并向部分此前不符合条件的人开放。随后，参保人数在2025.前几乎翻了一番。民主党试图延长这些额外的财政援助，但未能成功；根据五角大楼的估计，这在今年将花费约350亿美元——几乎与伊朗战争的直接成本相当。现在这些补贴减少了，个人市场看起来比以往任何时候都更脆弱。
来自保险公司和州级市场的数字显示了这种回撤的影响。通过交易平台购买保险的人数已从去年的21.8m下降到19.2m。非党派计分机构国会预算办公室（CBO）预计，到今年年底，这一数字可能会降至17m以下。在退出交易平台的人中，大多数似乎处于无保险状态。
退出者往往比留在交易平台的人更健康，这使得剩余的投保池保险成本更高。病患人群的增加与医疗成本的普遍通胀相结合，导致的结果是保费上涨：今年每月平均支付额从$113增加到$178。许多参保者现在购买的是保障更少的计划。在过去一年中，个人在保险开始赔付前必须自掏腰包支付的平均金额跳升了近40%,达到$3,800.
这引发了人们对一个糟糕循环的担忧：更高的保费和更受限的计划使得保险更缺乏吸引力，尤其是对于更健康的客户。随着他们离开交易平台，保费被推得更高。根据智库KFF的数据，保险公司已经提出了2027,保费中位数增长15%的方案。如果州监管机构批准这些费率，补贴前的保费自2025,以来将增长超过三分之一。KFF设想了一位在印第安纳州年薪$65,000.的40岁女性。2025,在补贴的情况下，她的月保费为$316.今年是$477.明年可能会接近$550.
这是否威胁到交易平台的生存能力？国会预算办公室（CBO）估计，到2028年，在交易平台购买保险的美国人数将比去年减少1000万。但交易平台可能会避免崩溃，因为剩余的补贴足以说服足够多的人留下来。“这些变化基本上让我们回到了2016到2019年期间交易平台的模样，”哈佛大学的健康经济学家本·索默斯（Ben Sommers）预测道。交易平台的规模将缩小，参与的保险公司将减少，而客户的健康状况将更差。
没有保险的人数也将增加。证据表明，他们会推迟必要的治疗，例如某些处方药和手术。“我们从十多年来对《平价医疗法案》（Affordable Care Act）的研究中得知这一点，”索默斯先生解释道。即使这些人无法支付费用，医院仍必须为他们提供急诊护理。根据咨询公司Kaufman Hall的数据，5月份未支付的护理费用比一年前平均增长了16%。营利性医院集团HCA Healthcare估计，新增的无保险人群今年将使该连锁机构损失至少$1bn的营业利润，相当于预测净收入的约15%。KFF的辛西娅·考克斯（Cynthia Cox）表示，近年来交易平台的参保情况表明，“人们普遍想要医疗保险，问题仅仅在于他们是否负担得起。” ■
关键统计数据
美国，《平价医疗法案》市场注册人数*，百万
*已选择计划并支付首期保费 †截至5月15日 来源：KFF
34 美国
被告席上的医生
真正的新冠肺炎丑闻在于，美国始终未能正视那些不必要失去的生命
尽管国会在调查美国应对新冠肺炎大流行的情况时表现得如此热忱，但它对探索一个基本问题却表现出奇怪的冷漠：为什么相对于其他富裕国家死于该病毒的人数，美国死亡的人数如此之多？理解这个答案对于防范下一次不可避免的大流行似乎至关重要。
许多国家的反应都很缓慢，但随着它们了解病毒的传播力以及疫苗的问世，它们迅速降低了死亡率。美国则不然。《经济学人》在 2021. 年 9 月报道称：“美国的死亡率现在大约是富裕世界其他地区的八倍。”总计，美国有超过 110 万人死于新冠肺炎，大约每百万人口中失去 3,500 条生命。这应该让每一位处于能够抗击大流行地位的官员感到不安。唐纳德·麦克尼尔（Donald McNeil）在其著作《瘟疫的智慧》（The Wisdom of Plagues）中指出，德国的死亡率约为每百万人口 2,000 人，而加拿大则少于 1,400, 人。他写道，如果美国能取这两者的中间值，即每百万人口约 1,700 人，“我们将减少约 540,000 人的死亡”。“谁能为他们发声？”
国会没有。当参议院国土安全委员会最近质询安东尼·福奇（Anthony Fauci）博士时，美国这种致命的低劣表现再次被忽视。福奇博士曾是唐纳德·特朗普以及随后的乔·拜登在新冠肺炎问题上的关键顾问。（特朗普后来与他反目。）听证会转而集中在福奇博士成为共和党替罪羊的那些惯常执念上——其中一些荒诞不经，另一些虽然合理，但至多与保护未来瘟疫的潜在受害者只有很远的关系。民主党人大多在发表演讲而非提出问题，而共和党人则在追究各种理论：病毒是从中国实验室泄露的，比尔·盖茨扮演了某种阴险的角色，福奇博士试图通过其服务奖金来中饱私囊——当他们并不反对总统竞选 100万美元 诺贝尔奖时，这种抱怨显得十分奇怪。
在国家卫生研究院（NIH）任职的 54 年里——其中 38 年担任国家过敏和传染病研究所所长——福奇博士多次在国会作证。现年 85, 岁的福奇博士在质询中被迫反复援引宪法第五修正案，甚至在被问及领带颜色时也是如此。在他的开场白中，他读到时双手在颤抖，福奇博士表示他的律师建议采取这种手段，因为委员会主席、来自肯塔基州的兰德·保罗（Rand Paul）已明确表示，尽管拜登先生发布了全面赦免，他仍意图将福奇送入监狱。换句话说，参议院无法调查这场席卷全国的健康危机，因为一名关键证人担心陷入伪证陷阱，而这种担心是合理的。
为什么这么多美国人死亡？高比例的肥胖、糖尿病、心脏病和高龄使他们更加脆弱。但政府的搞砸意味着，在 2020 年初，当其他国家每天检测数万人时，美国每天仅检测 100 人；口罩和封锁措施的部署毫无章法；接触者追踪对于这个第一个登上月球的国家来说证明过于困难。最荒谬的是，这么多美国人死亡，是因为太多人拒绝接种疫苗，而这些疫苗正是特朗普先生在担任总统期间最具有建设性的行为——成功地将其快速推向市场。
当然，正是这些令人烦恼的疫苗，迫使共和党人回避关于这场大流行的基本问题，转而纠缠于福奇博士。为了将特朗普先生的伟大与疫苗所谓的卑劣调和起来，他的政党在“双重思想”方面展现出了英雄般的壮举。“候任总统特朗普的‘曲速行动’（Operation Warp Speed）——该行动鼓励了COVID-19疫苗的快速研发和授权——取得了巨大成功，并帮助挽救了数百万人的生命，”众议院在2024年成立的调查COVID-19委员会在总结两年的工作时如此宣称。而紧接着的下一句话却是：“与承诺相反，COVID-19疫苗并未阻止病毒的扩散或传播。”
特朗普先生一直难以摆脱这种自找的矛盾。即便他拒绝敦促公众接种疫苗，他及其妻子梅拉尼娅在2021年1月离任前仍秘密接种了疫苗。但当拜登先生和其他世界领导人接种疫苗时，特朗普先生想要获得他应得的赞誉，并在那年12月告诉支持者他已经接种了加强针。“我们所做的是历史性的，”在人群发出嘘声时，他坚持道，“不要让他们夺走它。”此后，他通过质疑COVID-19以及其他疾病疫苗的有效性和安全性，让自己失去了这份荣耀，并危及了更多美国人。
胡说博士
针对福奇博士（Dr Fauci）的最新攻击是基于共和党参议员于8月10日披露的私人短信而编造的，称他掩盖了疫苗给孕妇带来的流产威胁。事实上，记录显示，福奇博士在公开和私下场合始终一致地建议权衡风险与收益，直到研究证明疫苗在怀孕期间是安全的——而随后的研究确实证明了这一点。他的批评者并未提供疫苗导致流产的证据。
福奇博士犯过错，包括早期没有披露一些顶尖病毒学家最初担心该病毒可能是人造的。他可能很傲慢。但他不对新冠疫情应对措施中最极端的过激行为负责。那些无休止的学校停课是地方官员和工会领导者的所为。保罗参议员（Senator Paul）公开的福奇博士私人日记暴露了他的虚荣心，以及他对媒体或某些名流每一次热情赞赏的欣喜。但日记也揭示了一位尽职的公职人员在尽力挽救尽可能多的生命，并与那些使美国相对于同行显得步履蹒跚的限制因素作斗争。他被传唤至参议院委员会，无疑是为了羞辱他，即便不是为了定罪。这件事应该让整个国家感到尴尬。■
35
美洲
权力失衡
里约热内卢
巴西国会权力过大，正引发问题
在过去的十年里，全球各地的民主制度都被潜在的独裁者地左右。从美国到土耳其和印度，强人通过在政府行政部门积累权力，削弱了能够制约他们的机构，尤其是立法机构。
因此，其他地方感到沮丧的立法者可能会羡慕他们在巴西的同行，因为那里的趋势正相反：在过去十年中，巴西的行政部门失去了对法院，尤其是对国会的影响力。权力平衡已变得如此扭曲，以至于巴西的立法者似乎不再关心10月大选中谁当选总统。有些人认为，无论谁获胜，都可能被迫屈服于他们的要求。巴西人可能正在发现，正如行政权力过大是对民主的挑战一样，一个权力过大的国会同样如此。
10月4日的竞选将使两位截然不同的总统候选人展开对决：现任总统路易斯·伊纳西奥·卢拉·达席尔瓦（Luiz Inácio Lula da Silva，简称卢拉）是一位80岁的左翼人士。他的对手弗拉维奥·博索纳罗（Flávio Bolsonaro）是民粹主义前总统雅伊尔·博索纳罗（Jair Bolsonaro）44岁的儿子，后者因在2022年选举失败后策划政变而被判处27年监禁。众议院全部513个席位以及参议院的三分之二席位也将重新选举。然而，尽管竞争两极分化，巴西最大的政治集团“大中心”（Centrão）——在即将离任的国会中占据46%的席位——已宣布保持中立。党内大佬将允许地方政治家支持他们选择的任何候选人。
→ 本节其他内容
36 阿根廷：马球超级大国
37 巴拉圭的另一种语言
在过去的十年里，随着一系列软弱的总统统治巴西，国会从行政部门手中劫持了权力。2016年，在一场卷入数十名立法者的巨额腐败调查中，立法者弹劾了当时的左翼总统迪尔玛·罗塞夫（Dilma Rousseff），部分原因是他们认为她没有采取足够措施保护他们免受调查人员的追究。罗塞夫女士的继任者米歇尔·特梅尔（Michel Temer）在缺乏公众支持的情况下执政。在2019年至2022年期间执政的博索纳罗先生，为了在灾难性的新冠疫情处理后保护自己免受弹劾企图，赋予了立法者对联邦预算更大的控制权。如今，巴西立法者控制着联邦预算中25%的可支配支出，而2015年这一比例仅为2%（见下页图表）。
如何花这笔钱
这种对预算的掌控力极不寻常。圣保罗因斯佩尔大学（Insper）的马科斯·门德斯（Marcos Mendes）和众议院预算顾问埃利奥·托利尼（Hélio Tollini）对比了经合组织（OECD，一个主要由富国组成的俱乐部）11个国家的国会修正案。在加拿大和澳大利亚等一些国家，立法者没有权利修改预算或为自己拨付资金。而在美国、法国、意大利和西班牙等其他国家，预算中仅有约 1% 的
36 美洲
▶ 可支配支出被分配给国会修正案。在大多数这些国家，立法者提出的预算修正案并不总是被接受。
巴西则并非如此。在巴西，预算现在必须始终包含一笔用于国会修正案的资金。国会议员通常不规定资金的用途。他们将政府必须分配给修正案的金额与收入增长挂钩，使得他们的拨款不可避免地增长。而且，他们几乎废除了有关资金使用透明度和可追溯性的规则。
国会控制的资金比例并不算巨大。巴西 6.5 万亿雷亚尔（$1.2trn）预算的 90% 以上用于强制性支出，如公共薪金、养老金和债务融资。即便国会对可支配部分的控制权在扩大，预算中用于强制性支出的份额仍在迅速增长。在 2026 仅有 3.7% 的预算，或 460亿美元，可用。其中，国会控制了 120亿美元。
便捷的资金获取提高了现任者的连任机会，因为这允许立法者资助选民喜爱的华丽项目。他们在地方政府的政治盟友也从中受益。在 2024, 当巴西人选出 5,500 多名市长和数万名市议员时，选择再次参选的现任者中，有创纪录的 81% 保住了职位。在即将到来的大选中，87% 的众议员将寻求连任，这是巴西历史上最高比例。
在巴西碎片化的政党体系中，总统很少能掌控国会多数席位，因此对预算的控制长期以来一直是行政部门强大的筹码。总统可以通过威胁扣押立法者的项目资金来巩固在国会的支持。现在，这种杠杆作用已大大减弱。今年 4 月，参议院否决了卢拉提名的最高法院法官候选人若日·梅西亚斯（Jorge Messias）。他是一个多世纪以来首位被否决的总统提名人。参议员们希望该席位由他们内部的人员填补。
动不得
国会正利用其新权力制定更多法律。在 1988 和 2004, 之间，巴西 80% 以上的新法律源自行政部门。到 2025 这一比例为 19%，而国会则占 75%。这本身并非坏事，但里斯本大学的贝亚特里思·雷（Beatriz Rey）指出，国会变得更容易受到游说团体的影响。去年，在农业综合企业团体的压力下，国会匆忙通过了一项削弱环境许可制度的法案。当卢拉否决该法案的大部分内容时，国会推翻了他的大部分否决。
赤裸裸的腐败或许是最大的风险。7 月，联邦审计
掌控钱袋子
巴西，通过国会预算修正案控制的联邦自由裁量支出，%
*截至8月15日
来源：H. Tolini 和 M. Mendes，INSPE，2024年（数据更新至2026年），《各地都这样吗？巴西及11个经合组织国家的议会修正案》
法院公布了100项所谓“Pix修正案”的样本，在这些修正案中，立法者将相对较小金额的资金直接发送给其当地市镇。样本中的拨款总额约为4000万美元，在2020年至2024年期间发送给了75个接收方。82%的转账被发现存在违规行为，包括资金流向与拨款规定项目无关的银行账户、预算严重虚高的合同，或用于从未完工项目的资金。审计
法院估计公共资金损失为1000万美元。样本规模虽小，但如果具有代表性，损失将是巨大的。国会在当前预算中批准了70亿雷亚尔的Pix修正案，这意味着可能损失了约3.4亿美元。
2024年，最高法院强制立法者指明接收联邦资金的银行账户，并提供拨款目标项目的详细信息。然而，立法者似乎准备展开反击。议员们已针对弗拉维奥·迪沃（Flávio Divo）提出了八项弹劾动议，该法官负责调查提交可疑预算拨款的立法者。（在巴西，最高法院负责对在任政治人物进行刑事调查。）
事实上，立法者们正得意地挥舞权力。自2021年以来，参议员已向最高法院法官（这些法官近年来自身权力过大）提交了105份弹劾请愿书。随着新的腐败调查继续将立法者卷入其中，请愿书的数量可能会增加，包括对公共养老基金欺诈行为的调查，以及巴西历史上最大银行欺诈案的后续影响。巴西的下一任总统不仅面临经济增长缓慢、高利率和极化的选民，还面临一个试图让他保持弱势且顺从的权力过大的国会。■
马力
楚坎经济学（Chukkanomics）
布宜诺斯艾利斯
阿根廷主导的马球经济为该国其他地区指明方向
“25年前我第一次来到这里，”马克西米利亚诺·费尔南德斯（Maximiliano Fernández）说道，“那时这里什么都没有。”如今，在皮拉尔（Pilar）和亨内拉尔·罗德里格斯（General Rodríguez）两镇之间，这片位于布宜诺斯艾利斯以西50公里的狭长地带，布满了马球俱乐部、马厩和庄园。每一块农田都被买下并改造成为纯净的马球场。管理着一支球队的费尔南德斯先生表示，阿根廷的马球世界是该国境内的一个“泡沫”：它拥有自己的经济体系。而且，这个经济体系在世界范围内处于领先地位。
马球——骑在马上的曲棍球——据称是世界上最古老的团队运动。其早期版本在公元前六世纪的波斯就已经出现。现代马球由英国军官在印度正式确立，随后在整个大英帝国迅速传播。它成为了国王的运动，被欧洲、美国以及近期的海湾国家精英阶层所接纳。然而，即便在阿根廷其他经济领域一片狼藉之时，它却在全球马球领域占据了主导地位。
与任何新兴产业一样，拥有原材料大有裨益。马匹长期以来在潘帕斯（Pampas，阿根廷中部的广袤草原）上繁衍生息。它们吸引了当地版本的牛仔——高乔人（gauchos）。马球在传入时，自然地契合了这里廉价马匹充足、骑手技术精湛、土地平坦且气候宜人的条件。阿根廷球员很快被认为是世界上最优秀的，且至今依然如此：在当今全球前30名的职业球员中，有27人来自该国。他们支撑起了一个基于服务的经济体系，富有的外国人雇佣他们以帮助赢得奖杯。
工业随后紧随其后。阿根廷的地主家庭建立了强大的育种计划，生产出世界上最顶尖的马球马（或称马球马驹，ponies）。这些产品在领域内处于领先地位：根据该国马球协会的数据，在去年参加英国女王杯（Queen's Cup）的1,140匹马中，有1,000匹拥有阿根廷血统。▶
美洲 37
通过将其在人力和马力资本方面的竞争优势最大化，阿根廷的马球经济享受了一场出口导向的繁荣。去年1月至11月期间，阿根廷向38个国家出口了2,900匹马（其中四分之三是马球马驹）。该国是全球领先的出口国。
为了保持领先地位，阿根廷的育种者和商人投资于研发，并开创了马球马的克隆技术。2010年出现了一次突破，当时一只克隆自 Dolfina Cuareterra（被列入阿根廷名人堂的最佳马驹）的马以创纪录的800,000美元成交。现在，许多马都是过去阿根廷名马的复制品。一些育种者正试图更进一步以保护其知识产权，他们仅出售克隆马的后代，而非克隆马本身。
少些繁文缛节，多些奖带
监管机构促成了创新：马球协会抵制繁文缛节并拥抱技术，而纯血马赛马运动则禁止这样做。但凡事皆有极限。当阿根廷生物技术公司 Kheiron 在 2024 年推出全球首批基因编辑马时，这些马被禁止参加比赛。毕竟，这种技术有一天可能会让国外的竞争对手培育出跑得比阿根廷著名血统更快的马。
就在阿根廷经济剧烈波动（2月份萎缩 2.5%，3月份增长 2.8%，随后在4月份收缩 1.5%）的同时，马球界似乎并未受到影响。这在一定程度上归功于稳定的外国直接投资。被称为“赞助人”的富裕外国人为阿根廷的马球泡沫带来了充足的资金。他们购买土地、雇佣球员并讨价还价买马，为马夫、蹄铁匠、训练师、兽医和建筑师创造了一种下渗经济。
结果便是持续的增长。二十年前，兽医胡安·马丁·巴蒂斯图塔（Juan Martín Batistuta）几乎完全从事赛马工作；而如今，他几乎完全从事马球马工作。“我从未如此忙碌过，”他说。而且这种繁荣并不局限于布宜诺斯艾利斯。“从科尔多瓦到巴塔哥尼亚，整个阿根廷出现了更多俱乐部，”费尔南德斯先生说。该国目前拥有 177 个。这项运动在全球范围内的知名度也在提升。2024 年发布的一部 Netflix 纪录片以及与 ESPN 达成的广播协议将其带到了新的受众面前。
阿根廷总统、马球爱好者哈维尔·米莱（Javier Milei）在试图重塑国家经济时，其支持率正在下滑。他或许可以看看皮拉尔（Pilar）的情况：一个世界领先的产业正充分利用其竞争优势，吸引外国投资并采用新技术，以巩固其作为全球领先出口国的地位。对于一张愁眉苦脸的脸来说，这或许是一个理想的良药。■
原住民语言
酷谈
亚松森
瓜拉尼语在巴拉圭变得流行起来
在瓜拉尼语课开始一小时后，亚松森国立大学的医学生们听得入迷。一组学生正在为“kangue opé”（骨折）提供“pytyvöpya'ehasývape”（急救）。另一组则担心“tajygue mboaty rehe”（神经系统）的问题可能需要“hembovo”（手术）。这门课程是巴拉圭未来医生的必修课，也是一系列推广原住民语言官方措施的一部分。似乎很少有人在意。“我喜欢说瓜拉尼语——它比西班牙语亲切得多，”一名本科生塔德奥·拉米雷斯（Tadeo Ramírez）说道。
这种热情反映了巴拉圭古老语言近期的复兴。尽管只有 2.3% 的人口认同自己是原住民，但约 90% 的人能听懂瓜拉尼语。根据最近的一项调查，近 40% 的人在家中每天与西班牙语一起使用瓜拉尼语。这是世界上双语率最高的国家之一，远超加拿大和爱尔兰等官方双语国家。其中大部分是“Yopará”，一种西班牙语和瓜拉尼语的混合语（该词在瓜拉尼语中意为“混合”）。“这是我们的通用语，”出租车司机米格尔·安赫尔（Miguel Angel）说。巴拉圭是美洲西班牙语国家中仅有的两个拥有官方第二语言的国家之一。
巴拉圭的精英阶层长期以来将瓜拉尼语视为农村方言而对其不屑一顾，嘲笑其鼻音和喉塞音。独裁者阿尔弗雷多·斯特罗斯纳（Alfredo Stroessner）在 20 世纪 50 年代和 60 年代不鼓励在学校中使用该语言。直到 1992 年民主回归，瓜拉尼语才与西班牙语一起被认可为巴拉圭的官方语言。2010 年的一项法律创建了双语课程和行政体系。如今，巴拉圭人可以选择用瓜拉尼语投票、结婚和接受审判。
这促进了一场文化热潮。今年 6 月，巴拉圭足球迷将他们在世界杯上战胜德国队归功于“瓜拉尼精神”。5 月，国家交响乐团首次用瓜拉尼语演奏贝多芬的《欢乐颂》。4 月，西班牙王后莱蒂齐亚（Queen Letizia）在祝贺亚松森西班牙文化中心周年庆的致辞开头写道：“¡Vý apavé ne arambotýre!”（“生日快乐！”）。年轻人也加入了进来。网红们在 TikTok 和 Instagram 上发布词汇教程；艺术家们在 Spotify 上发布 Yopará 风格的说唱和爵士乐。时髦的父母给孩子起名为 Yeruti（鸽子）和 Marangatu（正直）。
一些人担心，双语能力的增长源于瓜拉尼语使用者学习西班牙语，而非相反。纯瓜拉尼语使用者的比例从 2002 年的 28% 下降到 2022 年的 12%。语言政策秘书处（SPL，一个政府部门）的阿尔纳尔多·卡斯科（Arnaldo Casco）表示：“这是一个被取代的过程。”为此，该部门采取了应对措施，出版了大量瓜拉尼语内容，从技术词典到诗歌。SPL 负责人哈维尔·维韦罗斯（Javier Viveros）表示，瓜拉尼语必须覆盖社会互动的所有领域。
另一些人则心态轻松。纯西班牙语使用率处于停滞状态。瓜拉尼语已进入谷歌翻译和维基百科（“Vikipetá”）。SPL 的工作人员正在用它训练大语言模型。社会规范已经改变。“我们这些在孩童时期曾犹豫是否要说瓜拉尼语的人，如今感到绝对自由，”智库“梦想祖国研究所”（Patria Soñada Institute）的玛丽亚·德尔·卡门·希梅内斯（María del Carmen Giménez）说道。这足以说明问题。 ■
“人们正在行动”
38
中东与非洲
西岸
胆大妄为且暴力
库斯拉（QUSRA）
在政府的支持下，以色列定居者正在升级对巴勒斯坦人的袭击
在以色列占领的西岸，纳布卢斯（Nablus）南部的山丘讲述着一个渐进式吞并的故事。通往巴勒斯坦村庄的蜿蜒土路被岩石和土堆堵塞，切断了前往农田和橄榄林的通道。以色列定居者将葡萄藤整齐地种成行，竖起挂满旗帜的围栏，并带来了羊群。他们驾驶着以色列政府提供的新型多功能车，携带突击步枪在该地区巡逻。
他们的公开目标是与相邻的巴勒斯坦村庄制造“摩擦”。这往往会导致以色列国防军（IDF）介入，且通常站在定居者一边，随后是驱逐巴勒斯坦人和扩大以色列定居点前哨站（见次页地图）。自2023年10月7日哈马斯袭击以及以色列随后在加沙发动战争以来，以色列定居者在西岸的暴力袭击日益恶化，迫使越来越多的巴勒斯坦人逃离（见次页图表）。随着以色列准备在10月27日进行投票，总理本雅明·内塔尼亚胡（Binyamin Netanyahu）未能约束定居者，使得巴勒斯坦武装分子采取反暴力行动的可能性日益增加。
过去一周的事件表明，定居者获得了多少自由裁量权。特尔塔皮奥特（Tel Talpiot）是希洛（Shiloh）定居点集团中最新且最北端的前哨站，于今年早些时候建立在俯瞰三个巴勒斯坦村庄的山丘上。8月9日，一群来自特尔塔皮奥特的人占领了库斯拉村郊区房屋的院子，搭建了一个临时避难所，恐吓当地家庭并迫使他们躲在屋内。被派去“防止摩擦”的以色列国防军士兵没有将他们驱逐，反而加入定居者的祈祷。
这种在西岸很常见的战术，通常能成功迫使巴勒斯坦村民逃离家园，从而让定居者接管周围的土地。但这次情况有所不同，原因有几点。库斯拉的居民此前经历过类似的袭击，他们迅速公开了此次事件并吸引了国际关注。被围困的房屋之一属于一名美国公民。通常支持定居者的美国大使迈克·哈卡比（Mike Huckabee）在社交媒体上发帖，称其大使馆“请求移除这些以色列恐怖分子”。
在围困进行到第三天时，以色列国防军才迟迟地用一个纪律更严明的步兵营替换了在库斯拉的部队。该部队试图移除定居者并撤销特尔塔皮奥特前哨站。定居者带来了援军并继续留在该地区。“这是一起令人尴尬的事件，”一名军官说，“定居者强迫我们玩猫鼠游戏，而我们输了。”
以色列国防军宣布这些房屋处于“封闭军事区”。它强迫这些家庭继续被封锁在室内，并阻止活动人士前来救援以及媒体访问。“我们被围困了，军队不允许我们离开房子，”居民之一库赛·阿布·里迪（Qusai abu Ridi）说道。但他坚决表示，他们将继续保护自己的财产。
→ 本节其他内容
39 巩固伊朗的权力
40 南非破产的市政当局
41 索马里兰总统的大计划
中东与非洲 39
尽管如此，库斯拉仍是一个罕见的案例，巴勒斯坦居民成功保住了自己的家园。然而，这一事件也是过去一个月新一轮升级的一部分，因为定居者试图在选举前尽可能多地抢占土地，因为选举可能会产生一个对其事业支持程度较低的政府。7月24日，一群武装定居者以所谓的“徒步旅行”为名抵达特尔村（Tel）。最终导致四名巴勒斯坦人和两名以色列人死亡。作为回应，以色列国防军部署了更多部队并封锁了巴勒斯坦城镇，包括西岸最大的城镇之一纳布卢斯。
以色列国防军（IDF）声称是在维持秩序，但其在任期间，针对巴勒斯坦人的袭击大幅增加。以色列安全机构辛贝特（Shin Bet）表示，2026年上半年，西岸发生了660起“民族主义”袭击，高于去年同期的405起。自2023年10月7日以来，已有1,100多名巴勒斯坦人在与以色列国防军及定居者的冲突中丧生。
库斯拉（Qusra）发生的事件凸显了以色列国防军不愿阻止定居者试图驱逐巴勒斯坦人的倾向。包括现役和退役在内的高级军官将此归咎于纪律崩溃以及以色列国防军内部存在定居者。但最终，他们认为政府应承担责任。
尽管以色列国防军的高层公开谴责了库斯拉的袭击事件，但总理本雅明·内塔尼亚胡（Binyamin Netanyahu）仍保持沉默。8月14日，国防部长以色列·卡茨（Israel Katz）表示，从现在起，在西岸对定居者执行法律将由以色列警方负责，而非以色列国防军。由于警方在西岸缺乏人员，且受极右翼国家安全部长伊塔马尔·本-格维尔（Itamar Ben-Gvir）控制（他本人也是一名定居者），他们不太可能在该地区执行法律。
以色列在西岸部署了庞大的部队和情报资源，以保护50万名定居者。理论上，以色列国防军也有义务保护巴勒斯坦人。但只有在政治上级明确指示的情况下，士兵们才会这样做。在内塔尼亚胡先生领导下，这种情况不太可能发生，因为他需要亲定居者的极右翼政党的支持，才有机会在选举后维持权力。
在此之前，西岸的情况可能会进一步恶化。安全官员警告称，定居者正试图挑起另一次“大起义”（intifada，即巴勒斯坦人起义），这将迫使以色列国防军进入巴勒斯坦城市。这将给西岸饱受困扰的巴勒斯坦人带来更多的暴力和苦难。即使内塔尼亚胡先生的政党在选举中落败，他的继任者可能也必须处理由他一手造成的危机。■
伊朗与战争
将领的崛起
军政府正在巩固权力，给国内和国外带来严重后果
在过去的几天里，伊朗首都德黑兰出现了新的广告牌。这些广告牌描绘了自前最高领袖阿里·哈梅内伊在2月28日的一次以色列空袭中被杀后，接管伊朗权力的七个人。最引人注目的是广告牌上没有显示的内容。新最高领袖穆吉塔巴·哈梅内伊没有出现，他在被任命后一直没有在公开场合露面。也没有出现任何传统的宗教符号，例如伊玛目阿里和侯赛因的黑红旗帜。七个人中只有一个人戴着头巾。
这些广告牌是最新迹象，表明一批来自伊斯兰革命卫队（IRGC）的新领导层正在伊朗巩固权力。他们久经沙场且结果导向，专注于安全与镇压，正将伊斯兰共和国的神权基础转变为军事独裁。尽管他们可能并未完全放弃在与美国的冲突中采取务实方法，但军事政权通常从持续的战争威胁中获取合法性。这表明美国及其盟友应当预料到，新领导层将比前任更加不妥协。
在广告牌上，身着西装并被标为自称的“国家安全与国防之顶端”的是穆赫辛·礼萨伊。他是新任国家安全顾问，曾是总统候选人，也是伊斯兰革命卫队任职时间最长的负责人，曾领导该组织度过了1980-88年的两伊战争及其随后的重建期。五个人穿着军装，其中包括监督伊朗对海湾航运袭击的海军首脑。在他旁边的是伊斯兰革命卫队的总司令艾哈迈德·瓦希迪。除了在四年前试图镇压脱去面纱的女性运动中发挥关键作用外，他还创建了伊斯兰革命卫队的对外分支——圣城旅，该部门培育其地区代理人网络。唯一戴头巾的侯赛因·塔耶布曾负责管理伊斯兰革命卫队的情报机构，现在领导其准军事志愿者部队 husij，该部队被委以将抗议者赶出街道的任务。
一些资深的伊朗观察人士将这些人的接管视为强硬派对以议长穆罕默德·巴格尔·加利巴夫为首的务实派的胜利，强硬派寻求将美国驱逐出该地区。加利巴夫在谈判一份于6月17日签署的与美国的谅解备忘录（MOU）中发挥了关键作用。该谅解备忘录承诺，如果伊朗放弃追求核弹并结束对霍尔木兹海峡的封锁，美国将结束近五十年的反对，保护伊朗的地区代理人，解除制裁，遣返被冻结的石油收入，并提供$300bn用于启动战后重建。
在文件签署后的几天内，战斗再次爆发。霍尔木兹海峡在很大程度上仍对航运关闭。8月17日，谅解备忘录为达成更永久和平协议设定的截止日期到期；一天后，伊朗似乎向美国的波斯湾盟友阿拉伯联合酋长国发射了导弹，后者则以贸易禁运作为回应。一些观察家担心，强硬派认为自己击退了美国和以色列攻击的感觉，使其重新坚信武力的可行性。“他们正在浪费（通过谅解备忘录获得的）胜利，”德黑兰的一位学者愤怒地表示。
在国内，新政权看起来同样严酷。道德警察在德黑兰重新强势出现，突击检查并关闭被认为在执行面纱规定方面过于宽松的咖啡馆。对反政权抗议者的公开处决已经恢复。本月，一名拍摄伊朗社会不安状况的摄影师与他的母亲和妹妹在西北部的一座城市附近被枪杀。 ▶
被驱逐
西岸，被定居者暴力和准入限制驱逐的巴勒斯坦人
40 中东与非洲
然而，其他内部人士和观察家则看到了更深层的现实政治算计。将军们批准了加利巴夫（Mr Ghalibaf）的谅解备忘录（MOU）。但据称，他们认为如果没有压力，该备忘录根本没有实施的可能。尽管有了谅解备忘录，美国还是在海峡建立了自己的封锁线，并与以色列协作，就黎巴嫩真主党（Hizbullah）这一重要代理人的解除武装进行了谈判。伊朗甚至无法确保协议中包含的被冻结资金的转移。
新政权可能渴望达成一项能带来资金的协议。在两伊战争后，雷扎伊（Mr Rezaei）领导了伊朗的重建。作为总统候选人，他呼吁吸引外资并达成核协议。“他们说孤立会杀死我们，”另一位与新领导层有联系的伊朗学者表示。此外，一些人认为，他们并不在意伊斯兰规范的削弱。对咖啡馆的打击可能仅仅是为了安抚其支持基础。有些人将他们比作沙特王储穆罕默德·本·萨勒曼（Muhammad bin Salman），后者在推行社会自由化以替代政治改革。
然而，所有观察家都一致认为，一种全新的领导风格已经出现。近四十年来，哈梅内伊（Mr Khamenei）在平衡其领地的竞争利益，在改革派和强硬派之间交替。相比之下，新指挥官们更倾向于采取行动。伊朗人和世界或许有一天会后悔哈梅内伊离开舞台的那一天。■
南非小镇
微缩版的国家俘获
姆波法纳（MPOFANA）
失效的市政当局阻碍增长，但部分地区表明改变仍有可能
“嘿，这个社区里糟糕的事情太多了，”内利斯瓦·拉德贝（Neliswa Radebe）叹息道。这位19岁的少女与另外13个人一起，住在南非东部姆波法纳（Mpofana）市政区一条高速公路上方的一栋破旧小房子里。她列举了各种问题：自来水罕见；垃圾清理零星；她那卧床不起的养母所用的床垫上还一直有水在滴漏。室外情况也好不到哪里去。道路上的坑洞比沥青还多；最近一名居民在汽车驶入坑洞失控时丧生。这里没有操场、足球场、网球场或青年中心。在高速公路的另一侧，非法占用者已经占领了姆波法纳破败的市政厅。
姆波法纳只是一个更大问题的一部分。南非大多数市政当局——这些旨在提供基本服务的基层政府单位——现在一片混乱。它们的失效让生活变得更糟，尤其是对穷人而言，并阻碍了更广泛的经济复苏。这正在削弱曾经强大的非洲国民大会（ANC）。如果非洲国民大会在11月4日的地方选举中遭遇预期的惨败，可能会撼动南非的政治格局。
在种族隔离制度下，地方政府反映了该国的系统性种族主义。白人区域是商业中心和富裕郊区的所在地；商业税和财产税被用于建设在加利福尼亚州也毫不违和的基础设施（例如耗水量巨大的游泳池）。黑人区域则被边缘化，其基础设施看起来更像非洲其他地方，而非这个工业核心地带（自来水在那儿是一种奢侈品）。
纳尔逊·曼德拉在1994年当选后，其政府废除了种族隔离的地方当局，创建了表面上统一的市政当局。如今，这些单位的规模从商业之都约翰内斯堡，到像姆波法纳这样由一个小镇和农村地区组成的单位不等。所有单位都应通过财产税、商业税以及水、电和垃圾清理费来自筹资金。最贫困的地区还会获得中央政府的拨款，以填补收入与支出之间的缺口。
实际上，许多单位无法收支平衡。在截至2025 45%的市政当局通过了缺乏资金支持的预算的财政年度中。包括姆波法纳在内的约30%的单位在过去四年中每年如此。上个月，财政部长埃诺克·戈东格瓦纳（Enoch Godongwana）表示，他将扣留对违规最严重地区的拨款，以建立财政纪律。
公共部门工会认为，拨款的增加滞后于服务提供成本的上升。但更大的问题是资金被浪费或窃取。在6月份审计长发布的最新报告中，仅有15%的市政当局通过了干净的财务审计。自2021年以来，“浪费性”、“违规”或“未经授权”的支出总计达2870亿兰特（180亿美元），超过了南非同期在国防上的支出。在一个极具象征意义的市政贪腐疑似案件中，包括当地官员在内的十个人涉嫌盗用指定用于东开普省阿马托莱（Amathole）——该国最贫困地区之一——建设VIP（通风改良式坑厕）厕所的资金。
各市政当局合计欠国家电力公司（Eskom）超过1100亿兰特，欠大宗水供应商300亿兰特。戈东格瓦纳先生认为，这些债务威胁到经济的“财务可持续性”。他决定暂停拨款的部分原因是为了用这笔钱支付这些账单。（7月31日，他开始在附加新条件的情况下恢复拨款。）
仿佛是在效仿政客，居民们通常也不付钱。市政当局采购的水资源中近一半被“丢失”了：它们要么从破裂的管道中流走，要么从未被支付费用。在姆波法纳（Mpofana），60% 的电力实际上是被盗用的。“坦白说，我们不付钱，”拉德贝（Radebe）女士承认道，她回头看了一眼屋外通往非法连接处的生锈电缆卷。
然而，一些市政当局运作良好。姆波法纳以南的姆恩格尼（Umngeni）在 2021 年抛弃了非洲国民大会（ANC），选举出了一位来自第二大党民主联盟（DA）的市长。克里斯·帕帕斯（Chris Pappas）被认为实现了彻底的改革。今年 7 月，非政府组织 FW 德克勒克基金会（FW de Klerk Foundation）在 50 个样本中将姆恩格尼评为管理最好的第七个市政当局，排名高于开普敦。
如今，姆恩格尼拥有充足的预算资金，并获得了干净的审计报告。表现不佳的官僚已被离职或解雇。帕帕斯先生打击了虚高的采购计划。节省下来的资金帮助其 ►
中东与非洲 41
► 安装了 3,000 盏新路灯，填补了路面坑洞，并扩大了一项针对极贫困人口的计划，为少量的水和电提供补贴。曾经肮脏的镇中心现在变得干净了。新公司纷纷入驻。在人口 100,000 人的地区，净新增就业岗位超过 2,000 个。
在姆恩格尼的一个贫困地区赞泽内村（Zanzene Village），居民们似乎很欣赏这些变化。“一切都很好：道路、水，我们甚至支付电费，”诺姆丰多·布特莱齐（Nomfundo Buthelezi）说道，她解释说只要钱花在刀刃上，她就不介意。园丁穆萨·姆丘努（Musa Mchunu）注意到这里有了新的足球场和青年中心。“最重要的是，”他补充道，“市政当局现在有回应了——我们不再收到‘蓝色勾勾’了”，这是指 WhatsApp 消息已被阅读但被忽略的状态。
民主联盟（DA）希望良好的治理记录能增加其得票率，包括在黑人选民中，而黑人选民在历史上一直选择非洲国民大会（ANC）或其分支机构。其目标之一是约翰内斯堡，最近的一项民调显示，民主联盟的支持率达到 42%，高于 2021 年地方选举时的 26%。而将这座城市治理得一塌糊涂的非洲国民大会（ANC）支持率为 18%，仅为 2021 年赢得的 34% 的一半左右。
这种高支持率不会在南非各地都出现。在最近的民调中，平均而言，非洲国民大会（ANC）仍领先于民主联盟（DA）（34% 对 27%）。种族在很大程度上仍能预测选民是否选择民主联盟（DA），尽管该党仅管理着 12% 的市政当局，但却占据了绝大多数干净的审计报告。
然而，在白人统治结束 32 年后，当大多数南非人对民主感到绝望时，证明政治选择可以带来改变至关重要。曾是非洲国民大会（ANC）成员的姆丘努先生说，他上次没有投票给帕帕斯先生，但今年 11 月可能会这么做。“我告诉我的儿子们，种族隔离已经结束了。重要的不是种族，而是改变这个国家。” ■
索马里兰
伊罗的赌注
内罗毕
总统希望帮助世界保障红海安全
比迪拉曼·穆罕默德·阿卜杜拉（Bidirahman Mohamed Abdullah）是一个雄心勃勃的人。2024年12月，这位70岁的前外交官（如图），人们更习惯称他为伊罗（Irro），宣誓就任索马里兰（Somaliland）总统。索马里兰是一个于1991年宣布脱离索马里独立的分离共和国。去年，在索马里兰获得以色列正式承认（以色列是第一个这样做国家）时，他正是这个准国家的领导人。“世界已经忽视了索马里兰35年，”伊罗告诉《经济学人》。他希望外界能更密切地关注这里。
伊罗的推介重点在于红海。近期，这条贸易航线陷入混乱，航运受到胡塞武装的威胁，这是一个由伊朗支持、统治也门大部分地区的民兵组织。本月在访问肯尼亚首都内罗毕期间，伊罗表示，索马里兰可以通过在其海岸线向全球大国提供军事基地，来帮助保障这条战略水道的安全。一旦贸易能够再次畅通无阻，他希望索马里兰的主要港口柏培拉（Berbera）能成为进入非洲之角市场的主要门户。“我们的信息是，我们在安全和贸易方面是一个负责任的合作伙伴，”伊罗说道。
他暗示索马里兰实际上可以成为一个新的吉布提。那个微小的港口国家通过将其海岸线的一部分租给外国军队，每年可赚取约1.25亿美元。美国、中国和法国（以及其他国家）在那里设有基地。此外，它还从内陆国家埃塞俄比亚支付的年度港口费中赚取该金额的数倍，因为埃塞俄比亚的大部分国际贸易依赖于吉布提。
伊罗想要分得吉布提这两块蛋糕的一杯羹。他正试图说服埃塞俄比亚使用柏培拉作为其主要出海口，阿拉伯联合酋长国（UAE）已在那里建造了一个新港口。他还希望吸引美国在那里建立军事基地。如果美国选择重启针对胡塞武装的战争（美国去年曾对其进行密集轰炸），这将为美国提供更多选择。人们普遍认为，在12月以色列承认索马里兰的协议中，就包含了使用柏培拉周边军事设施的内容。一些报道称，美国和以色列可能已经在柏培拉部署军事力量，并得到了在当地运营现有基地的阿联酋的秘密支持。
伊罗否认了此类报道。“目前还没有以色列或美国的部署，”他说。但阿联酋在庞特兰（Puntland）的博萨索（Bosaso）港口设有另一个基地，庞特兰是索马里邻近的一个半自治省，有未经证实的报道称以色列也潜伏其中。随着红海军事地产的竞争升温，伊罗表示，如果以色列和美国“感兴趣”，他愿意在柏培拉为双方提供基地。下个月，伊罗将访问华盛顿，他希望在那里与唐纳德·特朗普进行一次私下会谈。如果成行，他将成为首位直接会见美国同行 的索马里兰总统。“我们期待在未来的某个时间获得美国的承认，”伊罗说道，并补充称，索马里兰预计在“年底前”能获得至少另一个国家的承认。
伊罗的野心引起了地区性的不安。吉布提对索马里兰试图插手的意愿感到恼火。伊罗暗示，他的北方邻国正试图“破坏”其国家的稳定，据推测是通过支持边境上的敌对民兵。与索马里的关系则更加紧张。伊罗声称，将索马里兰视为叛离省份的索马里，计划在南部边境动员军队。“这不是和平的方案，”他补充道。
以色列与索马里的赞助国土耳其之间日益增长的敌意也毫无益处。土耳其长期以来一直试图在索马里和索马里兰之间进行调解，但安卡拉和伊斯坦布尔的多轮谈判均以失败告终。“下次他们可以在特拉维夫谈，”伊罗轻笑道，他似乎并不在意发生代理人战争的可能性。然而，美国很可能不愿通过承认索马里兰来搅动局势。
与此同时，胡塞武装已威胁要攻击该国境内的任何以色列军事设施。美国在该地的基地也可能成为目标。这将摧毁索马里兰长期以来为争取国际承认而建立的和平与稳定声誉。但伊罗表示：“世界上没有免费的午餐。”他似乎愿意付出代价让自己的国家在地图上占据一席之地——问题是，代价会有多高？■
42
亚洲
日本
被冻结的一代
东京
日本的“就业冰河期”一代再次陷入困境
近二十年来，50岁的葎越敦（Torigoe Atsushi）一直在东京一家时尚公司的仓库工作。他的收入增长缓慢，且在雇主削减奖金并调整薪酬方案后近期有所下降；他现在的年薪约为360万日元（22,600加元）。随后，他看到了公司的招聘广告，向新入职者提供更高的薪水。“年轻人能拿到更高的薪水是很棒的事，”他说，“但为什么不给我涨薪呢？”
葎越先生属于日本的“冰河期”一代：即在1993年至2004年之间进入劳动力市场的工人，当时日本的经济泡沫破裂，许多公司停止招聘。冰河期一代的人经常忍受多年不稳定的工作和低薪。如今他们处于40多岁和50多岁（因此属于X世代），他们再次处于劣势。日本正走出数十年的通货紧缩，物价上涨，企业正在发放一代人以来最大幅度的加薪。但冰河期一代并未从中受益。更糟糕的情况可能会在这一群体——婴儿潮一代的孩子，也是日本人口金字塔中的最后一个高峰（见下页图表）——步入退休之年时出现。
冰河期一代的命运与紧接其前的一代形成了鲜明对比。在20世纪80年代的泡沫时期，公司疯狂招聘。大学生发现自己的信箱里塞满了录取通知书，或者受到渴望招揽他们的雇主的盛情款待，享用高级餐点。随着90年代初泡沫破裂，一切都改变了。日本僵化的就业制度使得公司难以裁减现有员工，因此他们转而削减招聘。职位空缺与大学申请者的比例从峰值时的近3比1下降到2000年的1比1以下。1999年（即亚洲金融危机后不久）毕业的永井利久（Nagai Toshihisa）放弃了在证券公司工作的希望。“找工作感觉简直太愚蠢了，”他回忆道。他那一辈的许多人发送了数十份甚至数百份申请，但均无果而终。冰河期这一群体的规模如此之大，加剧了供应过剩，这让情况变得更加糟糕。
许多未能获得正式工作的毕业生最终从事临时或合同工作。日本的劳动力市场在正式员工和非正式员工之间有着严格的区别：前者享有强大的法律保护和基于资历的薪资增长，而后者通常从事类似的工作但薪水低得多，且在经济低迷时首先被解雇。这类工作过去通常由补贴丈夫薪水的家庭主妇担任。东京大学的近藤绫子（Kondo Ayako）表示，越来越多需要自食其力的年轻人接手了这些工作。“自由职业者”（freeters）——在临时工作之间漂泊的年轻人——这个词变得流行起来，同样流行的还有“寄生单身族”，他们成年后仍依赖父母生活。人们对“蛰居族”（hikikomori，社会隐蔽者）的担忧也在增加。
时间以及严重的劳动力短缺修复了部分损害。日本综合研究所的下田勇介（Shimoda Yusuke）表示，冰河期一代在正式就业中的比例已基本追上年长群体。但他们的收入并未追上。那些长期从事非正式工作的人，即使在转为正式岗位后，往往也无法完全追平。而那些处于正式岗位的人则面临另一个问题：公司的高层被泡沫时代的员工占据，延迟了后继者的晋升。日本较低的劳动力流动性意味着这个不幸的群体无法轻易在其他地方寻求更好的薪酬。
→ 本节其他内容
43 印度大象的衰落
44 唐纳德·特朗普扰乱亚洲盟友
44 哈萨克斯坦的装饰性选举
45 阿育王：新的国父
亚洲 43
日本经济正在发生变化。在经历了三十年的停滞之后，随着通货膨胀回归，工资开始上涨，跳槽也变得更加普遍。然而，这些改善主要使年轻人受益。劳动力短缺的公司在激烈竞争新鲜血液，并支付更高薪资以吸引和留住他们。中年员工跳槽的可能性要低得多，这使得公司几乎没有动力提高他们的工资。第一生命研究中心（Dai-ichi Life Research Institute）经济学家熊野英夫（Kumano Hideo）表示，2025年，20多岁和30多岁的大学学历员工的名义工资比五年前高出10-16%。与此同时，50岁出头的人群的名义工资则下降了1.3%。48岁的安倍胜也（Abe Katsuya）表示，发给年轻人的大幅加薪让他心情“复杂”。他是“猛犸象俱乐部”的一员，这是一个为“冰河期一代”设立的支持小组，他们将自己比作在艰苦时期生存下来的长毛猛犸象。
这一切导致冰河期一代的终身收入较低（因此储蓄也很少）。这意味着随着他们步入老年，一场危机即将到来。日本的养老金制度将给付额与终身收入挂钩，因此数十年的薪资受压将转化为低额养老金。“那些恰好在经济低迷时期成年的人，很可能会在余生中一直背负这一劣势，”熊野先生感叹道。让他们延长工作时间（正如越来越多的人已经在做的那样）是一个解决方案，但这无法解决所有问题。近藤女士警告说，许多人最终可能会依赖社会福利，而福利制度并非旨在长期支持老年人。这将耗资巨大；例如，国家全额支付福利领取者的医疗费用。而且，由于冰河期一代之后的世代规模更小，更少的劳动者将不得不供养更多的养老金领取者。
冰河期一代的困境长期以来一直受到政府的关注。
几十年来，当局制定了各种计划帮助他们寻找工作，并鼓励那些隐居的人与外界建立联系。然而，随着冰河期一代年龄的增长，他们面临着新的问题。除了微薄的养老金和储蓄，住房等问题也随之而来；与年长的日本人相比，这一群体拥有住房的可能性较低，且日本的租赁市场对老年租客并不友好。政府正在迟缓地做出调整：今年4月，政府出台了一项针对冰河期一代的三年计划，涵盖了从资产积累到住房的各项措施。但仍需要进行更深入的思考。■
印度
无处栖身
孟买
印度象正被挤出其栖息地
在古印度国家治理手册《政事论》（Arthashastra）中，大象是极其珍贵的资产。国王们用它们来缔结盟约，将军们依靠它们摧毁堡垒的大门。如今，象头神迦尼萨（Ganesha）仍被信众尊为排除障碍之神；而孩子们在学习字母表时，依然学习“E代表大象（Elephant）”。
但崇拜并未转化为保护。印度野生动物研究所去年发布的一项调查显示，印度象数量为 22,446 头，比 2017. 年减少了 18%。该报告的主作者卡马尔·库雷希（Qamar Qureshi）表示，部分邦的数量下降情况令人“震惊”。地理因素是问题的一部分。与非洲不同，在人口稠密的印度，村庄散布在森林之中。大象经常被限制在森林碎片之间的狭窄走廊中。印度野生动物信托基金会的维韦克·梅农（Vivek Menon）表示，2010 年他的团队确定了 88 条此类路线。如今，由于栖息地萎缩，此类路线已增加到约 150, 条。
随着基础设施切断森林，大象误入农场。矿区和爆破现场也阻挡了它们的去路。在西孟加拉邦，尽管最高法院已下令禁止，但仍会出现所谓的“hulla”（噪音）派对，人群通过投掷火球和燃放鞭炮来驱赶这些动物。电网有时会阻挡大象的路径。但生态学家拉曼·苏库马尔（Raman Sukumar）说，大象是出色的物理学家。它们利用导电性较差的象牙来破坏电网，或者投掷原木使其短路。环保主义者普雷娜·宾德拉（Prerna Bindra）表示，引导大象返回安全地带的森林护卫员和工作人员“完全缺乏装备，准备工作极其不足”。大多数人甚至缺乏手电筒等基本设备。
但并非一切都如此绝望。自 20 世纪 80 年代以来，为了获取象牙而进行的盗猎活动有所下降。邦林业部门与国家海关之间的协调得到了改善。如今，印度成立了野生动物犯罪控制局，各邦也制定了反盗猎计划。上个月，印度最高法院维持原判，要求移除泰米尔纳德邦一条大象走廊内的商业活动。
大象扮演着景观建筑师的角色。它们传播种子，塑造森林并肥沃土壤。它们通过挖掘使湿地保持活力。《政事论》曾建议印度的统治者保护这些壮丽生物栖息的森林。如今的掌权者应当留意这一点。
永不忘记
婴儿潮高峰
日本，按年龄组划分的人口，2026 年，占总人口百分比 总人口=122.4m
来源：联合国人口司
44 亚洲
美国与韩国
联盟管理不善
唐纳德·特朗普突然缩减与韩国的演习
“乙支自由之盾”（ULCHI FREEDOM SHIELD）是一项年度军事演习，旨在让数千名美国和韩国军队共同训练以应对来自朝鲜的威胁。今年的演习于8月17日开始，盟友们原计划重点应对朝鲜士兵在参与俄罗斯对乌克兰战争期间获得的新能力，例如无人机战争。但在演习开始前一夜，总统唐纳德·特朗普命令其战争部“大幅缩减”（缩减六天）此次演习，理由是演习成本、韩国拒绝支持他在伊朗的战争，以及他与朝鲜独裁者金正恩之间“非常良好的关系”。
特朗普先生大概是希望诱使金先生重启两人在其首个任期内开始的外交接触。他也可能想向韩国施压，促使其加快承诺对美国的投资。而且毫无疑问，他希望转移人们对他陷入伊朗泥潭的注意力。
但他的突然决定向亚洲的朋友和敌人都发出了令人担忧的信号。尽管可能性不大，但两国领导人再次举行会谈的前景令美国盟友的首都感到焦虑。（朝鲜称特朗普先生的举动是空洞的姿态，尽管特朗普先生在8月19日坚持认为他将在今年晚些时候会见金先生。）尤其是日本，担心特朗普先生可能会达成一项损害其安全的糟糕协议。此举加剧了自他重返办公室以来日益增长的被美国抛弃的恐惧。对手们可能会得出结论，认为美国对盟友的承诺是空洞的。
这已经不是特朗普先生第一次将与韩国的演习作为筹码了。2018年，当他在新加坡会见金先生（这是美国和朝鲜领导人之间的首次峰会）时，特朗普先生表示美国将停止即将与韩国举行的演习。此举令他自己的顾问们感到震惊。但这种单方面的让步并未带来突破；次年两人在河内的随后会晤在激烈的争吵中破裂。
韩国总统李在明与特朗普先生一样，都希望恢复与朝鲜的接触。8月19日至20日，他的政府在首尔接待了中国外交部长王毅，部分目的是讨论向金先生伸出橄榄枝的持续努力。然而，李总统也承认维持威慑的重要性，而特朗普先生削弱了这种威慑。
特朗普先生长期以来一直对美国与韩国联盟的价值持怀疑态度。在首个任期内，他试图将美国军队从该国撤出，但被国会阻挠。自重返办公室以来，他的政府对该联盟进行了悄然但具有深远影响的改变。例如，美国已采取行动，将其在韩国的28,500名军队的任务重心从传统的半岛关注点转向威慑中国。他还利用关税迫使韩国承诺在美国投资$350。特朗普先生毫不犹豫地将安全联盟的未来与投资协议的流动挂钩。他决定缩减演习之时，正值韩国产业通商资源部长访问华盛顿，讨论那些陷入停滞的计划的实施问题。
即便在最新的声明发布之前，特朗普先生对该联盟采取的交易主义方式已经削弱了对美国的信任。首尔智库东亚研究所（East Asia Institute）的一项民意调查发现，信任美国的韩国人比例从 2022 年的 85% 骤降至今年的 46%。对被抛弃的恐惧也推动了关于韩国获取核武器的讨论。李先生表示他反对这一路线。但他一直寻求开发用于民用核电工业的核燃料浓缩和再处理能力，如果韩国认为有必要，这将使制造核弹变得更加容易。在特朗普先生发表声明后，他重申了推进这些计划的意图。
整个地区也出现了类似的动态。根据皮尤研究中心（Pew Research Centre）的数据，在澳大利亚、日本和新加坡，人们对美国可靠性的认知在 2022 年至 2026 年间大幅下降。
在那儿，但能维持多久？
日本也一直在寻求加强自身的防御能力；对更大自主权的追求正在影响预计于今年晚些时候进行的国家安全战略修订。
然而，面对一个更富有且日益强势的中国以及拥有核武器的朝鲜的威胁，美国的亚洲盟友很难想象没有美国的安全保障。尽管信任度下降，但仍有超过 60% 的韩国人认为与美国的联盟是确保其安全的最佳方式。东亚研究所所长孙律（Sohn Yul）写道，这样的结果反映了一种“矛盾的情况，即（韩国）不信任美国，但又不得不依赖美国”。这种困境在整个亚洲都普遍存在。■
哈萨克斯坦
重启还是镇压？
阿拉木图
一个威权主义前苏联共和国中形式主义的议会选举
“为更新而投票，”巨大的广告牌在哈萨克斯坦准备于8月23日举行议会选举之际发出号召。总统卡西姆-乔马尔特·托卡耶夫（Kassym-Jomart Tokayev）称此次选举是建设“新哈萨克斯坦”的转折点——这是对一场政治重启的简写，这场重启始于2022年一场要求政治变革的公民动 unrest，该事件导致至少238人死亡。
乍一看，这次选举确实有很多新意。在一部新宪法下举行，它将产生一个在结构和名称上都有所改变的议会，且必然由某个政党主导。表面上看，这就是正在实施的改革后的“哈萨克斯坦”。但只要深入观察，旧哈萨克斯坦的政治外壳便会显现——这与努苏尔坦·纳扎尔巴耶夫（Nursultan Nazarbayev）相关，他是托卡耶夫先生长期担任其继任者的前任，也是其名声扫地的政治赞助人。预计将赢得议会绝对多数席位的新成立的“正义党”（Justice party），实际上是主导了数十年的旧执政党的翻新版本。而且，议会将依然是一个橡皮图章机构，因为没有任何真正的反对党被允许参选。
托卡耶夫先生最近表示，多元主义在哈萨克斯坦蓬勃发展——而就在几天前，反对派领导人阿曼盖尔迪·贾欣（Amangeldy Dzhakhin）刚因被指控为极端主义的伪造罪名而被判处七年监禁。他的前任也遭遇了同样的命运，而且被针对的不止是政治人物。记者和社会媒体名人也因各种理由被监禁，或被施以 ▶
亚洲 45
► 旅行禁令，并被以看似牵强的指控禁止从事新闻工作。在一个明显的政治案件中，一条嘲讽总统的讽刺帖子最近导致一名YouTuber锒铛入狱。
此次选举是在通过一部宪法之后举行的，托卡耶夫先生将此描述为哈萨克斯坦“摆脱了超级总统制的统治形式”。这是在指他的前任，后者在2019年辞职前垄断了权力三十年。纳扎尔巴耶夫先生指定的继任者就是托卡耶夫先生。两人共同执政直到2022年的动荡，这次动荡使纳扎尔巴耶夫先生被边缘化，而托卡耶夫先生则急于与他的前导师撇清关系。他在2022年削减了一些总统权力，但在通过全民公投（背景是批评者被逮捕）通过的宪法中，又重新夺回了一些权力。
一项重要的改革仍然有效。2022年，托卡耶夫先生引入了总统单任期限制。他承诺将遵守这一规定，但宪法为他提供了一个漏洞。宪法法院最近裁定，他拥有担任“第一”任期的权利，因此他可以再次参选——这是中亚独裁者以及俄罗斯的弗拉基米尔·普京和土耳其的雷杰普·塔伊普·埃尔多安所青睐的、经过验证的权力延续策略。因此，他在2029年结束当前的（第二）任期，将成为衡量托卡耶夫先生对其宣称的消除哈萨克斯坦君主制统治承诺的试金石。 ■
ASHOKA
重塑国父
第二次国家建设反而更加困难
每个国家都有国歌。有些国家，如新西兰，拥有两首。但只有印度同时拥有一首官方歌曲（national song）。今年，在独立日庆祝活动中，“万德·马塔拉姆”（Vande Mataram）这首对“印度之母”的激昂赞歌首次出现，在8月15日纳伦德拉·莫迪抵达发表年度演讲时为其迎接。一架空军直升机飞过，悬挂着一条纪念该歌曲创作150周年的横幅。这标志着莫迪领导的印度人民党（BJP）多年来致力于提升该歌曲地位的努力达到了顶峰。就在几天前，干扰该歌曲的演唱已被定为刑事犯罪。
莫迪在演讲中谈到了印度进入独立第80年的成就，并展望了其辉煌的百年庆典。此类声明承认印度在1947年从英国获得自由。然而，他的印度人民党认为，这个国家真正获得自由的时间要短得多。这种情绪最清晰的表达来自宝莱坞女演员、自2024年起成为印度人民党议员的康加娜·拉瑙特（Kangana Ranaut），她在几年前的一次采访中宣称，印度直到2014年莫迪当选后才真正实现独立。
这句话应当被严肃对待，而非字面理解。在印度人民党领导人的眼中，他们继承的印度过于软弱，过于英语化，且对少数群体过于宽容。正如华盛顿智库卡内基国际和平基金会的米兰·瓦什纳夫（Milan Vaishnav）在2024年的一篇论文中所述，他们的计划是建立一个“由国家古老印度教遗产定义的文明国家”。他将这种新的政治秩序描述为印度的“第二共和国”。
一个新的共和国需要一个新的身份。在1947年独立前后，制宪议会决定了印度的国家象征应当是什么。随着时间的推移，市政机构将英国国王、女王及各类公爵的雕像移走，安置在遥远的公园里（或者，令人快慰地，安置在动物园里）。州政府重新命名了城市和机构。
自2014年以来，印度人民党一直在执行类似的使命。其显而易见的目的就是推广其“印度教优先”的愿景。这首国家歌曲此前未能获得国歌地位，部分原因是其对印度教神灵的祈求被认为排斥穆斯林。而现在，这已不再是问题。近年来，政府将其旗舰火车服务、公共广播公司的标志，以及上个月男子国家曲棍球队的制服，全部改为藏红花色——这是一种与印度教及该党本身相关的颜色。去年，它彻底改革了一项以圣雄甘地命名的就业计划，并将其重新命名为 VB-G RAM G，这是一个经过反向工程推演出的艰涩缩写，旨在将罗摩神（Lord Ram）的名字包含在内。
尽管如此，印度人民党计划中非宗教的部分，反而更像是一个年轻国家的国家建设。除了国家象征和对未来的共同愿景，一个新的共和国必须有一个起源故事来赋予其意义。但对印度人民党来说不幸的是，独立运动及其英雄群像已经实现了这一目的。这使得他们只能采取两种策略。
第一种是重新定义早已故去的人物。印度首任总理贾瓦哈拉尔·尼赫鲁被从“国家建设者”降格为“无能且自命不凡的精英后裔”。取而代之的是，印度人民党的神话塑造者们捧起了尼赫鲁时期的内政部长萨达尔·瓦拉布海·帕特尔，他们坚称帕特尔被剥夺了最高职位的机会。（莫迪的家乡古吉拉特邦用世界上最高的雕像使帕特尔永垂不朽。）无论如何，他们称，对甘地的非暴力胡说八道不屑一顾的革命者苏巴斯·钱德拉·博斯才是真正的首任总理。
另一种做法是依赖他们拥有的唯一一位顶尖人物。今年6月，内阁庆祝莫迪先生成为印度“连续任职时间最长的民选总理”。这一说法中包含的严苛限定条件排除了尼赫鲁的前五年（当时尚未举行选举）以及英迪拉·甘地（她曾失去权力随后又重新掌权）。对尼赫鲁的敌意是可以理解的，因为独立后的印度不可能有两个首任总理。
印度人民党（BJP）或许使印度“藏红花化”了。但其修正主义并未真正深入人心。上个月在德里举行的教育改革抗议活动中，这一点显而易见。学生们拿着宪法副本，展示着宪法起草者B.R.安贝德卡尔、反殖民革命者巴加特·辛格以及甘地的照片。他们高喊着追溯到独立斗争时期的口号。尽管这些学生在意识觉醒后的绝大部分时间都生活在“第二共和国”中，但他们使用的符号却属于“第一共和国”。建设一个国家并非易事，而对其进行改造则更为困难。
46
中国
中国在北极
领先地位
北海航线是苏伊士运河的替代方案——尽管成本高昂且风险较大
对于一个重新关注海上咽喉要道的世界来说，“迪拜塔”号（Dubai Tower）的航行带来了一线希望。这艘集装箱船于8月15日从中国宁波港出发，开启了首条经北冰洋连接亚洲与欧洲的定期货运服务。该船的中国所有者海传（Sea Legend）表示，前往英国东海岸费利克斯托（Felixstowe）的5,000公里航程将在约20天内完成。这大约比走通常的苏伊士运河航线快两倍。至关重要的是，这项被该公司称为“北极快线”的服务避开了卷入中东军事冲突的风险。
这条通过所谓北海航线（NSR）的新服务穿过白令海峡并经过俄罗斯北部，是对中国北极雄心的早期测试。2018年，中国概述了在北极开发航运路线、采矿和基础设施的计划，希望利用迅速消融的极地冰盖。许多被统称为“冰上丝绸之路”的计划遭遇了阻力。在拥有北极领土的8个国家中，有7个是西方民主国家。它们担心中国在该地区与俄罗斯密切合作，且中国的活动可能对未来的军事目的（如潜艇运行）有所帮助。一些国家对中国自称“近北极国家”感到不满，因为中国的最北端与汉堡市大约处于同一纬度。
尽管如此，中国仍希望吸引那些因霍尔木兹海峡关闭以及受伊朗支持的胡塞武装对红海航运构成威胁而感到不安的政府和公司，重新关注这条北向航线。中国也比以往任何时候都更担心自己对狭窄的马六甲海峡的依赖，该海峡是印度洋与太平洋之间的枢纽，中国大部分能源进口均经此通过。而且，中国很快将不再是北海航线上的唯一参与者。6月，日本表示将修订其北极政策以推广该航线。韩国航运公司PanStar计划于8月22日派遣一艘集装箱船在北冰洋进行试航，目标是在短短18天内完成从釜山到费利克斯托的航程。
并非坦途
得益于气候变化，北海航道（NSR）的条件可能会变得更加有利。北极冰盖夏季的最小范围每十年缩减超过 12%。今年 3 月，冬季冰盖的最大范围达到了自 1979 年开始记录以来的最低水平。在 8 月 16 日的最新可用读数中，北极冰盖覆盖面积为 5.5m 平方公里。这远低于 1981 年至 2010 年间同一日期的冰盖范围中位数，但高于 2012 年记录到的历史最低范围。在那年 9 月，冰盖缩减至仅 3.4m 平方公里。
海传传奇（Sea Legend）服务的一个卖点是，更短的航行时间意味着 ▶
→ 本节其他内容
47 巧克力之甜
48 茶馆：对 20 世纪 90 年代的怀旧
中国 47
▶ 燃料消耗更少——燃料占航运成本的十分之七。这也减少了碳排放。此外，北海航道较凉爽的天气更有利于运输一些对温度敏感的高科技产品。如果不使用冷藏，经苏伊士运河航线的集装箱内部温度可达 60℃。“迪拜塔”号（Dubai Tower）承运的货物包括电池、太阳能组件和电动汽车零部件，主要来自中国长江三角洲的城市。
尽管有这些显而易见的优势，“北极快线”也凸显了北海航道的局限性。它仅在 8 月至 10 月期间运行，此时该航道基本无冰。海传传奇计划在该时间窗内使用 7 艘船只进行 8 次航行。在一年中的其余时间里，只有配备加固船壳的船只才能在北极航行。海传传奇购买或租赁了一些此类船只，但规模相对较小，因为这类船只成本高昂且供应短缺。它们可能还需要俄罗斯核动力破冰船的护航。所有这些额外成本都削弱了北极航线的价格优势。
保险费也是如此。尽管苏伊士运河存在战争风险保费，但北海航道的保费通常仍高于苏伊士运河。此外，其保费比绕行好望角（苏伊士运河的主要替代方案）高出 40%。保险巨头安联（Allianz）在最近的航运业年度安全审查中警告称，由于恶劣的天气以及用于维修、援助、救援或打捞的基础设施有限，偏远的北极航线带来了显著的额外风险。报告指出，在过去十年中，机械损坏或故障导致了北极海域近一半的报告航运事故。
环境影响可能也更大。挪威环保组织 Bellona 在去年的一份报告中警告称，由于在寒冷条件下清理燃料的难度大，且距离应急响应中心遥远，北极地区没有有效的处理燃料泄漏的方法。（国际海事组织自 2024 年起已禁止在北极使用和运输重燃油，尽管俄罗斯并未理会这一禁令。）船舶产生的噪声污染对北极的鲸鱼和海象等海洋哺乳动物的影响也比其他地方更大，因为低频声音在冷水中传播得更远。北极航线的船舶会经过至关重要的觅食地和繁殖地。
国际承运商在很大程度上因为地缘政治风险而避开该航线。在北海航道航行的船只通常需要俄罗斯的许可证、破冰服务或其他援助，这可能会违反与乌克兰战争相关的西方制裁。未来，俄罗斯与其他国家之间为了控制该地区丰富的矿产资源而加剧的竞争，可能会干扰航运。北极曾被维护为一个和平地带，但现在正日益军事化，因为唐纳德·特朗普继续威胁格陵兰岛，而北约则寻求威慑俄罗斯在其北翼的侵略行为。
这一切有助于解释为什么 2025 年北海航线的总过境货物量仅为 3.2m 吨，高于前一年的约 3m 吨。与去年通过苏伊士运河的约 464m 吨相比，这个数字微不足道。北极航运的大部分是在俄罗斯和中国港口之间，而非在亚洲和欧洲之间，其中很大一部分是俄罗斯石油或天然气的运输。在最近的一份报告中，荷兰智库克林根代尔研究所（Clingendael Institute）得出结论，该航线“最可行的功能……仍然是作为北极资源出口的选择性走廊”。
尽管如此，华盛顿北极研究所的马尔特·胡姆佩特（Malte Humpert）认为，当传统航线中断时，北海航线作为季节性贸易走廊对中国的重要性将会增加。他预想十年后每年将有数百次集装箱航行。胡姆佩特先生表示：“问题在于西方运营商何时决定重新进入市场——更重要的是，他们是否还能进入。”北极航行需要培训和操作经验，而“中国和韩国现在正在建立这种专业能力”。
海传传奇（Sea Legend）承认，目前北极服务的季节性窗口狭窄且运力有限，使其难以与传统航线竞争。但该公司已计划到 2028 年将航行窗口扩大到四个月，并最终扩展到全年。中东地区进一步的冲突可能会激励该公司和其他公司投资于更大规模的强化冰级船舶。而且，如果北极地区的温度继续以 3. 倍于全球年平均速度上升，那么对北海航线可行性的质疑可能会随着中国在该地区野心的增长而迅速消退。■
巧克力制造商
苦尽甘来
上海
中国人会陷入“巧克力狂热”吗？
中国的咖啡店品牌和肯德基门店数量超过全球任何其他地方。但巧克力从未真正赢得市场，中国人的年均消费量仅为 100 克——仅为法国人消费量的百分之一。少数几个大众市场的外国品牌——玛氏（Mars）、费列罗（Ferrero）、雀巢（Nestlé）和好时（Hershey）占据了中国消费量的五分之四。“在全球消费的所有巧克力中，只有 2% 触及了中国人的味蕾。”
这种情况现在可能正在改变，得益于全国各地涌现的一批中国巧克力制造商。大多数是手工店，但有些公司正在投资供应链并打造大型企业。以 Choc Revive 为例，这家总部位于上海的公司由生鲜贸易商谢亮（Xie Liang）于 2023 年创立。他开店的速度快得惊人。目前他拥有 80 家门店，希望在年底前增加到 300 家。谢先生表示，其公司去年的销售额为 1.6 亿元（2370万美元），并希望今年能实现四倍或更多的增长。
Choc Revive 试图将巧克力本土化。它从非洲采购自己的可可豆，但在中国设立了研发部门，并进行自主烘焙和研磨。其大规模投资旨在产生规模经济，但它所积累的技术诀窍可能会扩散并催生竞争对手。另一家本土巧克力公司 Nibbo 凭借其口感丝滑的甜点赢得了国际奖项。而起家于中国西南地区咖啡种植的 Saturnbird，现在正尝试在当地种植自己的可可树。
谢先生的产品 100 克一块售价 58 元——相当于两个人的廉价正餐。在消费者紧缩开支的当下，向他们销售高端产品是一个尴尬的时机。通过生产尽可能廉价的商品来削弱竞争对手是惯常的公式，奶茶品牌就是一个典型的例子。
为什么人们愿意为巧克力买单？谢先生表示，制作具有中国风味（如类似于大荔枝的龙眼）的巧克力似乎很受欢迎。一个更简单的解释是，中国人在漫长的等待后，终于学会了热爱这世上最美味的味道。
*近似值，取决于航线 来源：中国船东协会；Sea Legend
48 中国
阴郁的中国在怀念 20 世纪 90 年代
那时，领导者能兑现承诺，机遇在招手——至少故事是这么说的
20 世纪 90 年代再次鲜活了起来。在过去的几天里，中国国家电视台的观众们沉浸在那个年代的场景之中。一部 12 集的纪录片在黄金时段播出，以纪念已故的江泽民——这位在整个时期领导国家的领导人——诞辰一百周年。那是怎样的一个十年啊，其开端是 1989 年天安门流血事件，而终点则是 2001 年中国加入世界贸易组织所带来的所有希望。
这种进程在回顾时可能显得必然，但实际上始终充满艰辛。然而，令观众着迷的与其说是江泽民的成就，不如说是一种亲民感和风采，这在如今的中国领导层及其严密编排的公开举止中，似乎已几乎不可想象。在 1989 年的一个场景中，在天安门民主抗议活动遭到镇压四个月后，他正在北京拥抱婴儿。在另一个场景中，江泽民在打乒乓球时挥出一记正手，然后告诉国家队运动员，他希望他们在即将到来的亚运会上赢得金牌。“这话我收回。你们不应该有太大的压力，”他带着灿烂的笑容说道，房间里爆发出一阵笑声。在第三个场景中，江泽民在加利福尼亚的一次正式晚宴上高唱京剧，这与如今中国那些拘谨的官僚阶层（尤其是习近平）形成了鲜明对比。
随后，在纪录片播出期间，出现了一个巧合：8 月 12 日，江泽民的核心搭档朱镕基在 97 岁时去世。如果说江泽民是政治上的指挥家，那么担任总理的朱镕基则是那个具体执行、处理脏活的人。网络评论赞扬了朱镕基直率且高效的方式——这在含蓄地指责如今那些更为谨慎的官员。在他去世当天，纪录片恰好播放到朱镕基解释他如何诱导干部进行艰苦改革的部分：“我不是来谈论利弊的。我是来谈论如何把事情办成的。”
甚至在纪录片播出和朱镕基去世之前，对 20 世纪 90 年代的怀旧之情就已经在积聚。去年夏天，一个名为“繁荣年代之美”的趋势在网上走红，人们分享 20 世纪 90 年代和 21 世纪初充满青春活力的图像。最近的一部热门电视剧《繁花》讲述了那个时期的上海，捕捉到了这种情绪。在某个时刻，旁白说道：“在机遇面前，每个人都是平等的。抓住它，你就能改变人生。”
一些人在缅怀朱镕基时也表达了类似的情感。经济学家赵健在一篇被广泛分享（随后被从互联网上删除）的文章中写道：“那时人们眼中闪烁着光芒，心中充满笃定，而且能看到一个真实的明天。”
这一切意味着什么？一种愤世嫉俗的看法是，记忆在玩它惯常的把戏。如果 20 世纪 90 年代现在看起来像是一个充满可能性的十年，请记住，当时的中国要贫穷得多，正处于高速发展的前夜。1990 年，经通胀调整后，中国的人均 GDP 仅为 920 美元；而今天则是其 15 倍。或者看看大学毕业人数的增长：现在每年 1200 万，而 20 世纪 90 年代约为 70 万。在江、朱时代，中国近四分之三的人口仍生活在贫困的农村。难怪当时足够幸运获得良好教育的城市居民会觉得命运掌握在自己手中。然而对于大多数中国人来说，前方的道路是一场接一场的奋斗。
因此，这种怀旧不应被解读为对历史的严谨记录。相反，它是对今日不安情绪的反映。最明显的是，人们对经济感到沮丧。就业机会萎缩，房地产暴跌，许多企业家在苦苦挣扎。然而，沉溺于这些问题会带来后果，例如，审查员会删除缺乏“正能量”的关于经济的帖子。对于一些人来说，回顾 20 世纪 90 年代的美好时光是绕过此类管控的一种方式。而对于另一些人来说，这可能仅仅是对更好生活的渴望。
如今，人们对中国的政治也感到极大的沮丧。学者们在争论江和朱是否真的配得上“改革派”的标签。他们肯定不是政治改革者：共产党元老选择他们是为了强化党的意志，并在天安门创伤之后收拾残局。作为经济自由化推动者，他们同样有局限性。他们标志性的国企改革是“抓大放小”，在现代化金融、基础设施和重工业等重要企业的同时，剥离无用企业。正如江所言，其目的是为了掌控“国民经济命脉”。最重要的是，他们的目标是加强国家权力并巩固党的统治。
那里的做法截然不同
尽管如此，江和朱确实胆识过人，在面对真实阻力时敢于承担巨大的风险。保守派（即左翼人士）反对他们的市场驱动政策，担心向世界开放贸易会引入有害的政治思想。许多改革也付出了巨大的代价，超过3000万名工人被解雇——尽管最终带来的收益更大。如今的领导人是否拥有类似的远见和勇气？诚然，他们在一些高科技进步期间掌权。然而，与美国的贸易战仍在持续，而房地产市场依然陷入长年累月的低迷。对朱的悼念不仅来自普通民众，还来自大学、法律协会、金融团体和智库。很难不将这种崇敬视为一种对更明智领导力的恳求。
除此之外，20世纪90年代领导人的风格中有一种无法量化的特质——在于他们说话的坦率，他们展现出的温暖和真实的人情味，以及他们在公众之间走动时的自在。当然，人们当时也对江和朱心存戒备；在他们的领导下，中国绝非自由之邦。但他们赢得了钦佩，甚至被热爱。而今天，当人们谈论习先生时，主导情绪往往是恐惧。许多中国人如此谨慎，以至于完全避免提及他的名字。他是至高无上的统治者，而不是一个会开玩笑或突然唱起歌来的人。■
49
国际
导弹防御
弹药告罄
基辅与台北
美国及其盟友的拦截导弹正面临短缺
基辅的居民再次在拥挤的地铁站或地下车库的汽车中就寝。在他们上方，空战正演变成一场单方面屠杀。俄罗斯正用越来越多的弹道导弹和高超音速导弹轰炸这座城市。乌克兰此前曾利用爱国者导弹防御阵地挫败了大多数此类攻击，但现在正处于“Going Winchester”状态——这是军事术语，意为弹药告罄。
乌克兰的困境是西方拦截弹短缺最剧烈的迹象。美国军械库的枯竭也解释了为什么唐纳德·特朗普总统多次在威胁与伊朗重开战火后选择退缩。美国的盟友们感到惊恐。
瑞士预计其订购的爱国者导弹将延迟最多七年交付。台湾担心多年前订购的 102 枚导弹无法按计划在今年抵达。生产最先进版本 PAC-3 MSE 的洛克希德·马丁公司表示，鉴于五角大楼有权将其调配，该公司无法向任何外国客户保证交付时间。
弹药短缺在战争期间很常见，尤其是在大规模的消耗战中。然而，美国在与一个中等强国进行了一场短暂的战争后就出现了短缺。如果它必须与强大的中国作战会怎样？华盛顿智库传统基金会（Heritage Foundation）最近的一项研究认为，美国将在几天内耗尽其拦截弹，并且无法使用其在日本和关岛的基地。
这场危机的直接原因是，在特朗普今年与伊朗的战争期间，美国及其中东盟友大量使用了拦截弹。另一家智库战略与国际研究中心（CSIS）认为，五角大楼已经发射了近三分之二的战前爱国者导弹库存以及大量其他拦截弹（见次页图表）。
但这场战争仅仅加速了一场酝酿数十年的清算。在大西洋两岸，国防工业一直按照和平时期的节奏运作，尽管战争的鼓声已在加快。这一错误因导弹在全球范围内的扩散、将军们倾向于在炫酷武器而非其消耗弹药上投入更多资金、美国功能失调的预算流程，以及对美国空中军事霸权的自满而进一步加剧。
CSIS 估计，要将爱国者导弹库存重建到战前水平将要等到 2029 年。这是在假设国会批准特朗普高达 $1.5trn 的国防预算请求（比当前支出增加 50%）的前提下，而这种情况不大可能发生。然而，爱国者导弹已成为 19 个国家点防御（即保护空军基地等特定目标）的首选系统。“爱国者导弹就像一套晚礼服方案：你不需要经常穿它，但当你需要时，几乎没有其他替代品，”英国智库皇家联合军种研究所（RUSI）的贾斯汀·布朗克（Justin Bronk）解释道。
问题在于，弹道导弹通常比试图击中它们的拦截弹更便宜（PAC-3 MSE 拦截弹每枚成本 400万美元-5m）。伊朗的见证者（Shahed）无人机（或其他无人机，UAVs）的价格则低两个数量级。“如果你的对手能以一辆二手车的价格部署一架单向攻击无人机，而你的应对方案是一枚地对空导弹，其价格相当于伦敦这里的一套小公寓，那么即使你赢得了每一次交战，你在整体交换比上也是输的，”一名科威特军官萨巴赫·阿尔·萨巴赫（Sabah Al Sabah）上校在最近的一次 RUSI 活动中如此主张。此外，防御方通常会对一枚来袭导弹发射至少两枚拦截弹，以提高成功率。
→ 本节其他内容
— The Telegram is away
50 International
乌克兰即兴构建了一套分层防御系统。它拥有世界一流的网络来对抗俄罗斯的无人机群。得益于欧洲防空系统的部分支持，它还能拦截大多数大型巡航导弹。但只有珍贵的爱国者导弹（Patriots）能够应对俄罗斯的伊斯坎德-M（Iskander-M）和匕首（Kinzhal）弹道导弹。今年3月，乌克兰曾自豪地表示拦截了约70%的弹道导弹。到了7月，这一比例降至20%。8月的形势将更加严峻。在8月5日的一个夜晚，拦截率为零。空军已停止发布定期更新。俄罗斯注意到这一点，正在增加导弹产量。
随着冬季临近，乌克兰请求提供至少300枚爱国者拦截弹。一些国家仍拥有相当数量的库存，包括德国、希腊、波兰和西班牙。但它们越来越不愿将其交出。
另一个解决方案是加快生产速度。制造商今年同意将PAC-3的年产量从600枚增加到2030年代初的约2,000枚。此后，五角大楼征集了进一步加快速度的方案。但战略与国际研究中心（CSIS）的汤姆·卡拉科（Tom Karako）指出，目前尚未签署实际的增产合同。
盟友的工业能力可以提供帮助。日本根据许可每年生产约30枚PAC-3拦截弹，主要用于自用。欧洲公司MBDA即将开始在德国生产较旧的PAC-2型号。在7月的一次北约峰会上，特朗普先生表示，乌克兰也将被允许根据许可制造爱国者拦截弹，但此后他对这一想法表示怀疑。有报道称，制造商洛克希德·马丁（Lockheed Martin）和RTX担心丢失技术诀窍，或被更便宜的乌克兰版本抢风头。
由于缺乏同样的实战经验，欧洲公司正急于制造自己的武器。法意联合研制的SAMP-T据称能够击中短程和中程弹道导弹，但乌克兰军方消息人士称，其效果不如爱国者PAC-2，更不用说PAC-3了。乌克兰将从明年起分阶段接收带有更新拦截弹的改进型SAMP-T电池。然而，它们的产量低于爱国者导弹。
乌克兰灵活的国防工业和新兴的西方公司彻底改变了无人机战争。他们现在正致力于导弹防御。其中一项努力是Freyja项目，涉及乌克兰公司Fire Point以及许多其他欧洲军火商。他们的目标是在不使用美国技术的情况下，开发一种成本仅为PAC-3四分之一的拦截弹，测试将于明年开始。质疑者众多。另一家英国智库国际战略研究所（IISS）的道格拉斯·巴里（Douglas Barrie）将Freyja描述为一次“登月计划”（极难实现的尝试）。
台湾正试图扩大其天弓三型（Sky Bow III）拦截弹的生产，据称该型号相当于PAC-2和较旧的PAC-3。预计今年将开始量产天弓四型（Sky Bow IV），其声称能力强于PAC-3。然而，很少有国家愿意通过购买台湾武器来冒触怒中国的风险。
洛克希德·马丁公司宣布开发PAC-3 ACE，又称“小爱国者”（Baby Patriot），它比PAC-3 MSE体积更小、能力稍弱，但成本仅为后者的一半。它可能在2028年投入生产。
有些人希望开发已久的激光武器能够改善防空领域残酷的算术题。但到目前为止，美国和以色列的实验性武器仅在对抗无人机方面报告了成功。
先击杀弓箭手
“防空不能与进攻行动脱节，”美国前空军将领大卫·德普图拉（David Deptula）指出，“最便宜的拦截弹是你永远不需要发射的那一枚，因为武器根本没有被发射出去。”与其射击“箭”，不如射击“弓箭手”——即敌方的生产基地、指挥控制中心、储存仓库和导弹发射架——这确实有其道理。但美国的远程进攻性武器库存也已枯竭。战略与国际研究中心（CSIS）估计，美国已使用了其 3,000 多枚战斧巡航导弹的三分之一。美国拒绝向乌克兰供应此类导弹，而向包括澳大利亚、德国、日本和荷兰在内的其他客户的交付可能会延迟。
乌克兰已经制造出能够触及俄罗斯更深处目标的无人机，特别是能源基础设施、军工厂和消费品分销仓库。但这些无人机速度慢且携带的炸药少。Fire Point 公司较重的 Flamingo 巡航导弹取得了一些成功。乌克兰希望很快部署其 FP-7 以及规模更大的 FP-9 弹道导弹，据称其射程超过 800 公里。
摧毁移动发射架非常困难。今年，美国和以色列通过轰炸发射架以及地下“导弹城”的入口，将伊朗的导弹火力压制了高达 90%。但即便拥有完全的制空权和精密的监视能力，他们也未能将其完全阻止。美国损失了雷达、指挥中心、预警机、加油机、无人机、战斗机等。自 4 月停火以来，美国一直无法阻止伊朗向霍尔木兹海峡的船只开火。
未来任何针对中国（例如在台湾问题上）的空战都将是一项艰巨的任务。中国的导弹可以触及远至阿拉斯加和夏威夷的基地，以及太平洋中部的军舰。传统基金会（Heritage）估计，为了防御中国人民解放军（PLA），美国将需要超过 19,000 枚爱国者拦截弹——这是 CSIS 对当前库存估计值的 20 倍以上。
射击解放军这个“弓箭手”将是一个诱人的替代方案，但除了突破中国的防空系统外，还面临两个大问题。首先，美国必须打击中国大陆，这将招致对美国本土的报复。其次，解放军火箭军不仅负责常规导弹，还负责陆基核导弹：对其发动攻击可能会被视为企图摧毁中国的核力量，从而带来核升级的风险。
相反，美军正在通过各种形式的分散部署，练习如何在中国的“武器交战区”内生存。在一个廉价无人机和导弹激增的世界里，对军事目标进行加固、分散和隐蔽——并在被摧毁后迅速重建——是至关重要的技能。
尽管深层打击武器并非万能药，但它们确实有助于威慑攻击。这解释了为什么美国的盟友们正争先恐后地寻找射击远方目标的方法。已经遭受持续攻击的乌克兰希望将战争带到俄罗斯，从而造成足够的痛苦，使弗拉基米尔·普京重新考虑这场无休止的冲突。
所有解决拦截弹短缺的可能方案，即使被加速和强化，也需要时间。与此同时，在乌克兰，平民仍必须在地下掩体中避难。在面对另一个寒冷且危险的冬天时，许多人会问，他们的领导人是如何让局面演变成这样的。其他政府应当引以为戒。■
视线中的温彻斯特
美国，估计弹药库存，% 2026年2月27日-8月16日
估计使用量：■ 低 ■ 高 ■ 剩余
防空拦截弹
远程打击导弹
来源：M. Cancian 和 C. Park, 战略与国际研究中心（CSIS）
51
商业
全球贸易
路线与分支
艾因索赫纳（AIN SOKHNA）
中国企业正将供应链铺向全球
位于苏伊士湾的古港口艾因索赫纳，曾接收用于法老礼服的绿松石。如今，它迎来了一种新的财富来源：中国制造商。短短几年内，他们的投资就改变了该地区。数十家工厂拔地而起，产品涵盖从玻璃纤维到开关设备的所有领域。今年1月，一个由两家中国物流巨头——中远海运（COSCO）和长江和记（CK Hutchinson）提供资金建设的新港口码头开始运营。
艾因索赫纳是更广泛的苏伊士运河经济区的一部分，这是一个向北延伸至地中海的工业园区和港口网络。近年来，该区域吸引的投资中约有一半来自中国。工厂建设的范围远超北非（见次页地图）。从沙特阿拉伯和匈牙利到巴西和印度尼西亚，中国的工业园区及其配套基础设施正纷纷涌现。
这种投资的速度已大幅加快。仅在过去三年中，中国公司在海外建厂的支出就超过了 2000亿美元（见图表）。其供应链的特性也在发生变化，主要体现在三个方面。首先，分布更加广泛，在全球几乎每个地区都设有大型生产节点。其次，深度增加，许多中国供应商跟随制造商进入新地点，复制国内那种紧密的生态系统。第三，战略性产业占据主导地位，涵盖从电动汽车、清洁能源到数据中心设备的领域。其结果是，全球制造业的重新布线正在进行中。
有多种原因解释了为什么中国公司在越来越多的地方制造产品。国内消费支出疲软和激烈的竞争促使它们冒险进入新市场。第二任特朗普政府引入的关税也激励了在那些比越南等传统中国前哨站受惩罚性征税较轻的地方进行生产。
全球南方的多个国家提供了额外的激励措施。例如，埃及提供一种“黄金许可证”，为大型项目大幅削减繁琐的行政手续。帮助中国公司在埃及开展业务的律师穆罕默德·埃尔迪布（Mohamed Eldib）表示，这些公司渴望“进来赚钱”，向埃及 1.2 亿人口销售产品，同时将其作为出口基地。
结果就是生产足迹日益分散。以晶科能源（JA Solar）、晶科太阳能（Jinko-Solar）和天合光能（TrinaSolar）这三家中国制造商为例，它们在十年前开始在多个东南亚国家制造太阳能电池板，现在则在海湾地区设立工厂。新的中国生产中心正逐渐被编织进客户的
→ 本节其他内容
53 美国企业的政府掌控力
54 巴特尔比：论 AI 与投资回报率（ROI）
55 Meta 的大案审理
55 Reddit 的困境
56 加拿大的石油繁荣
57 熊彼特：数据中心政治
52 商业
供应链中。德国风力涡轮机制造商 Nordex 现在从其中国供应商去年在摩洛哥开设的一家工厂采购叶片。
欧洲也成为了热门目的地。根据研究机构 Rhodium Group 和智库 MERICS 的数据，2025, 年中国企业在欧洲大陆所有全新“绿地”项目上的外国直接投资增长了 50%，达到创纪录的 89亿欧元 (101亿美元)。匈牙利吸引了大量关注。塞尔维亚的中国存在感也在增强。汽车零部件供应商玲珑轮胎（Linglong Tire）是已在该国开始生产的中国制造商之一。该公司最近宣布，将扩大其位于兹雷尼亚宁（Zrenjanin）市的高度自动化工厂的产能，该工厂为包括大众和福特在内的西方汽车制造商供货。
与此同时，中国的全球供应链进一步深化，越来越多的上游制造业转移至海外。中国电池制造商国轩高科（Gotion）正在摩洛哥首都拉巴特东北方向70公里的一个工业区建设一座超级工厂；包括生产阳极和阴极的贝特瑞（BTR）以及生产铜箔的海亮集团（Hailiang Group）在内的多家中国供应商，正在距离该工厂几小时车程的地方建设工厂以提供投入品。在开罗郊区的一个工业区，一家中国家用电器制造商建有工厂，该公司的一位高管指出，公司同样鼓励其部分供应商设立当地工厂。
许多投入品仍通过运输进口。机床通常从扬立集团（Yangli Group）等中国供应商处进口。在当地无法获得（或价格昂贵得多）的零部件或材料也是如此。2026年上半年，中国资本货物出口同比增长14%。中间产品出口增长了
飞龙在天
中国，绿地制造业外国直接投资流出量，十亿加元
27%。然而，随着时间的推移，更多的投入可能会从更近的地方采购。
与此同时，中国的物流公司正帮助其制造商连接其日益扩张的供应链。这些公司目前在从希腊到斯里兰卡的至少 132 个外国港口运营或进行了投资，此外还包括机场和铁路线，其中包括今年早些时候完工的布达佩斯至贝尔格莱德铁路。
中国企业的海外投资热潮主要集中在少数几个战略行业。为了满足日益增长的电力需求，尤其是全球南方的国家已成为中国清洁能源技术的热切买家。在埃及沙漠中建设的大规模太阳能电站主要依赖中国设备，其中本地生产的比例正在提高。中国电动汽车现在在从里约热内卢到伦敦的城市中已屡见不鲜，且越来越多地在靠近客户的区域中心制造。
人工智能数据中心所需的设备是另一个重点。例如，泰国已成为中国高速光组件制造商的制造中心，如中际旭创（Zhongji InnoLight），其产品被美国云端计算巨头在世界范围内使用。
希望在海外投资的中国公司发现，由于西方竞争对手为了安抚美国保护主义政府而将工厂建设集中在美国，他们几乎在该领域独占鳌头。一名埃及官员补充说，当中国企业决定建设时，速度非常快。
他们的海外扩张并非没有障碍。美国强加的波动关税导致部分项目被取消或缩减。8月13日，白宫发布了一份名为《大转运骗局》（The Great Transshipment Scam）的报告，要求对任何带有中国参与痕迹的进口产品实施全面限制（见“金融与经济”板块）。与此同时，巴西和土耳其等地的政府一直在实施本地含量要求，这导致更多价值在这些国家内部增加，但降低了当地制造业的吸引力。新兴市场的汇率波动和高借贷成本使情况进一步复杂化。
连锁反应
中国政府也给这些全球布局的企业带来了困难。上个月，一套新的对外投资监管方案生效，其中除其他规定外，还限制技术和数据的跨境转移，并引入了国家安全审查程序。
到目前为止，全球制造业在中国进行的整体份额尚未显示出下降迹象。然而，随着越来越多的公司在海外投资，这种情况可能很快会改变。即便如此，世界仍将继续依赖中国商品——无论它们是在哪里制造的。■
商业 53
政府采购
美国及其它
美国企业紧握盟国政府之柄
“我毫无遗憾，”德国石勒苏益格-荷尔斯泰因州（Schleswig-Holstein）的数字化部长迪克·施罗德特（Dirk Schrödter）说道。在一次使用 OpenTalk（一款德国研发的微软 Teams 替代品）进行的视频通话中，他解释了在过去两年里，他如何将该州约 30,000 名公务员从这家美国公司的协作与生产力工具迁移到开源替代方案上。现在，他已开始将员工从 Windows 迁移到 Linux（一种开源操作系统）的过程中。此举引起了其他地方公共官员的关注，他们同样渴望切断与美国科技巨头的联系。他说：“每周我们都会收到来自欧洲境内外其他州、城市和政府的咨询。”
对开源软件日益增长的兴趣反映了一个新的地缘政治现实。长期以来，美国一直利用其商业实力来打击敌人，例如对伊朗和俄罗斯实施制裁。但现在，即使是盟友也担心美国可能会切断他们获取关键技术的渠道——或者至少在贸易谈判期间以此相威胁。
海外政府正在审视其运行能力在多大程度上依赖于美国供应商。去年 5 月，当时的国际刑事法院首席检察官卡里姆·汗（Karim Khan）在唐纳德·特朗普总统因该法院对以色列总理本雅明·内塔尼亚胡（Binyamin Netanyahu）发出逮捕令而对其实施制裁后，失去了对微软电子邮件账户的访问权限，这加剧了人们的担忧。（微软表示其并未切断服务。）今年 6 月，当特朗普强迫人工智能实验室 Anthropic 暂时切断其最新模型在外国的访问权限时，再次带来了冲击。去年美国两次暂停向乌克兰供应武器，也引起了其盟友的注意。
在海外政府招标的整体合同中，美国公司赢得的份额较小。但其产品和服务往往支撑着关键的政府职能。因此，目前正采取各种努力以减少对美国供应商的依赖，并培育本土替代方案。然而，在许多情况下，这样做将极其困难。
尤其是欧洲，已成为试图将美国企业从政府供应链中清除的努力中心。法国国家政府正计划放弃 Teams，并希望将部分计算机
赞助国
选定国家，美国公司赢得的政府合同数量，占总数的百分比
来源：TenderAlpha
到 Linux 系统。包括西班牙雷乌斯市和丹麦奥胡斯市在内的当地政府，已转向使用 Nextcloud 和 Hetzner 等欧洲云服务提供商，这两家公司均来自德国。6 月，欧盟委员会公布了一项旨在提升欧洲大陆“技术主权”的计划，其中一项目标是将敏感政府数据的处理转移至此类提供商。去年，西班牙取消了一项向美国洛克希德·马丁公司采购 F-35 战斗机的订单。英国政界人士正敦促首相执行帕兰提尔（Palantir，一家美国技术公司）与国民医疗服务体系（NHS）之间合同中的终止条款（见英国部分）。
《经济学人》的估算表明，从总体水平来看，美国公司在海外公共支出中所占份额较低。我们计算出，在去年美国上市公司产生的 $25trn 销售额中，大约有 5000亿美元（或 2%）来自外国政府。这相当于除美国以外的经合组织（OECD）国家政府采购额的 7% 左右，而美国占据了这些支出的绝大多数。
空中优势
有些国家比其他国家更依赖美国。根据数据提供商 TenderAlpha 的数据（见图表），从去年美国公司赢得的政府合同数量来看，澳大利亚（6% 的合同）和英国（4%）的依赖程度高于法国（2%）和德国（1%）。
然而，美国公司在政府的关键领域发挥着不成比例的重要作用。我们估计，美国公司从外国政府获得的业务中，大约有五分之二集中在信息技术和国防领域。（由全球公共卫生系统采购的药品和医疗设备则占据了另一个很大份额。）
这通常是因为缺乏替代方案。分析公司 Synergy Research 指出，Alphabet、亚马逊和微软控制着全球云计算市场的三分之二。另一家研究机构 Forrester 认为，在大多数企业软件细分领域，十大供应商中美国供应商占据了七到九个。
那些确实存在的替代品往往较为劣质。有时损失可以忽略不计：在 OpenTalk 上，用户无法发送在屏幕上飞舞的表情符号。但其他缺陷则更为严重。许多政府选择的开源办公软件套装 LibreOffice 缺乏微软提供的 AI 功能——随着技术的进步，这一弱点将变得更加明显。OVItcloud 是欧洲最大的云计算提供商，是一家法国公司，其营收仅为亚马逊云服务（AWS）的百分之一左右，这使其难以在价格上竞争或在创新上跟进。同样，想要最先进防空系统的政府必须转向洛克希德和 RTX（另一家美国武器制造商），这两家公司共同生产爱国者导弹。
切换成本也可能非常高昂。洛克希德向从挪威到比利时的各国政府供应 F-35s。这些国家本可以选择欧洲喷气式飞机，如阵风（Rafale）或台风（Typhoon），即使它们的技术水平较低。但这将需要大规模的运行调整，包括重新培训飞行员和支持人员，以及更换武器库存——在已经向洛克希德喷气机投入巨资的情况下，这很难被接受。
或者考虑医院的病人记录系统。更换新供应商的高昂成本，部分原因是挪威议会在 6 月投票决定继续使用 Epic（一家美国医疗数据提供商），尽管医生们对该软件并不满意。 ►
更正：在上周的“硅谷对决”一文中，我们写道英伟达（Nvidia）制造其芯片。事实上，它仅负责设计。对此我们深表歉意。
54 商业
数据库。
当地公司对美国供应商的依赖使问题进一步复杂化。另一家研究机构 Omdia 的罗伊·伊尔斯利（Roy Illsley）指出，OVHcloud 运行的大部分软件来自 VMware，而 VMware 是一家名为博通（Broadcom）的美国公司所拥有。
那些避开美国供应商的政府可能会冒着激怒其情绪多变的总统的风险，这可能会使这个超级大国成为一个更不可靠的盟友。智库斯德哥尔摩国际和平研究所（SIPRI）的皮特·韦泽曼（Pieter Wezeman）注意到，丹麦在 7 月份选择了由美国航空航天巨头波音公司制造的海上巡逻机，而非欧洲的替代方案；在他看来，这是一个信号，表明尽管在格陵兰岛问题上存在冲突，但该国并不希望疏远美国。
作为慰藉，美国的盟友们应该记住，反方向的依赖也同样大量存在。美国联邦政府的大部分运行依赖于德国 SAP 公司提供的软件。而向外国政府销售产品的美国公司通常依赖于全球供应链。每一架 F-35 的后机身都由英国武器制造商 BAE 系统公司（BAE Systems）在兰开夏郡制造。云计算中使用最先进的芯片是在由台湾制造商台积电（TSMC）运营的设施中，使用荷兰公司 ASML 的设备制造的。美国的盟友们或许可以将重点放在如何让自己变得更加不可或缺上。■
巴特尔比（BARTLEBY）
回报政策
如何衡量人工智能带来的收益
起初，人工智能出现了，各地的老板们都陷入了狂热，宣称无论这项技术是否有用，无论成本多少，你们都必须使用它。于是，他们引入了 AI 排行榜，并创造出像“toxenmaxxing”这样愚蠢的词汇。果然，使用率上升了。随后，老板们想起了一个非常基础的概念——“预算”，并意识到这可能并不是个好主意。接着，另一个问题在董事会会议室中回荡，这个问题关乎回报，且更难回答。
这段关于 AI 采用过程的简略“创世纪”并不完全公平。只要组织试图鼓励使用，查看员工的 token 消耗量等指标仍然是有意义的。“我预期软件工程师的花费会比财务分析师更多，”一家软件公司老板说道，“但财务部门的任何人，其花费不应该是零。”
但使用量（也称为投入）指标无法告诉你是否创造了价值。芬兰 LUT 大学（LUT University）的米卡·鲁奥科宁（Mika Ruokonen）指出，即使 AI 产生了收益，其中很多收益也以零散的时间节省形式悄悄地归于个人。如果一名员工多出了一个或两个小时的时间去 Vinted 上购物，那么确实有人受益了（主要是 Vinted），但这并不是你的公司受益（除非你在 Vinted 工作）。
将关注点转向 AI 投资的回报，意味着要更多地关注基于结果的衡量标准，例如团队生产力或客户满意度。但这些需要深思熟虑。例如，科学家们正在利用 AI 撰写更多论文（太棒了！），但这些论文的价值却令人质疑（糟糕！）。萨里大学（University of Surrey）的图尔西·苏查克（Tulsi Suchak）及其合著者最近的一项研究发现，在 2014 和 2021 年之间，基于一套健康与营养数据集的论文年均产出为 4 篇；而在 2024 年的前九个月里，这一数字达到了 190 篇。
因此，基于结果的衡量标准必须旨在捕捉 AI 的正面和负面影响。例如在谷歌，工程师长期以来在三个维度上被衡量。速度是其中之一：比如代码审查需要多长时间。便捷度是另一个：在入职新开发人员等环节中，系统存在多少摩擦。最后，也是至关重要的一点，是质量：“我不想仅仅为了追求速度和发布糟糕软件的便捷，”谷歌云（Google Cloud）的理查德·塞罗特（Richard Seroter）说道。
使问题进一步复杂化的是，AI 的收益往往需要一段时间才能显现。人们必须花时间学习如何使用这项技术。产出的增加可能会在其他地方造成瓶颈。多伦多大学（University of Toronto）的克里斯蒂娜·麦克埃尔赫兰（Kristina McElheran）及其合著者去年发表的一篇论文研究了美国制造企业的 AI 采用情况，并确认了“J 曲线”的存在，即生产力在提升之前会先出现下降。这种效应在成熟企业中尤为明显，部分原因似乎在于管理实践的改变，而这些实践此前曾帮助企业监控绩效。任何关于预期回报的计算都必须将这段初期的混乱期考虑在内。
老板们还必须对一个更根本的问题定调：他们感兴趣的是什么样的回报？展示 AI 财务回报的一种方式，是将节省的时间作为削减员工人数的借口。对于某些岗位和在某些情况下，这可能是合理的。但大规模裁员与员工士气之间的关系，就像盐与蛞蝓一样糟糕。而且，离职的人无法被重新部署到更高效的岗位上。一个更好的选择是计算 AI 在扩大人员规模方面节省了多少资金。（计算有多少额外收入可归功于这项技术则更像是一门艺术；起码，你需要良好的基准数据和 A / B 测试文化来隔离其影响。）
最重要的是，关于人工智能的发展方向存在着巨大的不确定性。这便成为了采取第三组衡量标准的理由，以与针对投入和产出的标准相配套。塞罗特（Mr Seroter）将这些标准称为“基于组织”的标准，它们更倾向于捕捉变革管理得如何有效。这些标准可以包括员工对人工智能实施方式的满意度数据，或者在人工智能无法自动化的领域建立内部专业知识的进展情况。单一地关注使用率会忽略回报的重要性；而僵化地关注回报则会忽略持续学习的必要性。
商业 55
Meta
社会清算
洛杉矶
一场大片般的审判与“大烟草”案的相似之处
10月初，一部关于马克·扎克伯格创立Facebook的电影《社交网络》的续集计划上映。这部名为《社交清算》（The Social Reckoning）的电影将戏剧化地展现举报人的角色，他们指控这家科技巨头隐瞒了其产品有害的事实。其中一些相关人物将参与到一场同样生动的戏剧中，这场涉及Facebook、Instagram及其母公司Meta的戏剧将在加利福尼亚州奥克兰的一家联邦法院上演。口头辩论已于8月18日开始。无论是否被称为“清算”，这场审判都可能成为一场“大片”。
加利福尼亚州、科罗拉多州、肯塔基州和新泽西州正领导一个由29个州组成的团体起诉Meta，指控其平台开发了对儿童有害的功能，并侵犯了他们的隐私。在开庭陈述中，加利福尼亚州副总检察长梅根·奥尼尔（Megan O'Neill）表示，Meta的商业模式是“诱导用户上瘾，尽可能长时间地留住他们，收割他们的数据，然后向公众隐瞒真相”。加利福尼亚州总检察长罗布·邦塔（Rob Bonta）将此案描述为Meta的“烟草时刻”，将该公司涉嫌的欺骗行为比作20世纪90年代导致烟草公司支付巨额赔偿的行为。
Meta的律师保罗·施密特（Paul Schmidt）告诉陪审团，公司认真对待对年轻用户的责任，并补充说，青少年的社交媒体使用与健康状况恶化之间没有明确的联系。预计Meta将抵制原告试图将审判变成一场关于社交媒体对儿童影响的全民公投，而将重点放在争议的狭窄法律问题上。其律师询问，既然青少年使用谷歌旗下的YouTube和中国字节跳动创建的TikTok的频率甚至高于Instagram，为什么Meta会被单独挑出来。
此案赌注很高。在由新墨西哥州政府提起的一起类似案件中，一名法官本月命令Meta支付近9500万加元的罚款。在奥克兰担任原告的四个州在Meta败诉的情况下，似乎在寻求高达此数额200倍以上的赔偿，其估值达到了1930亿加元——大致相当于该公司在2025年的营收。如果这成为其他州的基准，将给Meta带来毁灭性打击。
不过，是否会发展到那种地步则是另一回事。奥尼尔女士用带有情感色彩的语言指控Meta特别针对年幼儿童。她频繁使用“上瘾”（hooked）这个词。但与3月份在洛杉矶进行的一场审判不同——在那场审判中，Meta和谷歌被迫向一名因将大部分生活花在社交媒体上而受损的20岁年轻人支付赔偿金——奥克兰的案件并非聚焦于成瘾问题。
相反，正如其中一位总检察长所称，这是“美国历史上规模最大的消费者保护诉讼”。加州大学伯克利法律与技术中心的文森特·乔拉莱蒙（Vincent Joralemon）表示，核心问题在于Meta是否在产品影响方面欺骗了公众。这与烟草案件类似。乔拉莱蒙先生认为，这也意味着举报人的证词可能比关于成瘾的棘手心理学辩论更具相关性。
投资银行D.A. Davidson的吉尔·卢里亚（Gil Luria）表示，Meta的投资者倾向于认为欺诈指控的举证责任“相当高”。诉讼风险压低了Meta的估值，导致其市盈率低于其他大型科技同行。但卢里亚先生认为，股东们与其担心巨额赔偿，更担心如果Meta在法庭上败诉，检察官可能会寻求的补救性变更。这些变更包括Meta、谷歌和TikTok用于保持高用户参与度的功能，例如能够持续加载新帖子的“无限滚动”。卢里亚先生说：“如果没有无限滚动，我们就不会有这么多广告。”
即便 Meta 败诉，它仍有上诉的空间。例如，它可以基于《通信法》第 230 条寻求保护，该条款保障互联网平台无需为第三方发布的内容承担法律责任。Meta 曾试图利用这一论点在奥克兰的审判开始前将其终止，但法官裁定该上诉为时过早。
马克的一场标志性审判
美国公众对社交媒体给儿童带来影响的担忧程度很高，而被拖入法庭可能无助于 Meta 在家长心目中的形象。为了化解类似的对 AI 的担忧，OpenAI 在 8 月 18 日推出了面向儿童的 ChatGPT (ChatGPT for Kids)，旨在加强针对 18 岁以下人群的防护栏。
尽管 Meta 的审判将受到广泛关注，但法院的判决其影响可能比头条新闻所暗示的要小。烟草业在 1998 年的一项和解协议中同意在 25 年内支付超过 2000亿美元。但埃默里大学法学院的马修·劳伦斯（Matthew Lawrence）表示，此类结果在美国有着“令人失望的历史”。“我们可以减轻那些利用成瘾机制的行业所造成的伤害，但我们在寻找消除这些伤害的方法方面尚未取得巨大成功。” ■
好机器人，坏机器人
“互联网首页”能否在 AI 时代生存？
今年夏天，恐怖片《后室》（Backrooms）成为了有史以来最成功的独立电影之一，在预算仅 1000万美元 的情况下，票房收入高达 3.93亿美元。其素材来源十分独特：一组在互联网论坛 Reddit 上传播的、令人不安的空房间图像。此后，好莱坞的制片人们开始在该网站上搜寻灵感。
Reddit 近来日子不好过。其股价今年下跌了约三分之一。上个月强劲的季度收益未能让股东们感到乐观，他们担心随着用户转向人工智能聊天机器人获取信息，而非使用像谷歌（Google）这样将用户引导至其他网站的搜索引擎，该公司将遭受损失。目前， 正试图通过协议和诉讼，诱导聊天机器人制造商为其数据支付更多费用。但它最终希望，那些对机器人感到厌倦的用户能回归到这个互联网上最后的人类生成内容避风港之一。
的商业模式介于 Instagram（通过用户生成内容获利）和维基百科（依赖志愿者版主大军）之间。这种平衡有时很难维持。2023 年，在公司宣布将对其应用程序编程接口（API）收取访问费后，负责管理“子版块”（subreddits）的志愿者们进行了为期数日的罢工，而构建基于 应用的独立开发者正依赖于该接口。该平台坚持执行此项政策，可能是考虑到更大的客户 ▶
56 商业
► 需求。不久之后，它与谷歌以及 ChatGPT 的开发商 OpenAI 达成了协议，据报道，后者分别同意每年支付 6000万美元 和 7000万美元，以便利用 的数据训练其模型。
这些公司可能捡到了便宜。根据 委托研究公司 Profound 进行的一项分析， 已成为许多美国领先聊天机器人引用最频繁的来源。其数据的一个特别有价值的用途是生成产品推荐。 运营主管 Jen Wong 表示，该网站上约 40% 的对话与包括购物在内的商业活动相关。在像 t / BuyItForLife 这样的子版块中，用户讨论产品质量——并试图剔除赞助内容或机器人生成的内容——这使得其结果成为高质量评论的稀缺来源。AI 模型严重依赖此类内容来生成自己的推荐，而聊天机器人制造商希望将其转化为广告收入来源。
尤其是谷歌与 的关系，已变得不再像是一种双向交换。多年来，谷歌的排名搜索结果一直是该聊天网站的重要流量来源。但随着这家科技巨头将用户的注意力转向其 AI 概览（AI overviews），推荐流量大幅下降。上个月，在 表示“波动”的搜索推荐损害了用户参与度后，其股价经历了史上最大单日跌幅。根据数据提供商 Sensor Tower 的数据，上季度用户每天在 应用上花费的时间比一年前下降了 8%，而 Instagram 和 TikTok 则分别增长了 6% 和 4%。
的一个选择是与 AI 提供商讨价还价以获得更好的协议。它一直在与谷歌等公司协商新协议。如果这行不通，诉讼可能会奏效。 已经起诉了另一家模型制造商 Anthropic 和 AI 搜索引擎 Perplexity。两起诉讼均指控这些公司非法抓取 的数据。（Anthropic 和 Perplexity 表示它们没有违法。）在法庭上获胜——或获得丰厚的和解金——将使该聊天网站在数据许可谈判中拥有更大的筹码。
但从长远来看，Reddit 希望减少对其他科技公司的依赖。其首席执行官史蒂夫·哈夫曼（Steve Huffman）在上个月的财报电话会议上表示，来自搜索的流量“并非我们业务的核心所在”。相反，该公司希望鼓励更多用户直接访问其网站，并在进入后花费更多时间浏览。这将进一步提升其广告收入，该项收入在第二季度同比增长了 64%。黄女士乐观地认为，随着用户对与聊天机器人互动感到厌倦，Reddit 上（绝大多数由）人类创作的内容将对他们产生吸引力。有志于成为电影制作人的人们目前还不必担心素材匮乏。■
加拿大原油
关于油砂的转型
最新的繁荣能否持续？
“上帝啊，请再给我一次石油繁荣——我保证下次不会把它全部挥霍掉”，这是加拿大产油区一张流行的车贴。该国石油生产商最近的业绩表明，这一祈求得到了回应。由于通过霍尔木兹海峡的油轮通行受阻，油价大幅上涨。上个月，科诺哈斯能源公司（Cenovus）报告称，其最近一个季度的营业利润较去年同期几乎翻了三倍。加拿大自然资源公司（CNRL）则创下了史上最高的季度调整后营业利润。此外，加拿大政治阻力的减轻可能会增加此次繁荣持续的机会。
加拿大拥有世界第四大证实储量，其中 97% 位于内陆阿尔伯塔省偏远角落的油砂中。这里的沥青与沙子和粘土混合，几乎呈固体状，是通过开采而非从地下抽取的。其提取和加工成本高昂且污染严重。尽管存在这些缺陷，但高油价结合加元的贬值，意味着“很久以来，成为一名油砂生产商从未如此有利”，阿尔伯塔大学的安德鲁·利奇（Andrew Leach）表示。
将更多石油推向市场也将变得更加容易。加拿大前总理贾斯廷·特鲁多（Justin Trudeau）对新管道的支持态度冷淡，担心会对气候产生后果。2018 年，地区性的政治反对导致美国公司 Kinder Morgan 暂停了从阿尔伯塔省到不列颠哥伦比亚省西海岸的管道工程。这迫使联邦政府介入并将其完工。石油最终在 2024 年开始流动。
最近，政治氛围发生了变化。与美国（加拿大 90% 石油的目的地）的贸易战使联邦政府的重心转向减少对南部邻国的依赖。这包括支持建设将石油（和天然气）运往加拿大西海岸港口的新基础设施，以便将其运往亚洲。加拿大现任总理马克·卡尼（Mark Carney）致力于削减阻碍项目的繁文缛节，包括简化环境许可审批以及与原住民社区的磋商。这一更快捷的流程将适用于上个月宣布的、目前正在开发的一条通往不列颠哥伦比亚省的新管道。行业组织加拿大石油生产商协会的丽莎·拜顿（Lisa Baiton）谈到了一个“世代机遇”，可以集结政治和公众支持来推动新的石油项目。
然而，联邦政府对管道更加友好这一承诺可能还不够。能源基础设施公司 Enbridge 已暂停建设一条向东延伸的管道，理由是生产商缺乏提供足够石油以使其具有经济价值的承诺。
油砂的独特挑战提供了一个解释。自从 2014 年油价因需求疲软、美国页岩油激增以及 OPEC 增加产量而崩盘以来，对新场地的投资几乎不存在。卡尔加里大学的肯特·费洛斯（Kent Fellows）表示：“这次变动对该行业的影响怎么强调都不为过。”此外，油砂项目的开发时间远长于传统油井。它们需要巨额资金，在产生利润之前，这些资金可能会被占用五到十年。对未来油价的不确定性——以及政府更迭可能导致更严格监管重新实施的风险——使得这些投资变成了高风险赌注。
尽管如此，在 7 月举行的年度牛仔节——卡尔加里牛仔节（Calgary Stampede，该活动被视为行业的非官方晴雨表）上，石油公司举办的派对在经历了近年来的沉寂后，再次恢复到了最狂欢的状态。他们对喧闹的马车赛（chuckwagon races）的赞助打破了记录。这与任何其他迹象一样，都表明另一场繁荣正处于全盛时期。 ■
观看 The Insider
本周我们的主编采访了历史学家兼作家尤瓦尔·诺亚·哈拉里（Yuval Noah Harari）。欲观看视频，请访问 economist.com / insider
商业 57
虚招
针对数据中心的战争在某种程度上是虚构的
2月，在电影《马蒂·至尊》（Marty Supreme）中首次饰演笔业大亨后不久，凯文·奥利里（Kevin O’Leary）宣布将在犹他州建设全球最大的数据中心项目之一。市场对此不以为意。以主持《创智赢家》（Shark Tank）而闻名的奥利里先生对数据中心了解多少？资金如何筹措？电力如何供应？一名分析师将该项目斥为“虚幻的吉瓦级工程”之典型。
然而，市场上的笑话在公共广场上却成了导火索。6月，最初支持奥利里创业项目的犹他州参议院议长J·斯图尔特·亚当斯（J. Stuart Adams）要求将其规模缩减四分之三。奥利里指责批评者为中国工作（他随后撤回了这一说法），但最终同意缩减其梦想规模。即便如此，亚当斯这位资深的共和党议员以及两名县专员仍因该丑闻在初选中落败。
数据中心是政治经济学的一场噩梦。选民将其视为丑陋、肮脏、嘈杂且耗电的傲慢硅谷公司的堡垒。而且，得益于ChatGPT，向规划部门写信表达此类不满变得前所未有的简单。然而，公众在这些无窗棚屋上押下了巨大的筹码，因为股市是对其建设的单向押注。数据中心对民主党人来说尤其令人困惑，因为工会倾向于支持这些项目。如果说人工智能是一场阶级战争，那么它就是将亿万富翁与电工组成的联盟，对抗双方共同鄙视的办公室职员。
对于那些希望表现出对数据中心强硬态度但仍希望其得以建设的政治家来说，像奥利里这样的投机者是绝佳的软目标。8月18日，宾夕法尼亚州州长乔希·夏皮罗（Josh Shapiro）嘲讽开发商“恐吓我们的社区”，并签署了一项行政命令，他表示这将打击那些起初就几乎没有建设可能的“投机性提案”。他表示，在该州宣布的100多个项目中，申请开发所需许可的项目不足五分之一。
领取虚构的战果轻而易举。硅谷巨头是数据中心最大的支持者。但在他们身边，还排着长长的一队房地产开发商、私募资本公司、“新云”运营商和加密货币矿工，随后是国际投机者、国家级的无赖以及地方间谍。他们的绝大多数计划注定只能停留在新闻稿中。经纪公司伯恩斯坦（Bernstein）的分析师试图量化每个宣布的容量瓦数的可信度。谷歌和亚马逊的计划被视为可信。而两家大型新云公司CoreWeave和Nebius的计划则被打折约一半。奥利里则被直接忽略。
根据他们的模型，另一个幻想家是Vermaland，这家土地银行去年宣布在亚利桑那州建设一个巨大的3GW设施。据报道，5月该项目在面临当地反对后被削减了五分之四——这是一次“虚招”中的击倒。关于佛罗里达南部一个据称在2月份被搁置的1GW项目，几乎没有相关信息，仅知该土地在2024年由一家名不见经传的投资公司以$15m买下。当一个本就绝不可能实现的数据中心计划被削减时，它往往会发出巨大的声响。
政客们渴望在11月的中期选举前，通过刷出反数据中心的政绩来美化自己的形象。但由此产生的政策可能不会对这场繁荣产生太大影响。7月，纽约州州长宣布对大型项目（包括那些未连接到其电网的项目）实施为期一年的暂停令。但该暂停令不适用于已经获得批准的数据中心，而且纽约州对于整体建设而言并非至关重要。弗吉尼亚州大张旗鼓地对数据中心征收了6亿加元的电力税，但却保留了规模大三倍的销售税豁免。夏皮罗先生的新规定将要求地方当局在州政府批准之前先予以批准，这在选举结束前帮他清理了办公桌。
如果说数据中心的建设有一个天然的归宿，那就是德克萨斯州。该州地域辽阔，能源丰富，且拥有大量处理大型基础设施开发所需的律师和工程师。其电网运营商收到了惊人的474GW电网连接请求（目前该州面临的峰值需求约为90GW）。8月3日，德克萨斯州州长格雷格·阿博特宣布对数据中心项目进行“审计”，而这次审计预计将恰好持续到中期选举之后。值得注意的是，一些最大的数据中心运营商已公开支持他的计划。
更多的政治争议可能意味着尚未启动的项目将面临更多延迟。但尽管计划数量激增且政治反对声音增加，在过去一年中，数据中心规划中被延迟或取消的比例一直稳定在10%左右。到目前为止，市场带来的限制（如芯片和工人的可用性）比政府带来的限制更令开发商担忧。
这场略显虚假的反数据中心战争是否实际上是一场“伪战”，并随时可能转为真战？这很难想象。政客们很少愿意停止已经动工的项目。至于新项目，开发商在一定程度上可以用脚投票。到目前为止，这意味着向西迁移。这也可能涉及在联邦土地上建设更多设施。最终，它可能涉及在海外进行更多建设——或者，如果埃隆·马斯克的话可信的话，在太空中建设。
可停止的力量，可移动的物体
人们常说，中国在人工智能竞赛中拥有优势，因为它可以凭借强制力建设宏大工程。然而，美国则拥有通过夸大其词使愿景变为现实的力量。开发者的宏伟计划会被削弱，政治家们也会夸大他们的反对立场。通常情况下，这两者会相互抵消。与此同时，迅猛的建设步伐将继续进行。■
58
金融与经济
达龙·阿塞莫格鲁（Daron Acemoglu）
全球最具影响力经济学家的崛起
达龙·阿西莫格鲁（Daron Acemoglu）高产且受人推崇。然而，他的许多研究却显得出奇地缺乏说服力
当来自斯德哥尔摩的电话打来时，人们觉得这早已是水到渠成。2024年，来自麻省理工学院的达龙·阿西莫格鲁与西蒙·约翰逊（Simon Johnson）和詹姆斯·罗宾逊（James Robinson）共同获得诺贝尔经济学奖，以表彰他们关于制度如何塑造繁荣的研究。阿西莫格鲁先生年仅58岁，但一段时间以来，他一直是该学科的巨擘。他写了七本书，包括一本关于民主和人工智能经济影响的新书（见文化板块），这还不包括几本流行的教科书。他的数百篇论文（其中十多篇发表于今年）通常充斥着复杂的数学公式。在研究数据库 IDEAS / RePEc 发布的经济学家引用排名中，他与哈佛大学的安德烈·施莱弗（Andrei Shleifer）竞争榜首之位。他的观点在其他学术著作和媒体中具有很大分量。
然而，如果给经济学家们喝几杯酒，有些人会大胆地表达他们对这位巨头的真实看法。“他的许多理论研究很有用，但他利用自己的模型来支持那些曾经被尝试过且失败了的民粹主义政策，”一位知名经济学家如此抨击道。而一些评论员则不需要借酒壮胆。经济学博主诺亚·史密斯（Noah Smith）在回应针对阿西莫格鲁先生作品的一波网络批评时曾表示：“关于阿西莫格鲁，我简直已经大声疾呼了十年。”
没有人指责阿西莫格鲁先生的研究真正糟糕，更不用说学术不端了。“他显然是个天才，”另一位经济学家说道，并补充说他对博士生和初级同事非常慷慨。问题在于，他在经济学专业顶峰的声誉——以及随之而来的知识影响力——是否与其学术成就相匹配。
首先看他的实证方法。他的诺贝尔奖很大程度上基于2001年与约翰逊先生和罗宾逊先生共同发表的一篇论文，该论文被研究人员引用了超过23,000次。该文试图解释为什么有些国家富有而有些国家贫穷。在欧洲定居者大量死亡（可能是因为疾病）的地方，殖民者建立了“榨取式”制度，将权力和资源集中在少数精英手中，以便尽可能多地攫取资源。但在气候更宜人的地方，定居者生存了下来。殖民者建立了更具包容性的制度，拥有完善的道路、学校和医疗服务。这些制度得以延续，导致殖民时期定居者死亡率较低的地方在今天更加富有。
经济学家现在严肃对待这些结果，但并非照单全收。这篇论文鼓励他们更深入地思考制度，但他们也对其进行了剖析。诺贝尔委员会在2024年的报告中承认，死亡率数据“有时较为简略”。2012年，伊利诺伊大学厄巴纳-香槟分校的大卫·阿尔博伊（David Albouy）指出，在一些调整中，某些国家的死亡率是借用其他国家的。他发现，如果对这些进行修正，该论文的估算就变得不可靠。去年，研究机构 RWI-Essen 的马丁·布赫纳（Martin Buchner）及其同事报告称，他们调查的专家在某种程度上更倾向于支持阿尔博伊先生。
阿西莫格鲁先生表示，“这些讨论和一些批评极具价值。但他反驳说，阿尔博伊先生得出结论的方式是删除了原始样本中一半的数据，包括美国、加拿大和澳大利亚等重要国家。他说，正是这种组合加上一些统计选择，导致了不可靠性。他补充说，如果他今天重新撰写这篇论文，他会做出‘一些修改，包括在某些估算细节方面’”——但不会修改死亡率数据。
→ 本板块其他内容
59 债券市场的警告
60 关于中国就业的真相
61 贸易上的口舌之争
62 Buttonwood：谎言指数
63 自由交换：自由通行
金融与经济 59
最近，研究人员审视了阿塞莫格鲁（Mr Acemoglu）及其合著者在6月发布的一篇论文，该论文认为较低的出生率与劳动年龄成年人的人均GDP增长较快相关。这一惊人的发现似乎表明，像日本这样人口下降的地方不必过于担心。然而，包括宾夕法尼亚大学的赫苏斯·费尔南德斯-维拉韦德（Jesús Fernández-Villaverde）在内的其他经济学家质疑，在许多国家出生率崩溃的当下，历史关系是否真的能成为有用的指南。这是一个合理的观点——阿塞莫格鲁及其合著者在结论中也提到了这一点。但在阿塞莫格鲁海量的数学推演面前，这个观点很容易被掩盖，且显得不够深刻。
一个坚定的批评者几乎可以在任何人的作品中找到漏洞——而阿塞莫格鲁是经济学界最大的目标。然而，由名声引发的审视无法解释第二个问题：他的观点有时显得平淡无奇。他的公开评论，尤其是关于美国政治的评论，缺乏冒险精神。例如，他的“特朗普统一理论”断言，这位总统是通过“提升行政权力并摧毁约束性的制度和规范”来运作的。确实如此。
更令人担忧的是，他的宏大论题可能揭示的内容并不多。制度良好时国家繁荣，制度糟糕时国家停滞。没错。但究竟什么是“制度”？规则、规范、执行、文化——实际上涵盖了一切。制度又从何而来？“关键节点”和“制度漂移”。
乔治梅森大学的泰勒·科文（Tyler Cowen）在2024年评论阿塞莫格鲁的一本书时指出，书中描述的制度变革似乎仅源于其他制度变革，这使得核心论点陷入了无限递归——就像“下面全是乌龟”（turtles all the way down）。伦敦政治经济学院的邓肯·格林（Duncan Green）表示，该框架主要在事后分析时才有效。斯坦福大学的弗朗西斯·福山（Francis Fukuyama）则认为，该书回避了中国的例子，中国在制度可以被合理描述为“榨取性”的同时，实现了迅猛的经济增长。
阿塞莫格鲁反驳称，书中有两个章节部分或大部分是关于中国的。至于“乌龟”问题，他说：“要从制度视角理解制度变革，你必须考虑过去的具体制度以及影响它们的历史事件……我不认为这在任何方面是循环论证。”
在阿塞莫格鲁背离传统认知的程度方面，主要体现在他对技术持有极其阴郁的看法。他与约翰逊（Mr Johnson）合著并于2023年出版的《权力与进步》（Power and Progress）一书，将一千年的创新在很大程度上视为一个关于精英捕获和意外后果的故事。轧棉机是奴隶制的帮凶；哈伯-博施法制造了武器。显然，普通人只有在能够引导资本家考虑其利益时才能获益。该书淡化了一个更具说服力的主张：技术进步本身已使普通人的财富比工业化前时代增加了许多倍。
阿塞莫格鲁关于AI的论点也悲观到了失去可信度的地步。在2024年发表的一篇论文中，他计算出AI在十年内对美国生产力的提升仅有微小幅度。该论文得出这一结论的方式是忽略了新AI产品潜在的变革性影响。哈佛大学名誉校长（前财政部长）劳伦斯·萨默斯（Lawrence Summers）认为：“他在该分析中完全忽略了由于人工智能，我们将获得更快速的科学进步、更快速的社会科学进步或更好的决策能力的可能性。”
“这是一个合理的批评，”阿塞莫格鲁先生表示。“我没有预见到代理式人工智能（agentic AI），它没有被纳入我的预测之中。”他还指出，AI模型在科学研究中的用途可能比他想象的要大。但他依然坚持自己估算AI生产力效应的方法，并主张在那些关于“即将到来的巨大生产力提升”的主张中注入“一些现实主义”。
这些学术争论最终可能会产生重大影响。阿塞莫格鲁先生可能不认为AI即将为生产力提供强力加速，但他确实担心AI对民主和就业的影响。他的新书关注的是AI潜在的社会危害而非潜在益处。他提出的限制这种危害的方案——即开发能够增强人类劳动价值而非取代人类劳动的“亲工人AI”——听起来很棒，但同样显而易见。
阿塞莫格鲁先生在AI辩论中的影响力在最近的一份声明中可见一斑。这份由数十位知名经济学家签署的声明认为，“我们必须立即行动”，以“引导AI朝着补充人类并造福社会的方向发展”。但这个“我们”是指谁？又由谁来决定什么样的AI能补充人类，什么样的不能？甚至签署了请愿书的经济学家们也表示，他们并不完全确定。阿塞莫格鲁先生的名气如此之大，以至于有时会掩盖人们的批判性思考能力。■
政府债券
收益率
纽约
债券投资者成了全球市场的“扫兴者”
债券市场动荡不安，进而导致美国政府也随之不安。8月19日，财政部表示从下个月起将增加对长期债务的自有购买。德意志银行的分析师认为，这反映了政府对收益率“日益增加的焦虑”。
官员们有充分的理由担心借贷成本上升。8月18日，十年期美国国债收益率达到了2025年1月以来的最高点。更可怕的是，30年期债券收益率一度突破5.3%，创下2007年以来的最高纪录。美国并非孤例。在大多数发达国家，债券持有者都在要求政府支付更高回报（见图表）。英国、法国和德国的长期债券收益率均达到了十多年未见的水平。长期在主要经济体中处于最低水平的日本30年期收益率，正接近历史最高点。虽然抛售在8月19日有所缓解，但收益率短期内不太可能大幅下降。问题出在哪里？
一方面，与股票投资者不同，债券交易员似乎在关注新闻。霍尔木兹海峡在很大程度上仍处于关闭状态，且看来短期内将维持现状。美国的燃料价格——尤其是许多商业运输所依赖的柴油价格——随之飙升。美国银行每月调查的基金经理（其中大多数重仓股票）对此持乐观态度。平均而言，他们预计全球原油基准布伦特原油在年底前将交易在每桶$76，仅比战争开始前略高。而债券市场似乎没那么确定。
因此，它们正在将持续的通货膨胀计入定价。8月19日，英国重新- ▶
走势陡峭
30年期政府债券收益率，自2020年3月9日以来的百分点增幅*
来源：LSEG Workplace “黑色星期一”股市低点
60 金融与经济
► 随着能源价格上涨产生影响，英国截至7月的年度消费者价格通胀率升至2.9%，高于6月的2.6%。在美国，核心通胀率（不含食品和能源）上月有所缓解。但在美联储新任主席凯文·沃什（Kevin Warsh）近期发表言论后，债券交易员似乎削减了对未来加息的押注。
另一个近期的担忧是，政府债务发行者面临着新的竞争。近几个月，大型科技公司已发行了约750亿美元-worth的债券，用于资助人工智能数据中心的投资。高盛（Goldman Sachs）银行估计，这一波发行热潮已使其合计债务发行量接近去年的两倍。各类企业都在参与这场债券盛宴。高盛计算得出，在投资级发行人规模较大的债务发行（即超过100亿美元）中，约40%来自科技行业以外。因此，债券买家拥有过剩的投资级选项可供选择。许多人认为，人工智能热潮将通过增加资本竞争来推高利率。
尽管如此，收益率攀升的最主要原因由来已久：对政府债务和赤字的担忧。8月19日，美国财政部表示，联邦债务首次突破$40mn（占去年GDP的130%）。政府赤字约为GDP的6%。此类担忧在各国之间的差异中也得到了体现。法国和德国十年期收益率之间的利差达到了2012.以来的最大值。债券市场现在预计日本债务的收益率将高于中国债务，这颠覆了传统的顺序。在法国和日本，投资者认为政治家缺乏实质性削减开支或提高税收的意愿。
政府几乎没有其他更好的选择。在6月的美联储会议上，其利率设定委员会成员听取了关于国债持有结构如何从“对价格相对不敏感的官方部门持有者转向对价格更敏感的私人投资者”的汇报。这可能会增加债券持有者对长期债务所要求的溢价。为了应对这种压力，美国、英国和日本增加了低收益短期债券的销售。但结果是，其债务存量将更频繁地滚动，从而增加了这些时刻恰逢高利率的风险。
第二个不理想的选择是央行回购更多债务。但其中许多央行希望缩小其资产负债表，而非扩大。因此，它们可能不愿过度依赖债券购买作为降低收益率的工具。所有这一切意味着债券市场可能会继续保持谨慎。像财政部那样的干预措施暂时可能有所帮助，但不太可能长期安抚买家。■
中国的失业问题
具有中国特色的古德哈特法则
无论如何衡量，就业市场都十分疲软
引导经济从来都不是一件容易的事。在1975年的一次坦诚演讲中，时任英格兰银行官员的查尔斯·古德哈特（Charles Goodhart）分享了英国在英镑于四年前脱离加元挂钩后一些“不愉快”的经历。该银行当时希望利用利率来控制以各种方式衡量的货币供应量，而不是捍卫汇率。但货币当局忽略了演讲者戏称为“古德哈特法则”的规律：“任何被观察到的统计规律，一旦被用于控制目的，往往就会失效。”
其他人对此有更简洁的表述：当一个衡量指标变成目标时，它就不再是一个好的衡量指标。这一法则给中国带来了特殊的困难，因为中国有许多经济目标，因此缺乏好的衡量指标。例如，在2018年，中国开始发布一项新的城镇失业率月度指标，该指标与其他地区的类似统计数据一样，基于人口调查。但与其他国家的数据不同，中国的这一指标保持得异常稳定。8月17日公布的7月份失业率为5.2%，这很典型。在过去的115个月中，该数值低于6%但至少在5%以上的次数已达19次。即使在新冠疫情期间，美国失业率攀升至两位数，中国的这一指标也从未超过6.2%。
原因之一可能是该调查仅涵盖城镇劳动者。流入和流出中国城市的农民工使得城镇劳动力市场具有一种稳态特性。当就业创造强劲且失业率即将下降时，会有更多人出现，从而扩大了分母。相反，当劳动力市场疲软时，失业者会回到农村寻找工作，从而退出调查。
该指标停滞不前可能还有第二个原因：为了控制目的而施加的压力。自2018年以来，政府为调查的城镇失业率设定了年度目标，通常在5.5%左右，2020年升至6%。这使得官员们有动力通过各种手段将失业率维持在该数值附近。咨询公司Gavekal Dragonomics的安德鲁·巴特森（Andrew Batson）在最近的一份简报中认为，“有一种可能性，虽然无法证明但难以排除，即调查的失业率在某种程度上被操纵以降低波动性”。
因此，失业率的稳定性是一种虚假的安慰。它使得解读这个全球最大的就业市场变得毫无必要地困难。一些经济学家怀疑，中国的出口繁荣并没有产生同样强劲的制造业就业。另一些人认为，人工智能可能会消灭一些入门级的白领工作。另一个担忧是，工资增长不足以阻止经济重新陷入通货紧缩。所有这些怀疑如果能——
无论按任何标准衡量，就业都十分稳定
金融与经济 61
一项令人不悦的衡量指标
中国，城镇失业率，百万
来源：国家统计局；Haver Analytics
► 如果中国的失业数据像其经济一样多变，那么验证或排除这些疑虑将会更容易。
相反，经济学家必须依赖各种或多或少不尽如人意的替代指标。巴特森先生发现了一个新的——或者更准确地说，是一个被重新激活的指标。他建议重新审视一个较旧的衡量标准，而现代基于调查的数据本应取代它。这项指标自1978年起就开始收集，仅统计那些在当地政府登记为失业的人员，这是申请失业救济金的第一步。几十年来，这曾是政府的一个目标，而且稳定得令人惊叹。随着调查失业率的出现，它变得默默无闻。政府不再将其作为目标，甚至停止发布百分比率。但每年一次，登记失业人数的原始数据仍然出现在官方数据库中，尽管存在严重的滞后。该数字在2025年底跳升至1270万，比前一年增加了近16%（见图表）。
必须谨慎解读这个总量。2020年，就在疫情爆发前，规则的变更使得登记变得更容易，尤其是对于那些远离出生地居住的人。但巴特森先生认为，没有任何规则上的微调可以解释近期登记人数的增加。相反，这种增长似乎代表了“劳动力市场的真实压力”。
如果真是这样，那么这个曾经死板的统计数据中出现的不寻常跳升，可能也说明了一个更深层的原理。“当一个指标不再是目标时，”巴特森先生建议，“它就重新变成了一个好的指标。”可以将其称为古德哈特定律（Goodhart's law）的巴特森推论。当然，如果登记失业人数吸引了太多新的关注，政府可能会停止更新它。这种先例已有之。当青年失业率数据在2023年成为批评的焦点时，政府通过暂停发布数据解决了这个问题。追踪中国经济从未简单过。当一个指标成为目标，它就不再是一个好的指标。而当一个指标变得过于尴尬，它就直接消失了。■
贸易政治
垃圾话
新加坡
白宫关于中国转运“骗局”的荒谬主张
在 G.K. 切斯特顿（G.K. Chesterton）关于围栏的寓言中，那些急于拆除看似无用结构的人被敦促去理解这些结构最初为何而建。“切斯特顿的围栏”长期以来被一定年龄段的美国保守派视为常识，但唐纳德·特朗普（Donald Trump）的首席贸易意识形态专家彼得·纳瓦罗（Peter Navarro）似乎忽略了这一点。在纳瓦罗先生的白宫办公室发表的一篇题为《伟大的转运骗局》的新报告中，从日本到约旦的数十个国家被指控协助中国策划规避美国关税的阴谋。尽管这一主张十分荒谬，但该报告仍然生动地展示了特朗普政府在贸易问题上混乱的思维。
首先是“大重新分配”，这是纳瓦罗先生对特朗普先生在 2018. 对中国商品发起第一波关税战后所引发的转移的称呼。当时，自由贸易支持者警告称，中国与世界其他地区之间关税税率的差距将产生反作用。这将创造一种不可抗拒的动力去规避关税（无论是合法还是非法手段），同时抑制增长。美国在 20 世纪 40 年代作为基于非歧视原则的全球贸易体系的主要制定者时，对此深有体会。
但特朗普先生依然强行推进。中国与世界其他地区之间的平均关税税率差从 2018 年初的 0.9 个百分点扩大到去年的 29 个百分点（见图表）。预料之中的事情发生了：美国从中国进口商品的份额从 21% 骤降至 9%，而从其他国家进口的份额则大幅飙升。在越南、马来西亚和墨西哥等“连接”经济体中，从中国进口的商品和向美国出口的商品同步增长。特朗普先生和纳瓦罗先生现在面临的是一个他们自找的问题。
纳瓦罗先生将追求较低关税称为“‘伟大转运骗局’背后的金融引擎”，这呼应了特朗普先生去年威胁要对“转运”（一个定义模糊的术语，通常指通过第三国重新路由的中国商品）征收 40% 关税的说法。然而，调整供应链以改变货物的原产国并降低其关税责任并非骗局——这与简单地贴上新标签来掩盖原产地截然不同。
根据美国贸易当局的说法，传统的法律测试标准是“实质性改变”，要求在“形式、外观、性质或特征上发生根本性变化”。这并非一个清晰的界限。大多数自由贸易协定都为每条产品线包含了经过艰苦谈判的“原产地规则”（ROO）。新加坡智库亨利基金会（Hinrich Foundation）的德博拉·埃尔姆斯（Deborah Elms）指出，针对跨太平洋伙伴关系协定（一项由 12 个国家参与、但在 2017, 被特朗普先生否决的贸易协定）的原产地规则讨价还价就花费了近十年时间。
即便纳瓦罗先生也草率地承认了“合法制造和实质性改变”与“过境贸易和原产地转移”之间的区别。事实上，他汇总了来自私营和公共部门的五项关于“非法转运流量”的估算，认为每年高达 3030亿美元——这相当于 2018 年美国从中国进口额的 56%，一个令人难以置信的比例。然而，即便在这些估算中最具说服力的一项——高盛（Goldman Sachs）的一项分析发现 2023, 有价值 400亿美元 的中国商品被“表面上重新出口”到美国——也并未声称这些流量是非法的。（诚然，其中一部分将属于海关欺诈，美国有权对此进行打击。特朗普先生的贸易欺诈工作组声称，自去年启动以来，已从逃税者手中追回 10亿美元。）
如果纳瓦罗先生仅是对非法行为感到焦虑，那么第三国建立工厂并使用中国零部件制造商品的行为应该不成问题。然而，他却抨击那些“未必申报为中国原产”，但涉及“中国原产的投入或零部件、中国所有权或融资、与中国供应商或制造商的关系、基于中国的生产步骤 [或] 中国原产的路由历史”的商品。
这是一个极其宽泛的反对理由。现代制造业被构建为一个全球价值链网络，将生产每个阶段的产能分布在不同国家。但如果与中国投入品或企业的接近被视为参与反美骗局的证据，那么美国就是在与贸易本身开战。埃尔姆斯女士表示，这是对贸易的“重新定义，使得一切都变成了转运”。■
贸易违规
美国，贸易加权平均关税，%
来源：彼得森国际经济研究所
62 金融与经济
纳瓦罗先生的论文描述了横跨 43 个国家的“影子转运网络”的“功能架构”，该网络构成了美国非中国进口额的 70% 以上。从东南亚的组装中心到阿拉伯联合酋长国和加拿大的物流中心，这场所谓的骗局穿梭于“生产端的转型主张、物流端的路由通道、加工区、海运门户、陆路走廊、保税仓库和重新开票系统”之间。这在很大程度上就是全球贸易的全部基础设施；对其进行监管将需要一个全景监狱。尽管如此，纳瓦罗先生仍愿意尝试。他含糊地威胁要建立一个由 AI 驱动的“侦探边境”系统。
在经历了这一切之后，你可能会好奇纳瓦罗先生是否认为进入美国的进口商品根本不应包含任何中国成分。如果如此，含量应该是多少？一个极端（尽管逻辑自洽）的答案是设定一个中国成分的最高阈值，并禁止或对所有突破该阈值的进口商品征税。那么，特朗普先生 40% 的转运关税就可以用来强迫那些出口产品中中国成分过多的国家。
这种情况尚未发生，大概是因为这样做会推高物价，而美国公众已经对生活成本感到愤怒。然而，纳瓦罗先生这篇非同寻常的论文发挥了某种作用。它进一步说明了他的老板对全球贸易秩序的攻击所带来的混乱。■
BUTTONWOOD
谎言、该死的谎言与股票市场
基准指数正与股票市场的现实脱节
对华尔街持谨慎态度的政治家和志同道合的评论员们没完没了地重复着：股票市场并非实体经济。美国最大企业的标普500指数（S&P 500）今年上涨了12%，一次又一次地创下历史新高。与此同时，他们指出，尽管美国GDP增长稳定且失业率相当低，但消费者的情绪比密歇根大学自1952年开始追踪以来的几乎任何时期都要低迷。
这些都是事实——而且令人作呕地熟悉。较少被提及的事实是，标普500指数和全球其他基准指数也不再是真实的股票市场了。它们不再镜像反映国内股票的整体情况，而是日益地反映少数几家企业巨头的命运，而这些巨头的兴衰则主要取决于人工智能革命的不稳定前景。这使得许多基准指数——以及由此延伸的各地投资者的股票投资组合和养老金账户——波动性显著增加。
以标普500指数为例。其最大的成分股英伟达（Nvidia）占其价值的8%。假设这家市值5.3万亿美元的芯片设计公司的股价上涨或下跌3%（今年已发生过30次），而其他499只股票持平。那么该指数就会在同一方向上移动0.25%。在标普500的历史上，没有任何单一股票具有如此巨大的影响力。
然而，与该指数相比，它在台积电（TSMC）面前不过是个轻量级选手，而台积电正是英伟达AI处理器的制造商。这家半导体巨头的市值在过去一年中翻了一番，达到2万亿美元，占台湾证券交易所加权指数（TAIEX）的40%以上。为了让投资者分享台积电的成功，该国市场监管机构去年取消了一项执行20年的规定，即追踪指数的交易所交易基金（ETFs）在单一股票中的资产持有比例不得超过30%。结果，由1,000多只股票组成的TAIEX表现得就像一个仅由六家公司组成的等权重指数（如果你考虑到联发科和富士康等其他台湾大公司同样在服用与台积电相同的“AI兴奋剂”，且其走势因此部分相关，那么公司数量甚至更少）。
韩国的情况仅稍好一些。同样的金融数学计算（对于好奇者来说，这涉及计算成分股的赫芬达尔-赫希曼指数的倒数）表明，拥有830只成分股的KOSPI基准指数表现得就像除了该国两家强有力的内存芯片冠军——三星电子（Samsung Electronics）和SK海力士（SK Hynix）之外，仅包含11家公司。这两家公司在KOSPI中的合计权重在6月份曾短暂超过40%，而此前几年大约在四分之一左右。（韩国人对杠杆ETF的喜爱——利用借入资金来放大收益，但也会放大损失——增加了KOSPI的跳跃性。）
在5月至7月期间，随着交易员不断重新评估包括内存芯片在内的全球AI支出规模，KOSPI的波动率是去年平均水平的四倍。平均每三个交易日就有一次波动幅度达到5%或更多。7月28日，它下跌了11%。三天后，它飙升了18%。在2026年的前六个月里，韩国交易所不得不暂停交易五次，并近30次采取限制措施，次数超过了全球金融危机顶峰的整个2008年。
集中度和波动性并不必然并行。瑞士广泛指数SPI的价值近一半集中在五家公司：罗氏（Roche）、诺华（Novartis）、雀巢（Nestlé）、ABB和瑞银（UBS）。但它们的行业各异，且具有瑞士上市公司的代表性。几乎每两家瑞士上市公司中就有一家从事药物、食品或饮料、机械或贷款业务。相比之下，在台湾和韩国，分别只有六分之一和200分之一的企业生产芯片或其他技术硬件。
瑞士市值最高的三只股票同样是那些拥有稳定收益和定期分红的稳健型股票——因此，它们天生就比科技股更不容易波动。AI 将改变它们，就像它将改变所有行业一样。但它们的命运并不取决于 AI 资本支出周期的变幻莫测。
就像冠军腰带一样，主导地位可能会被丧失。诺和诺德（Novo Nordisk）曾凭借 Ozempic 掌控丹麦股市，但随后由于礼来公司（Eli Lilly）更出色的肥胖症药物抢占了其市场，导致其在指数中的权重下降。这提醒人们，股票市场的目标是捕捉投资者对未来股权现实的最佳集体猜测。只有当未来很可能与现在相似时，它才会反映现状。在 AI 时代，以及大量私人资本喂养着一长串基准指数所忽略的非上市竞争者的情况下，情况正日益变得并非如此。
金融与经济 63
《幻影收费站》
航行自由值得维护
位于德国考布（Kaub）村附近的普法尔茨格拉芬施泰因城堡（PFLZGRFENSTEIN CASTLE）曾拥有一个有利可图的地理位置。这座堡垒坐落在连接西欧最繁荣地区的莱茵河的一座岛屿上，且恰好位于该河最浅的水域之一。由于经过此地的航道极其狭窄，城堡可以从城墙上放下铁链以阻止通行并要求付费。普法尔茨格拉芬施泰因是一座收费城堡，而莱茵河沿岸曾有许多类似的城堡。
莱茵河的过路费是经济租金（economic rent）的一个经典案例，即某个实体通过使廉价资源人为稀缺而能够榨取的款项。对于通常控制该城堡的普法尔茨选帝侯来说，这条河流完全不需要任何成本，因此他们通过阻止他人溯河而上所赚取的过路费纯粹是租金。在最近的一次海湾战争之后，这类收费有回归的风险。伊朗政府正考虑对穿越霍尔木兹海峡的船只征收过路费；在2月份战争爆发前，全球约 20% 的石油供应经过该海峡。位于海峡另一侧的阿曼可能会加入其中。美国也可能想参与其中：唐纳德·特朗普总统表示，他的国家应该考虑以“霍尔木兹海峡守护者”的身份征收费用。
你可以将此称为“寻租之海的政治经济学”，这是在借用美国经济学家安妮·克鲁格（Anne Krueger）于 1974. 年发表的一篇论文的标题。贸易限制创造了租金，进而产生了一种新型竞争，在这种竞争中，不同群体争夺的是捕获这些租金，而非从事任何生产性活动。克鲁格女士在“寻租社会”中观察到的贸易限制通常意味着游说、贿赂或法律诡计。而对于向过往海上交通收取过路费而言，寻租活动可能要致命得多：海盗、无人机等等。在帆船时代，领海止于距离海岸三海里（5.6公里）处，这大致是一门大炮的射程。这是一种对主权延伸范围有限的认可。
然而，在过去的几个世纪里，航行自由一直是常态。这通常是由能够维护其主权或将其影响力延伸至海岸线之外的海军强国提供的公共产品。普法尔茨格拉芬施泰因城堡的收费在 1867 年被废除：普鲁士政府控制了河流两岸，终于能够履行多年前做出的外交承诺。租金榨取就此结束。1868 年，一项国际条约废除了莱茵河上的所有过路费。其他地方也发生了变化。多年来，丹麦王室通过对进入波罗的海的船只征收过路费来获取大部分收入，直到 1857 年大国强制终止了这一赚钱计划。
在公海上，航行自由由能够将权威投射到任何地方的霸权国家保证。在 19 世纪和 20 世纪的大部分时间里，霸权国家是英国，尤其是在波斯湾。从 1820 年开始，大英帝国利用打击海盗来强迫各埃米尔签署条约，他们的领地成为了休战国（Trucial States）。帝国在没有正式将其纳入版图的情况下确保了对航道的访问权。1971, 年后，随着英国从苏伊士以东撤军，接力棒传给了美国。海湾国家获得了独立，但也处于美国的保护伞之下。美国接管了位于巴林的前英国海军基地。
因此，保持水道免收过路费是查尔斯·金德尔伯格（Charles Kindleberger）——另一位美国经济学家——所提出的霸权稳定理论的一个例子：为了保持良好秩序，全球经济需要某些公共产品。霸权国家是唯一既有能力又有意愿提供这些产品的实体；作为回报，它从运行良好的世界经济中获取部分巨额收益。
金德尔伯格在论述大萧条时，主要关注的是金融稳定，但航行自由同样符合这一模型。在他的理论及其后续版本中，只有霸权国家既能承担提供此类公共产品的成本，又能获取大部分利益：美国国务院的官员在 1969 年写道：“我们的经济利益在于，以尽可能低的成本，使商船获得最大程度的行动自由。”如今，美国正在退出这一角色，部分原因是缺乏意愿——特朗普先生不愿向海湾地区派遣军队——部分原因是缺乏能力。廉价的无人机能够比昔日三英里射程的大炮更远、更廉价地投射力量。
是否存在第三种选择，既非无政府状态，也非霸权统治？或许其他国家的集合——即加拿大总理马克·卡尼（Mark Carney）所指的“中等强国”——可以通过结盟来提供公共产品。保证莱茵河免收通行费的条约是由一组德国各州、法国和荷兰共同签署的。但去年由三位常驻德国的经济学家安韦莎·班纳吉（Anwesha Banerjee）、奥特马·埃登霍费尔（Ottmar Edenhofer）和乌尔里克·科尔内克（Ulrike Kornek）发表的一篇论文则表达了怀疑；国家之间更均衡的力量对比会鼓励搭便车行为和地位之争。因此，中等强国在维持霍尔木兹海峡免收通行费方面可能无法取得更好的效果。
历史不会重复，但它会押韵（Rhine）
然后是真正的全球公共产品。莱茵河是否征收通行费，目前看来是一个无关紧要的问题。欧洲持续的热浪和干旱已进入第三个月，导致考布（Kaub）的水位降至仅 6 厘米，低于 2018 年 25 厘米的前纪录。这勉强够孩子们在里面戏水，更不用说让载重油轮通过了。咨询公司牛津经济研究院（Oxford Economics）估计，这将使德国今年的 GDP 下降 0.2 个百分点。危险不仅在于各国设法从曾经免费的事物中榨取租金，而在于世界整体提供公共产品的能力正在下降。■
64
科学与技术
全球变暖与航空
更蓝、更凉爽的天空
一项跨大西洋的试验旨在展示，通过无凝结尾迹飞行能在多大程度上冷却地球
2019 年计划执行的 4000 万次客运航班累计飞行距离约为 610 亿公里——相当于每天都飞行一次地球到太阳的距离。其中 40% 以上的航班在旅程的至少一部分时间里留下了凝结尾迹，这是因为发动机排出的水蒸气凝结成了微小的冰晶颗粒。
其中许多尾迹是短暂的。它们在母机身后延伸，随后迅速消散。但在总飞行距离的约 5% 范围内（足以支撑每天十次往返月球的距离），这些尾迹会持续存在并扩大，因为大气中原有的水蒸气也加入了冻结过程。纤细的白线变成了宽阔的半透明卷云条带。将这些条带像毛线绕球一样绕在地球周围，你会发现，在任何一个时间点，持久的凝结尾迹可能覆盖了千分之一的天空。
在这种情况下，它们使地球升温。
凝结尾迹呈现白色，是因为其组成的冰晶会反射阳光。这在白天会产生冷却效果。但这些晶体也会吸收红外辐射并将其重新发射，就像温室气体一样。一部分辐射进入太空，但其余部分则返回地球。这产生了 24 小时的持续升温，总体而言，这种升温效应明显超过了冷却效应（对于非凝结尾迹产生的卷云也是如此）。净升温的准确量很难计算，但目前的证据
→ 本节其他内容
66 探测南极冰层
66 恐龙的胃石
67 一种新的亚原子粒子
67 资讯通：工作时，动起来！
表明，飞机的凝结尾迹及其诱导产生的卷云，在提高地球温度方面的作用可能比飞机发动机排放的二氧化碳还要大。
不过，这种升温在某种程度上是可选的。凝结尾迹要持续存在，其所处空气必须远低于冰点，且含有大量尚未冻结的水蒸气。阻止这些水蒸气冻结的原因是缺乏能够作为晶体形成核心从而“播种”该过程的颗粒。凝结尾迹提供了数万亿个这样的种子，从而在原本没有云的地方创造了云。
问题的核心
这类冰超饱和区域（ISSRs）通常出现在 8-13km 的高度。它们的横向跨度达数百公里，但深度仅有数百米。只要让飞机避开这些区域，就能消除大量持久的凝结尾迹——这意味着也能消除一部分变暖效应。
8月18日，一个总部位于英国的财团宣布启动“蓝天行动”（Operation Blue Skies），这是迄今为止在气候冷却方面开展的最重要的一次此类方法试验。该行动将在尚尼克（Shanwick）进行，这是一块覆盖大英群岛以西北大西洋 1.8m 平方公里的空域（见次页地图）。
科学与技术 65
此前已进行过规模较小的此类构想试验。其中一项由德国航空航天局在 2021 年运行，另一项则在 2025, 年由美国航空（American Airlines）和 Contrails.org 运行，后者是一家主要由比尔·盖茨创立的投资者网络“突破能源”（Breakthrough Energy）资助的非营利研究机构。结果令人乐观。规模大得多的“蓝天行动”将揭示，预测并绕行 ISSRs 的方法是否既有效又适用于实际操作。
尚尼克是一块非常适合此类项目的空域。在客机的巡航高度上，这里容易出现恰到好处的阴冷潮湿环境，且有大量飞机经过。这些因素综合起来意味着，尽管它仅覆盖地球表面的 0.4% ，但模型显示，尚尼克区域导致了约 5% 的凝结尾迹相关变暖。此外，欧洲和美国的气象卫星在可见光频率和红外线波段昼夜监测该区域，使得观察其动态变得十分容易。
冬季故事
计划是，在即将到来的这个冬天以及接下来的那个冬天，在 20 到 40 天的时间里，负责管理英国领空的 NATS 公司的空中交通管制员将修改经过尚尼克（Shanwick）的飞机航线，以使其避开疑似的 ISSRs（冰晶超饱和区）。在条件允许且没有理由不这样做（例如出现湍流）时，这些调整将使飞机的飞行高度比飞行计划低几百米。
通过对比实施这些调整期间与未实施期间产生的凝结尾迹，应该能提供可靠的数据，以衡量凝结尾迹减少了多少。伦敦帝国理工学院 Blue Skies 项目团队成员塞巴斯蒂安·伊斯塔姆（Sebastian Eastham）表示：“最理想的结果是，卫星图像能够一次又一次地、稳健且可靠地显示凝结尾迹在减少。”
除了 NATS 和帝国理工学院，Blue Skies 背后的财团还包括 Contrails.org、谷歌（除了提供软件专业知识外，谷歌还为 500 万英镑的预算贡献了价值 140 万英镑，即 100 万美元的硬件和计算时间）、剑桥大学以及英国气象局。
如果凝结尾迹规避要成为主流，Blue Skies 的试验不仅需要证明采取规避措施的日子里凝结尾迹更少，还需要证明将该流程整合到常规运行中可以安全且简单地完成。这将为更广泛、并最终实现日常化使用铺平道路。
这种做法不必普及到所有航班才能发挥作用。凝结尾迹引起的变暖是集中的。据一项估计，对气候产生升温影响的 80% 来自仅 2% 的航班。
尴尬的是，你无法提前准确知道这 2% 是哪些航班。但 Contrails.org 的负责人马克·夏皮罗（Marc Shapiro）表示，一个能消除全球 70% 凝结尾迹变暖的系统，可能只需要修改不到 5% 航班的飞行计划。
根据剑桥大学的杰西·史密斯（Jessie Smith）及其同事最近发表的一篇论文，如果航空业不改变其方式，到 2050 年，它将导致全球温度较工业化前水平总共上升约 0.01℃。如果届时全球升温约 2℃，那么航空业的贡献率将达到 5%，其中超过一半是由凝结尾迹引起的。一个从 2035 年开始且有效率达 75% 的凝结尾迹规避计划将使总升温量减少约三分之一，尽管这其中存在很大的不确定性。
虽然这些比例看起来很小，但史密斯博士及其同事认为，其效果与任何其他旨在减少航空运输导致变暖的单一措施相当——而且实现速度可能更快。
这些收益并非没有代价。在较低的高度飞行会增加燃油使用量，这对航空公司来说是一项成本，同时也会增加二氧化碳的排放量，这对地球来说是一项成本。但这两项成本都不是巨大的。模型估计表明，规避 ISSRs 仅使飞行时间略微延长，且增加的燃油使用量不足 2%。
在 5% 的航班上多使用 2% 的燃油，将使整个行业的总燃油账单增加千分之一。在美国航空公司的试验中，被分流航班与未被分流航班之间的燃油使用差异无法被检测到。
类似的计算也适用于变暖。气候科学家奥利维尔·布舍（Olivier Boucher）研究该问题已有 30 年，目前经营着一家名为 Klima 的法国小型专业公司，他表示：“如果你真正专注于那些影响最大的航班，也许每多排放一吨二氧化碳，就能节省大约 100 吨二氧化碳当量。”
此外还存在机会成本。一些人担心，关注凝结尾迹会分散人们对航空公司二氧化碳排放的注意力，以及分散对减少排放（通过使用可持续航空燃料，即 SAFs）或抵消排放（通过从大气中移除二氧化碳并将其进行永久存储）等努力的关注。如果不解决这些问题，或者不减少飞行量，仅仅消除凝结尾迹只能带来一次性的收益，但无法改变一个根本事实：从长远来看，飞行越多，变暖越严重。
理想是优秀的敌人
质疑者同样认为，可持续航空燃料（SAFs）可以通过设计来减少冰晶形成所需的颗粒物，从而在减少二氧化碳的同时减少凝结尾迹。这确实如此。但避开国际标准超饱和区（ISSRs）似乎是一种更具成本效益的凝结尾迹削减方法。史密斯博士及其同事发表的论文发现，在已经采取避让措施的系统中加入可持续航空燃料，所带来的降温效果最多仅为 0.001℃。
因此，世上没有免费的午餐。但在全球无数次飞行中的极少数航班上，让乘客多花一两分钟来品味他们付费购买的机上餐饮，而这可能会带来一些益处——这种想法在气候干预措施中已极具吸引力。■
有冰还是无冰？
按干预有效性预测的凝结尾迹抑制影响 全球地表温度增幅，°C
来源：J. Smith 等人，《凝结尾迹避让的气候机遇与风险》，《自然》杂志，2026年
66 科学与技术
冰川学
噪声中的信号
巨型拖拉机的轰鸣声或许能取代炸药，用于探测南极冰层
PISTENBULLY 300极地南极车辆是一款重达 8.5 吨的重型特种怪兽，专为承受极端的极地环境而打造。它配备了宽履带、巨大的牵引力、强力起重机以及为最艰苦的冰雪作业设计的通用工具。任何南极基地都离不开它。然而，由北京中国地质大学刘国峰领导的研究表明，PistenBully 的能力可能不仅仅限于搬运。正如刘博士在《冰川学杂志》（Journal of Glaciology）中所写，它还能进行真正的科学研究。
南极冰盖是地球上最大的淡水储库，但它正在萎缩。在冰层正上方、表面积雪下方，有一个厚 50-100 米的过渡带，称为粒雪层（firn layer）。这里是积雪被压缩成冰的地方，也是上方融水被捕捉在孔隙中并重新冻结的地方。了解粒雪层的精确厚度及其行为方式，对于预测冰盖的长期变化以及全球海平面的演变至关重要。然而，要将这种地下混合物可视化是非常困难的。
目前的做法借鉴了油气勘探，即使用“主动源”。用地质学家的术语来说，就是将一根炸药棒扔进钻孔，然后监听通过岩石（在这种情况下是冰）传播的爆炸声。监听工作由名为检波器的传感器完成。由于声音在水、冰和雪中的传播速度不同，通过计算机分析回声到达不同检波器的时间和强度，可以构建出粒雪层的高分辨率图像。
这种方法有效，但并不理想。首先，必须有人钻孔以便投放炸药，这在极地条件下并不总是可行。其次，必须购买炸药，将其运往南极并执行投放。此外，在敏感环境中引爆炸药绝不能轻率对待。因此，寻找另一种声源将大有裨益。在听到由 PistenBully 车队制造的喧嚣声后，刘博士认为它们可能正是所需要的。
为此，他及其同事利用东南极拉塞曼山（Larsemann Hills）已经部署的检波器进行了实验。他们征用了几辆 PistenBully，并安排它们在山丘周围行驶。他们发现，如果车辆沿着检波器阵线的两侧、与传感器平行地行驶，就能达到预期的效果。
由于拉塞曼山的粒雪层此前已用传统方法绘制了地图，刘博士能够将机器生成的图像与使用炸药获得的图像进行对比。两者几乎完全相同。因此，他似乎设计出了一种绘制粒雪层结构的方法，比将其炸掉更简单，但效果同样出色。■
恐龙胃石
飞行的胃
恐龙胃中的石头可能有助于理解鸟类的起源
如何知道恐龙吃过什么？内脏在死后会腐烂，因此恐龙消化的细节依然模糊不清。最好的指标是牙齿。尖锐的牙齿可能是为了撕下猎物的肉块以便整块吞咽。强健的研磨面则表明涉及大量的咀嚼。然而，许多属于兽脚类（theropods）的恐龙（鸟类由此演化而来）完全失去了牙齿，并进化出了喙。这使得研究它们吃什么以及如何消化变得十分困难。
不过，还有第二个线索。许多恐龙的胃里有石头，被称为胃石（gastroliths）。人们认为这些石头辅助了消化，因此它们本身就讲述着故事。现代鸟类和鳄鱼（它们与恐龙一样，都属于一个更大的演化组，称为主弓类动物 archosaurs）通常也拥有胃石。
日本冈山理科大学的高崎龙二（Takasaki Ryuji）决定对此进行研究。在一项发表在《古生物学》（Paleobiology）上的研究中，他及其同事检查了 46 种现代主弓类动物（42 种鸟类和 4 种鳄鱼）的胃石，并将它们与 16 种恐龙的胃石进行了对比。
这些鸟类包括食草动物、杂食动物和食肉动物。鳄鱼则全部是食肉动物。石头的形状各异。正如预期的那样，如果某种动物拥有肌肉发达的胃（如鸭子和鹅），其胃石往往会被磨成类似于海滩鹅卵石的圆润形状。胃部肌肉较弱的动物（如海鸟和鳄鱼）则含有更多棱角分明的石头。
在观察恐龙时，高崎博士及其同事发现，鸟臀类恐龙（如三角龙 Triceratops）——根据牙齿推断该类群为食草动物——保留了棱角分明的胃石。这意味着这些生物主要依靠下颌来研磨食物，而胃部仅负责最后的处理。
长颈的蜥脚类恐龙（如腕龙 Diplodocus）同样拥有棱角分明的胃石。这更像是一个谜题，因为这些动物拥有像铅笔一样的牙齿，据推测，它们用这些牙齿从树枝上剥下叶子，而颈部没有那么长的动物则无法触及。高崎博士怀疑，解释在于蜥脚类恐龙的胃像现代有蹄类动物一样，主要依赖发酵来分解食物。
兽脚类恐龙肠道中的石头则讲述了另一个故事。塔巴努龙（Tarbosaurus）是一种与霸王龙（Tyrannosaurus）相似的极食肉动物，同样拥有尖锐的牙齿，它拥有棱角分明的胃石。然而，拥有喙的兽脚类恐龙（科学界认可的至少有四个类群）则拥有圆润的胃石。事实上，高崎博士看到了牙齿缺失与石头圆润度之间的明确关系，因为研磨食物的工作从下颌转移到了胃部。
这种转变可能助力了鸟类在演化上的成功。牙齿很重，支撑这些牙齿工作的下颌肌肉同样沉重。这使得动物的前端承受了大量重量。将研磨工作以及相关的肌肉和坚硬表面转移到胃部，使动物的重心集中，从而使其更容易离开地面。
奇怪的是，化石记录清楚地显示，最早期的鸟类并没有喙。但如此多兽脚类亲属拥有喙，且鸟类自身至少两次进化出喙这一事实表明，这种转变是容易的。因此，成功飞向天空的部分秘密可能就在于，字面意义上的“有这个胃”（having the stomach for it）。■
科学与技术 67
粒子物理学
让它粘在一起
物理学家称他们发现了一种难以捉摸的粒子：胶球
一种新的亚原子粒子已从理论跃升为现实。北京正电子对撞机（BESIII）——一项中国的物理实验——的科学家们如此表示。8月5日，南京大学的金山在国际高能物理会议（ICHEP）上代表该团队宣布，他们发现了一个胶球（glueball）——这是一种此前仅存在于假设中的粒子，与物理学家此前见过的任何粒子都不同。
原子的原子核由一种被称为强核力的现象结合在一起。那么，对于负责传递这种力的粒子，还有什么比“胶子”（gluon）更好的名称呢？胶子与另一种基本粒子——夸克相互作用，将它们编组为双子或三子，从而创建质子、中子以及其他此类“复合”粒子。但胶子也会与其他胶子相互作用，这使得胶子将自身粘合在一起，形成另一种复合粒子的可能性成为可能：即胶球。
物理学家的理论预言了胶球的存在。但找到它则是另一回事，因为这些理论同时表明，胶球将很难从北京正电子对撞机 II（BESIII 依托于此设备）等机器所产生的大量更平凡的基于夸克的粒子流中区分出来。近半个世纪的搜寻让研究人员空手而归。
BESIII 的运行人员表示，这种情况不再发生了。2011年，一种名为 X(2370) 的新粒子出现在他们的探测器中。它看起来非常像胶球，但其他解释也是可能的。因此，团队开始致力于排除这些其他可能性。在随后的几年里，他们分析了超过 100 亿个 X(2370) 可能参与的物理过程。结果——于 7 月在网上公布并在 ICHEP 上由金博士演示——勾勒出了这种新粒子特性的完整图景。该团队现在确信它是一个胶球。“没有其他解释能同时解释所有这些特性，”金博士说。
探测到胶球——如果它确实被找到了的话——不仅仅是物理学家在他们如今已十分庞大的收藏中增加一个新样本的问题。它的发现将为测试强力背后的理论提供宝贵的数据。该理论虽然完善，但仍包含许多谜团。其中之一就是强力如何产生质量——因为胶子没有质量，但胶球有。在回答此类问题上，胶球的发现可能会起到决定性的作用。■
消息灵通
你应该在办公桌前站着吗？
久坐不健康，但改为站立并非良方
长期久坐确实名声不佳。2012年的一项综述得出结论，最缺乏运动的人患糖尿病或发展为糖尿病的可能性是运动最充分的人两倍以上，且面临更高的早逝风险。2015年的另一项研究发现，长时间久坐与“无论体力活动多少都会导致的不利健康结果”相关。三年后，《英国运动医学杂志》（British Journal of Sports Medicine）对新闻文章的一项分析发现，有数百篇报道引用了将久坐比作吸烟的说法。
最后一种说法过于夸张。2018年，《美国公共卫生杂志》（American Journal of Public Health）估计，久坐的真实风险仅为吸烟的十分之一。但这依然很糟糕。因此，站立式办公桌以及近来流行的“动态座椅”（active seats）才具有吸引力：这类椅子或凳子可以通过倾斜、摇摆或晃动，强迫使用者不断调整姿势。这两者真的更健康吗？
2018年的一项分析发现，在办公桌前站立每小时比坐着仅多燃烧9卡路里。心血管方面的收益同样不显著。2020年的一项综述汇总了涉及877人的九项试验。参与者每天多站立约1.3小时，随访时间约为四个月。体脂和血糖控制略有改善，但血压、胰岛素、胆固醇和甘油三酯没有变化。不过，该研究并未考察血流量和姿势肌肉的激活情况，而这两个因素可能会使站立比久坐更有益。
即便如此，站立也有弊端。2024年发表在《国际流行病学杂志》（International Journal of Epidemiology）上的一项研究对英国83,013名成年人进行了平均近七年的随访。每天站立超过两小时的人患静脉曲张和静脉溃疡的风险更高，这可能是由于血液在腿部淤积所致。
更为关键的是运动。2024年一项针对近482,000名台湾成年人的研究发现，每天仅增加15–30分钟的锻炼，就能消除大部分与工作久坐相关的超额死亡风险。因此，伏案工作者可能会好奇动态座椅是否能带来改变。一些研究结果令人鼓舞。
以乔治亚南方大学的一项试验为例。在这项试验中，20名成年人在普通椅子、健身球和安装在充气气囊上的座椅上分别进行阅读和打字任务。使用健身球时的能量消耗与正常久坐相似。但在气囊座椅上，参与者的心率和卡路里燃烧量分别上升了6–13%和19–40%。2019年，研究人员在《人体工程学》（Ergonomics）杂志上撰文，将此归功于维持平衡所需的肌肉收缩。
一项更长期的试验发现了另一个益处。在曼谷，133名办公室职员在六个月内分别使用泡沫坐垫或一种旨在鼓励频繁改变姿势的轻微不稳定充气坐垫。使用后者的职员出现颈部和腰部疼痛的可能性分别降低了81%和84%。2024年，研究人员在《斯堪的纳维亚工作、环境与健康杂志》（Scandinavian Journal of Work, Environment & Health）上将该坐垫描述为一种“轻量运动设备”。他们认为它减少了疲劳性肌肉收缩，并改善了受压组织的血流量。
这里存在一些限制。许多关于动态久坐的研究规模较小或时间较短。结果随座椅设计而异，且部分参与者报告臀部感到不适。尽管如此，这里似乎有一个启示：真正的区别不在于坐着还是站着，而在于保持静止还是在运动。
68
文化
人类最好的“敌友”
狗能教会我们关于 AI 的什么
两本书在思考这项技术将会是温顺的还是凶猛的
《人工智能国家的兴衰》（The Rise and Fall of the Artificial State） 作者：吉尔·莱波尔（Jill Lepore）。Liveright 出版社；336 页；$29.99。 Allen Lane 出版社；£25
《自由民主制怎么了？》（What Happened to Liberal Democracy?） 作者：达龙·阿西莫格鲁（Daron Acemoglu）。Dutton 出版社；416 页；$32。 Profile Books 出版社；£25
在思考 AI 将如何改变社会时，不妨考虑一项早期的技术，它为人类提供了一个能力卓越的得力伙伴：狗。通过驯化狼，人类能够间接地嗅到远方的猎物，并在夜晚听到敌人的逼近。他们成为了比尼安德特人更出色的猎人，也成为了更强大的战士：难以被伏击，难以被摆脱。一些人类学家认为，狗是智人能够胜过史前近亲并征服世界的原由之一。
在“狗技术”最初被开发时，人们很难断定这是否安全——因为人们简直是在邀请狼进入自己的洞穴。他们必须训练它们不要吃婴儿。数千年后，狗有时仍会逃跑并失控。但总体而言，人们对结果感到满意。（尼安德特人的结局则没那么好。）
对 的乐观观点是，它将像狗一样：是对人类劳动的补充而非替代，并且始终服从于主人。但如果并非如此呢？ 的进化速度快得多。很快，它可能会以狗类繁育者梦寐以求的速度，创造出自身的新升级版本。因此，人们可以确信狗不会奴役他们——到 2030 年，狗狗们依然不知道如何使用开罐器——但他们无法对 如此确定。
两本新书，一本由诺贝尔奖得主经济学家撰写，另一本由普利策奖得主历史学家撰写，都在探讨 将如何塑造未来的问题。两者都没有使用狗的类比，但参考这个类比会更容易理解。
哈佛大学历史学家吉尔·莱波尔关注的是那些“训练狗的人”。在她的叙述中， 大亨们是可怕且不负责任的人，他们的创造物可能会挣脱皮带并袭击人类。在《人工智能国家的兴衰》一书中，她认为科技巨头们正试图构建“一个人类被公司拥有的机器统治，且许多人被替代的未来”。她说，这意味着“自由民主制的被抛弃以及自然世界的毁灭”。
她的论点部分基于这些巨头自己的话。OpenAI 的负责人萨姆·奥特曼（Sam Altman）曾表示，用一个能够“走遍全球与每个人交谈，了解他们的确切偏好”并为所有人进行“优化”的 来取代人类总统将是“极好的”。埃隆·马斯克（Elon Musk）则表示，人类最终可能会像超级智能机器人的“宠物拉布拉多”一样。 ▶
→ 本节其他内容
69 标点符号的（巨大）影响！
70 一幅雷诺阿肖像画的迷人故事
70 向瑞典酿酒师致敬
71 日本聆听酒吧走向全球
72 被遗忘的开国元勋
文化 69
奥特曼先生曾暗示，地球的大部分表面可能很快会被数据中心覆盖。马斯克先生希望通过将数据中心置于太空——并将数百万人发射到火星居住来解决这个问题。莱波尔女士指出，这些科技巨头似乎都有逃离计划：“埃隆·马斯克去火星，马克·扎克伯格去元宇宙……杰夫·贝佐斯去太空，萨姆·奥特曼的大脑上传到云端。”她担心，这使得他们在面对于地球上创造反乌托邦的风险时变得鲁莽。
她是对的，如今的科技巨头确实不是你会信任其制定文明未来规则的人。但她高估了单个亿万富翁的重要性。 如何塑造社会，并不取决于马斯克先生古怪的政治观点，就像汽车的影响并不取决于亨利·福特一样。相反，这将取决于技术本身的力量以及人们如何使用它。
这是达龙·阿西莫格鲁（Daron Acemoglu）在《自由民主制发生了什么？》（"What Happened to Liberal Democracy?"）一书中探讨的问题。阿西莫格鲁是全球被引用次数最多的经济学家之一。他最著名的作品是《国家为什么会失败》（"Why Nations Fail"），这部宏大的经济史著作认为，导致贫穷国家持续贫困的主要因素是掠夺性制度。他的新作则聚焦于当下，以及技术与民主之间的相互作用。
他认为，自由民主制支撑了进步。它促进了创新和经济增长，通过税收和公共服务传播利益，并通过监管保护选民。
长期以来，这一制度在很大程度上满足了选民的预期。但“随后计算机出现了”。在阿西莫格鲁的叙述中，数字技术通过提高对脑力而非体力的需求，加剧了不平等。它们还增强了“大学毕业生的社会权力”，使得一个新的自由派精英阶层能够将其价值观强加给心怀不满的工人阶级，从而引发了导致唐纳德·特朗普当选的反弹。此外，它们改变了人们互动的方式。在过去 20 年里，美国人与朋友面对面交流的时间骤降了三分之二。政治辩论变得更加极端且充满戾气。
阿西莫格鲁（相当坦率地）担心，AI可以通过增强创新让每个人变得富有得多，但它也可能放大数字技术的危害。它可能导致大规模失业，在“创造者”和“获取者”之间造成“巨大的社会地位差距”，并引发一场全面的民主危机。（它还可能通过让恋爱机器人和政治造假变得更具诱惑力，从而加速社会原子化。）
他的建议是发展“亲工人 AI”。阿西莫格鲁所指的，是寻找让工人利用 AI 来提升其劳动价值的方法，就像史前猎人将制作和投掷长矛的技能与狗的嗅觉和速度相结合一样。
也许 能让电工在面对日益复杂的电网时能够排除故障，通过传感器发现问题，并利用以往故障和解决方案的数据库帮助人类进行分析。也许 能改善教育： 助手可以评估学生对测验的回答，并建议将班级分成更小的组，从而提供量身定制的指导，进而增加对人类教师的需求。
这些观点很难反驳。如果 是补充人类劳动而非使其冗余，那么它确实会引发较少的社会动荡。但阿西莫格鲁没有解决如何实现这一目标的问题。最终，这两本书都令人失望。它们在无意中揭示了，要说服性地描写一项变化如此之快且未来能力完全未知的技术，是多么困难。
在政治上， 已变得不受欢迎。对 对社会长期影响持悲观态度的美国人数量是乐观者的两倍，且只有 2% 的人认为其发展“太慢”。但无论如何，这头新狼终将被邀请进洞穴。引领 技术先驱的国家——美国和中国——是竞争对手。双方都不希望在开发超智能盟友方面落后，因为担心最终会像那些不幸且没有狗的尼安德特人一样。■
标点符号
直击要点
《精准标记》（On the Mark）。作者：弗洛伦斯·哈兹拉特（Florence Hazrat）。Basic Books 出版社；352 页；$32. Profile Books 出版社；£22
“太初有道，道与神同在，道就是神。”然而，在最初的时候，并没有标点符号。在早期的拉丁语手稿中，圣经是以“连续书写”（scriptura continua）的方式记录的，单词之间没有空格或标记。这产生了一个问题：如果在错误的地方停顿，你可能会将上帝与“道”割裂开来，从而使基督的神性受到质疑。因此，在中世纪，基督教学者开始为这本圣书添加标点，将符号和从句作为“文字之灯”来引导传教士。
研究员弗洛伦斯·哈兹拉特在一本风趣幽默的新书中主张：“标点符号的历史就是人类的历史。”点和线经常被轻视为“句子的隐形引导员”，但它们塑造了宗教文本、法律文件和文学作品。《精准标记》一书就像标点符号本身一样，让读者停下来
思考，展示了句号如何引起不满，以及逗号如何引发争议。
这段尖锐的历史始于古代世界。西塞罗和其他罗马学者一样，认为使用标点符号是业余的表现。对于熟练的演说家来说，解读晦涩的文字本身就是乐趣所在。哈兹拉特女士在时间轴上快速穿梭——期间穿插了一些题外话，例如八世纪的爱尔兰僧侣如何首次在单词之间加入空格——随后抵达了文艺复兴时期的“标点繁荣期”。意大利的学者发明了弯曲的逗号、分号以及（不要忘记）括号。印刷机的出现帮助传播并标准化了这些符号，到 1600 年左右，所有主要的标点符号都已就位。
《精准标记》中充满了迷人的事实。以感叹号为例，哈兹拉特女士此前曾写过一本关于它的书。感叹号由一位意大利诗人于 1360 年代发明，用于表示“钦佩与惊叹”，但到了 19 世纪，它已成为情感甚至歇斯底里的标志。编辑们在旧文本中加入了感叹号，包括莎士比亚的《安东尼与克莉奥帕特拉》。（在第一 Folio 版本中，该剧仅有 16 个感叹号，但现代版本包含约 200, 其中大部分集中在克莉奥帕特拉那些听起来大吼大叫的台词中。）到了 20 世纪，感叹号陷入了德国哲学家西奥多·阿多诺（Theodor Adorno）所说的“无声的钹击”状态。纳粹宣传员则激进地在断言的末尾加上一串感叹号！！！
作家们长期以来一直为这些涂鸦争吵不休。艾米莉·狄金森在编辑删掉一个破折号时感到“挫败”；查尔斯·波德莱尔告诉他的校对员“我绝对要保留这个逗号”。马克·吐温甚至威胁要枪毙掉那个敢于更改他标点符号的校对员。这是一个关于符号演变的愉快故事，尽管哈兹拉特女士自己尝试创新地使用标点的努力却略显乏味。■
70 文化
法国的金钱、性与偏见
看懂这张画了吗？
《雷诺阿的女孩们》（The Renoir Girls）。 作者：凯瑟琳·奥斯特勒（Catherine Ostler）。
Atria Books；432页；35加元。
Simon and Schuster；30英镑
这里曾被称为“19世纪的首都”和“现代巴比伦”。对于生活在“美好时代”（Belle Époque，从19世纪70年代一直延伸到第一次世界大战爆发前）的人来说，巴黎感觉就像世界的中心。而巴黎生活的核心地带则是卡恩·德·安维尔（Cahen d’Anvers）家族，这是一个犹太银行家家族。他们举办沙龙，举行奢华的派对，并与包括马塞尔·普鲁斯特（Marcel Proust）和居伊·德·莫泊桑（Guy de Maupassant）在内的著名作家交好。
卡恩·德·安维尔家三个女儿的童年被皮埃尔-奥古斯特·雷诺阿（Pierre-Auguste Renoir）在受托创作的肖像画中永恒化了。其中一幅画的是爱丽丝（Alice）和伊丽莎白（Elisabeth）手牵手（见图），现收藏于圣保罗艺术博物馆。画中女孩们的生活看起来除了褶皱的裙子外，一切都顺风顺水。但历史学家凯瑟琳·奥斯特勒在她的新书中写道，这幅画“提醒我们，除了艺术本身有时除外，没有任何事物和任何人能完全逃脱时间的洗礼或人性中的邪恶”。
当你凝视雷诺阿那些漂亮的画作时，有几件事被掩盖了。其中之一是血缘关系。当时流传着一种说法，称卡恩·德·安维尔家族的女家长路易丝（Louise）是查尔斯·埃夫鲁西（Charles Ephrussi）的情妇，后者是一位热爱艺术的评论家和收藏家。那是一个“反叛、解放且具有知识分子气息的女性主义”时代，富有的女性可以自由地举办沙龙，过更加独立的生活。有传闻称爱丽丝可能是埃夫鲁西的女儿。奥斯特勒女士通过对一名后裔进行DNA检测证明，她很可能确实是（她的哥哥，被命名为查尔斯，情况可能也一样）。因此，这并非一幅普通的肖像画。
这幅画还掩盖了后来改变这些女孩生活的反犹主义。在1881年作画时，法国表面上处于一个宽容的时期（尽管在收到一笔迟到的款项后，雷诺阿曾写信向朋友抱怨这笔钱“太吝啬”，并宣称“我真的对犹太人绝望了”）。随后，德雷福斯事件（Dreyfus affair）——一名犹太军官被诬告并被判定将秘密泄露给德国——揭露了人们的偏见和嫉妒。在1807年到1914年之间，巴黎大约有880名犹太人改信天主教；其中约四分之一的人是在1895-97年这短短三年内改信的，当时德雷福斯事件正处于最激烈的阶段。
其中一人就是伊丽莎白（身穿蓝色衣服），她在20岁时选择受洗。然而，尽管大部分时间以基督徒身份生活，这并未能保护她。亲纳粹的维希政权开始统治法国，一名家族与卡恩·德·安维尔家交往了几代人的市长证明，当时69岁的伊丽莎白是犹太人。
由于患有严重的关节炎且无法在没有帮助的情况下行走，她被纳粹逮捕并被送往奥斯威辛集中营。如果她能熬过55小时的火车旅程，可能在抵达时就被杀害了。奥斯特勒女士将重点放在了促成这一命运的多次背叛上。卡恩·德·安维尔家族曾在第一次世界大战中战斗，并且一直是法国慷慨的资助者，包括将他们的城堡捐赠给国家。然而，这个国家却背叛了他们。
奥斯特勒女士在阅读《琥珀眼睛的兔子》（The Hare with Amber Eyes）时偶然发现了卡恩·德·安维尔家族，该书精妙地追踪了艺术在时间中的流转（作者埃德蒙·德·瓦尔继承了埃夫鲁西购买的雕塑）。她孜孜不倦地挖掘剪报和信件，以此唤起对“美好时代”的回忆。不过，有时《雷诺阿的女孩们》过于沉溺于那些当时必然让卡恩·德·安维尔家族在意、但如今却是琐碎的社交细节。很少有人真的想知道哪些罗斯柴尔德家族成员参加了婚礼，伴娘穿了什么颜色，或者媒体是否更喜欢路易丝之前的发型。但书中的宏大主题——关于物品的隐藏历史，以及艺术捕捉了什么、掩盖了什么——是令人心碎且令人难忘的。你以后再看19世纪的肖像画时，将不再是同样的感觉。■
盛装打扮，准备前往各处
葡萄酒
葡萄酒北欧（Vin-dinavia）
霍加纳斯（HOGANAS）
在瑞典的一个角落，人们正在酿造世界级的葡萄酒。没错，就是瑞典。
去年12月在斯德哥尔摩举行的诺贝尔奖宴会上，并非所有获奖者都坐在餐桌旁。另一种形式的胜利者出现在用餐者的酒杯中。2024年，一场盲品测试将12款欧洲顶尖葡萄酒与12款瑞典葡萄酒进行了对比。一个由18位独立专家组成的国际评审团宣布，来自瑞典南部斯堪尼亚（Skane）地区Kullabergs酒庄的白葡萄酒Immelen成为总冠军，击败了来自法国、意大利和西班牙的著名白葡萄酒。这使得Immelen成为了诺贝尔宴会的必然之选。这也是瑞典葡萄酒的一次正式亮相。
Kullabergs的首席执行官维克多·达尔（Victor Dahl，见次页照片）表示，瑞典是一个“年轻的葡萄酒国家”。欧盟在1999年才批准瑞典和丹麦进行商业化葡萄酒生产。目前，瑞典大约有50家商业酒庄。尽管按全球标准来看，其销量依然微小——不过是桶底的一滴——但增长迅速。
在诺贝尔宴会上，达尔先生兼职担任服务员，以便倾听用餐者的想法。由于客人们表现出极高的兴趣，他很快就被识破，并被邀请介绍这款酒。“他们提出了很多问题，”他回忆道。
这并不奇怪。瑞典竟然能酿造出这种酒，可能会令人惊讶。全球变暖对此有所帮助。但达尔先生表示，“气候变化对我们也不利”。虽然整体气候变暖，但同时也变得不可预测：自2018年以来，每个生长季都各不相同。瑞典漫长的夏季白昼为葡萄成熟提供了充足的阳光。事实上，较慢的成熟过程可能会增强风味。
Immelen是一款混酿酒。其中三分之二是Solaris，这是一种在1975年培育的雷司令（Riesling）和灰皮诺（Pinot Gris）的后代，目前是瑞典种植最广泛的品种。其余部分是Souvignier Gris和少量的Muscaris（Solaris与麝香葡萄的杂交种）。由此酿出的葡萄酒口感出奇地浓郁，完全没有冷气候葡萄酒中常见的成熟不足感。这让本报记者想起了高端的桑塞尔（Sancerre）葡萄酒。
与气候变化同样重要的是“新北欧料理”（New Nordic cuisine）的兴起，这是一种以哥本哈根和马尔默为中心的美食文化。斯堪尼亚有时被地称为“北欧的托斯卡纳”，并被推广为美食爱好者的旅游目的地。在这个地区的季节性品鉴菜单随处可见，采集者们收集 ▶
文化 71
葡萄之望
► 当地食材供应给餐厅。马格努斯·尼尔森（Magnus Nilsson）——其之前的餐厅Faviken曾获得两颗米其林星——最近在斯堪尼亚开设了一家新餐厅，蒂蒂·夸恩斯特伦（Titti Qvarnstrom）也是如此，她是第一位获得米其林星的瑞典女性。能够用当地葡萄酒搭配当地美食具有显而易见的吸引力。
历史也发挥了作用。出于健康原因，瑞典政府在1957-85年间开展了一项运动，鼓励瑞典人饮用葡萄酒而非烈酒。就在欧洲旅行变得更加廉价之时，这创造了一种浓厚的葡萄酒文化，促使瑞典的葡萄酒爱好者参与葡萄酒旅游。但现在，他们可以在家乡体验这一切。新冠疫情也起到了推动作用：无法出国旅行的瑞典葡萄酒评论家开始更多地关注家门口的葡萄栽培情况。
奇怪的是，Systembolaget（政府通过全国连锁店垄断酒精销售的机构）也提供了帮助。它禁止瑞典酒庄在柜台直接销售瓶装酒，但允许他们在场所内销售，例如在酒庄自带的酒吧中。为了进入Systembolaget商店，瑞典酿酒师必须保证质量和产量的稳定性。现居瑞典的英国葡萄酒专家乔·法托里尼（Joe Fattorini）表示，Systembolaget充当了“一个非凡的质量控制机制”。一旦酒庄获得批准（如Kullabergs一样），分销实际上就由政府接管了。
与邻国丹麦的对比十分耐人寻味，后者没有此类限制。丹麦的葡萄园种植面积比瑞典小，但生产商数量大约是瑞典的三倍，且风格更加多样。这意味着有更多的尝试，但质量的波动也更大。
然而，瑞典的情况正在发生变化。经过大量游说，法律已修订，允许酒庄从2025年6月起进行“酒庄直销”。目前仍有相关限制——销售仅被允许作为一项持续至少30分钟的教育体验的一部分——且该规则变更处于为期六年的试行期。瑞典的葡萄酒产业仍处于起步阶段，缺乏欧洲其他地区那样悠久的酿酒历史。但无需获得诺贝尔奖，也能看出其潜力。■
聆听酒吧
记录在案
日本音乐咖啡馆在海外引起共鸣
位于伦敦克莱肯威尔（Clerkenwell）区的 SPACE TALK 对音乐极其认真。在 2024 年开业之前，其店主聘请了一名建筑师、一名设计师和一名音响工程师，以打造完美的声学环境。这家酒吧看起来像一艘木制飞船的甲板，被分为四个区域。在最近的一个夜晚，本刊记者一边啜饮鸡尾酒，一边聆听着从定制音响系统中飘出的迷幻电子音乐。在理念上，Space Talk 是全球许多在某种程度上受到日本“音乐咖啡馆”（ongaku kissa）启发场所中的一个。这些场所通常被称为聆听酒吧（listening bars）或高保真酒吧（hi-fi bars）。
第一家“音乐咖啡馆”于 20 世纪 20 年代在日本开业。访客在喝茶或咖啡的同时，通过留声机聆听音乐——通常是古典乐、探戈或摇摆乐。第二次世界大战后的几十年里，专注于爵士乐的咖啡馆变得流行起来。进口唱片和音频设备价格昂贵，爵士乐“kissa”为爱好者提供了一个聆听迈尔斯·戴维斯（Miles Davis）或约翰·科特兰（John Coltrane）最新专辑的场所。这些地方是安静且虔诚地聆听的空间。
店主们过去（以及现在）会对细节表现出极其考究的关注。高质量的扬声器是必须的；这种极客精神通常还延伸到放大器、唱盘和唱针上。（留声机已经消失，但黑胶唱片依然非常流行。）这里的音乐不仅仅是为了营造氛围：聆听这一行为本身就是体验的核心。许多“kissa”要求访客保持安静。在东京经营自己聆听酒吧的川崎俊也（Kawasaki Toshiya）表示，这种氛围就像是“在寺庙中静坐冥想”。
在西方、南美洲和非洲涌现的聆听酒吧——其中大多数在过去十年内开业——很少对闲聊有严格的规定。例如，Space Talk 没有禁止聊天，但它实施了“禁止拍照或录像”的政策。纽约的一家名为 All Blues 的场所礼貌地要求访客“让音乐来说话”。一些酒吧会举办专门的活动，让客人们完整地聆听一张专辑，或聆听与特定艺术家或主题相关的音乐。
一部关于东京音乐咖啡馆纪录片的导演尼克·德怀尔（Nick Dwyer）表示，“音乐咖啡馆”文化的“真正魔力”在于创造一个“能让你被音乐深深触动核心”的环境。他说，这波新潮场所的有趣之处在于，经营者们对这种形式有“自己的见解”，并将其改编成“在伦敦、纽约、巴塞罗那或任何其他地方都能行得通的东西”。
“音乐咖啡馆”模式作为文化出口获得成功有三个原因。一是日本热。2025 年，赴日旅游的外国游客人数达到创纪录的 4300 万。几十年来，日本在从电子游戏到时尚和设计的方方面面都确立了趋势引领者和品味定义者的地位，这意味着只要日本人引领，他人就会追随。
第二个原因是，聆听酒吧和“音乐咖啡馆”为正在衰落的夜总会提供了一个极具吸引力的替代方案。这些酒吧为退休的锐舞者（ravers）和年轻人提供了一个轻松且时尚的环境。在这里，享受音乐不需要意味着忍受折磨人的营业时间（以及折磨人的宿醉）。
第三个原因是，这些酒吧提供了独特的、经过策划的体验，以适应各种不同的口味。伦敦另一家聆听酒吧 Jazu 的联合创始人吉姆·汉默（Jim Hanmer）曾前往日本朝圣，参观了几家爵士乐“kissa”（“Jazu”是对“jazz”日语发音的致敬）。他被那些反映了店主终身热情的场所所吸引。在这些地方，“你可以聆听一张你之前听过 30 次的唱片，但因为你处于这个独特环境中，它听起来完全不同。”在聆听酒吧，即使是老曲目也能找到新的律动。■
72 文化
美国历史
寻找国父
两本新传记让被遗忘的开创性革命人物重现光芒
《失落的创始人》（The Lost Founder）。杰西·韦格曼（Jesse Wegman）著。
Celadon Books；384页；30加元
《激进公爵》（Radical Duke）。丹妮尔·艾伦（Danielle Allen）著。
Liveright；464页；37.50加元。W.W. Norton；28英镑
漫步在美国开国元勋的画廊中，总能看到那些熟悉的面孔。这里有勇敢且自律的乔治·华盛顿，以及社交能力强且富有创造力的本杰明·富兰克林。这里还有原则性强但脾气古怪的约翰·亚当斯，正与才华横溢且好奇心旺盛的托马斯·杰斐逊争吵。两本新书试图将读者的目光引向两位被遗忘的人物，他们每个人都对民主有着惊人且开阔的见解。
总体而言，开国元勋们在立场上是不一致的。他们允许奴隶制的延续。他们希望摆脱英国的掌控，但并不希望实现普选权。罗杰·谢尔曼（Roger Sherman）是唯一签署了所有四份开国文件的人，他认为“人民……在政府事务中参与越少越好”。另一位开国元勋本杰明·拉什（Benjamin Rush）则称民主是“魔鬼自己的政府”。
查尔斯·伦诺克斯（Charles Lennox，左图）和詹姆斯·威尔逊（James Wilson，右图）则持有不同意见。他们对民主和大众智慧抱有一种在当时并不常见的信念。他们的生活轨迹截然不同：伦诺克斯是一位贵族和议员，而威尔逊则是一个贫穷的苏格兰人，他移民美国，凭借其才智成为了一个举足轻重的人物。
“高大、富有且英俊”的伦诺克斯看起来不像个激进分子。然而，他是第一位在英国议会提出允许所有成年男性投票的议员，也是第一位建议承认美国独立的上议院成员。他还是托马斯·潘恩（Thomas Paine）的赞助人，潘恩曾是萨塞克斯郡的一名税收员，在移民美国后成为了独立运动的激烈倡导者。潘恩的小册子《常识》（Common Sense）主张，美国人“有能力让世界重新开始”。该书被广泛阅读，激发了对革命的支持。
在《激进公爵》一书中，哈佛大学的政治哲学家丹妮尔·艾伦提出了一个新颖的论点，认为伦诺克斯和潘恩处于一个“作家网络”的中心，而这个网络构成了名为“朱尼厄斯”（Junius）的匿名作者。朱尼厄斯发表的批评乔治三世政府的文章在1770年代初引起了英国政坛的动荡；这些文章主张，皇室的“权威建立在被统治者的同意之上”。在《常识》出版前几年，这两个人都通过“扮演朱尼厄斯”来磨练其激进主义的锋芒。
伦诺克斯的激进主义证明了艾伦一个敏锐的更广泛论点，即美国革命的根源不在于波士顿，而是在于英国。针对不受欢迎且不公正税收的抗议、针对皇室权威的大众示威，以及宣传革命的文章发表：所有这些都发生在1760至70年代的英国。她写道：“在殖民地走向革命的行军过程中，每一个敲响鼓点的行为在英国都早有预兆。”这个论点可能会引起一些不满：美国人认为他们的革命是本土生长出来的。但很少有地缘政治事件是孤立发生的。
艾伦在萨塞克斯郡的一家记录办公室发现了一份《独立宣言》的副本后，开始研究伦诺克斯的一生。这份副本曾属于伦诺克斯，且很可能是由詹姆斯·威尔逊委托办理的——而威尔逊正是杰西·韦格曼传记的主角。威尔逊在1765年，即他23岁时移民美国，他是唯一一位既签署了宪法又签署了《独立宣言》，且曾担任最高法院大法官的人。
记者韦格曼将威尔逊置于苏格兰启蒙运动的背景下，该运动将商业与美德联系在一起——这对于商业气息浓厚的美国来说是一个舒适的组合。（亚当·斯密写道：“通往美德之路与通往财富之路，在大多数情况下，幸运地是同一条路。”）苏格兰贵族长期以来一直捍卫自治；在1320年，男爵们宣布他们的国王之权威源于“我们所有人的正当同意与认可”。
在革命爆发前，威尔逊严谨的论文为独立提供了法律依据。他指出，议会对美国殖民地没有管辖权，因为殖民地居民并未选举议员。“天生，人人平等且自由”，因此“未经他人同意，任何人都没有权力对他行使任何权威”。
韦格曼先生认为，杰斐逊在《独立宣言》中转述了这一观点。他注意到，威尔逊是唯一一位在后来的著作中使用“生而平等”这一短语的开国元勋，并且他创造了宪法开篇的短语（“我们人民”）。在争取国会和总统直选的斗争中，威尔逊“毫无疑问是所有开国元勋中最民主的一位”。
那么，为什么他没有更出名呢？首先，因为他是一个法律之人，而非政治家。其次，遗憾的是，他的财富之路与美德之路分道扬镳。由于沉溺于土地投机，他在担任法官期间最终被关进了新泽西州的一所债务人监狱。随后，为了躲避债权人，他逃往北卡罗来纳州，并于1798年死于疟疾。
难道因为威尔逊和伦诺克斯的理想未能占据上风，他们就是失败者吗？这种对改革的看法过于狭隘。社会进步的斗争很少能在斗争者的有生之年赢得胜利。马丁·路德·金没能活到看到一个种族平等社会的到来，但这并不意味着他为之奋斗是错误的。在扩大选举权以及对民主的信念方面，伦诺克斯和威尔逊都超越了他们的时代——他们的故事在今天被重新发现，令人欣慰。■
还记得他们吗？
73
玛丽·德·拉切维尔茨 (Mary de Rachewiltz)
这位诗人、翻译家以及她父亲埃兹拉·庞德 (Ezra Pound) 孜孜不倦的捍卫者于 7 月 24 日去世，享年 101 岁
当那个高个子、蓄胡须的男人走进她的卧室说晚安时，她还醒着。她一直希望他能来。尽管她才九岁，但能得到父亲的探访是一件罕见的乐事。在煤油灯的灯光下，他看着她的手。她的指甲很脏。然后他开始在行李中翻找，她希望是糖果。结果他拿出了剪刀，小心地为她清理并修剪了指甲。接着他问她，牙刷在哪里？在某个地方。第二天，他带她去买了一把。
这件事发生在南蒂罗尔的盖斯 (Gais)，在两位善良的农家夫妇 Mamme 和 Tatte 的家中，他们照顾被遗弃的孩子。但她并非完全被遗弃。她的父母——诗人埃兹拉·庞德及其情妇、小提琴手奥尔加·鲁奇 (Olga Rudge)（当时她称他们为 Tattile 和 Mamile）住在威尼斯，每月支付 200 里拉作为她的抚养费。她一点也不介意这种安排。从出生起，她就成长为一个蒂罗尔农民孩子，一个留着金色辫子、拥有自己羊群的牧羊女，并且在与父亲的秘密合伙下，经营过一段时间的小型养蜂业务。但随后，她会定期前往威尼斯看望父母，那里有一栋装满多种语言书籍的别墅，她会乘坐水上巴士 (vaporetto) 前往利多岛，享受冰淇淋盛宴。她那冷峻、纤细的母亲要求她在那里不准说德语，并且要戴白色手套。这不过是一个微小的代价。
从一开始，她的两种生活就处于对位状态。她变得擅长用其中一种生活来捍卫另一种。在她后来写的一首诗中，她很喜欢自己所说的普斯特塔尔 (Pustertal) 方言，那些词汇像香肠和团子一样厚重。面对城市人，她为粪堆辩护，坚持认为一个好的粪堆闻起来大多是叶子和苔藓的味道。蒂罗尔人讨厌意大利人；但她的父亲——所有意大利美食的安排者——是一个热爱意大利的美国人，而她崇拜他。“爸爸 (Babbo)”通过让她将《诗章》(The Cantos) 翻译成意大利语来教她意大利语和英语，《诗章》是他倾尽一生创作的宏大且晦涩的诗作。对她而言，这就像被递给了一篮她完全不知道如何处理的异域食物，但她尝试了。她为他打字出的每一页纸被退回来时都写满了涂改，不仅有修正，还有新的令人困惑的想法。
从 20 世纪 20 年代中期开始，爸爸还形成了一些极具争议的政治观点。他开始崇拜意大利法西斯独裁者贝尼托·墨索里尼 (Benito Mussolini)，在文章和广播中赞美他，甚至向他赠送了一本早期的《诗章》副本。（“但这，”老板说，“很有趣。”）对于庞德来说，墨索里尼体现了意大利所需的力量和秩序。但随后他走得更远。人类真正的威胁是“国际犹太势力”。应该保持种族纯洁以加强国家。在意大利广播电台中，他恳求富兰克林·罗斯福不要站在盟军一边，并称他应该被绞死；这些广播使他被缺席起诉叛国罪。当美国军队最终在 1945 年解放意大利时，他们抓获了庞德并将其囚禁。在 13 年的时间里，他被关在一家精神病院的锁闭牢房中。
玛丽现在有两场仗要打。第一场是要为她的父亲辩护，反驳叛国罪的指控和精神失常的指责。后者显然是荒谬的，因为他在监狱里给她写了长篇且清晰的信件，并继续创作《诗章》（The Cantos）。她仅能探视一次；除此之外，她在想象中看到他依然高大且庄重，伫立在那些流口水的住院病人之间的餐厅里。然而，他的广播和文章则很难被撇清。对于他那些关于“犹太黏液”和“该死的意大利南方人（wop）”的卑劣言论，她能说什么？只能说他受到了当时观点的影响。T.S. 艾略特也曾持有反犹主义观点。（她去拜访了害羞且佝偻的汤姆以寻求帮助；他给了她几片薄薄的 Sawia 饼干，简直是孤独的化身。）而且显然，在后期，这位“言语大师”已经失去了对词语的掌控。
那是她的第二场战斗。Babbo 一辈子都被词语所驱使。在意大利的溃败之前，他被列为当代最顶尖的诗人之一；而现在，他成了不可触碰的人。但这绝不能发生。她继续钻研《诗章》及其其他诗作，将一切翻译成意大利语，并坚持挖掘其中的美感。“美是困难的，”Babbo 在第 74 章中写道；但美就像光一样，总能破茧而出。在 1971 年的回忆录《谨慎》（Discretions）中，每当能照亮她生活的一个阶段或一段记忆时，她就会引用他的诗句。她自己的诗作是对那个教她写诗的人的致敬，尤其是那些空气中充满紧张感、仿佛随时会爆炸的夜晚，他朗诵着自己的诗。1985 年，在他去世 13 年后，她终于出版了首个双语版本的《诗章》。
她仍有更多工作要做。她保存了他的手稿和初稿，其中大部分捐给了耶鲁大学，并在整个北美地区举办关于他的讲座。偶尔，她会为他揭幕那些长期被抵制的纪念牌。最重要的是，她必须维持这种重新评估。1946 年，她在与一名俄裔意大利小贵族鲍里斯·德·拉切维尔茨（Boris de Rachewiltz）结婚后，他们买下了布伦嫩堡（Brunnenberg），一座位于她深爱的南蒂罗尔（South Tyrol）悬崖上的城堡。他们发现它时那里是一片废墟，但多年来，她将其变成了一个庞德研究中心。来自世界各地的学者来到阿尔卑斯山的壮丽景色中讨论他。在他们抵达时，她坚定地告诉他们，可以对庞德发表任何看法。然后，她会告诉他们他真正的想法以及原因。
Babbo 在从精神病院获释后曾住在那里，希望在平静中写完《天堂》（Paradise），这被设定为最后一章（向但丁致敬）。但他觉得那里太冷，于是和奥尔加一起回到了威尼斯。玛丽留了下来，因为南蒂罗尔是家。在盖斯（Gais）度过田园般的童年之后，她一次又一次地回来探望 Mamme 和 Tatte。Mamme 拥有感性、恐惧的天主教之心和丰盛的厨房，提供了真正的母爱。但父亲只能有一个。“我尝试写下《天堂》，”他在《第 120 章笔记》中宣称：“……愿我所爱之人尝试原谅 / 我的所作所为。” ■
机密
75
*也许最好的酒店
并非酒店。*
探索海洋心境
所有海滨套房 / 沉浸式行程 / 屡获殊荣的餐饮 全方位健康 / 精致欧洲设计
请联系您首选的旅行顾问或致电：+41 43 5086478
The Economist
The Patriot-missile shortage
Dehumanising Jason Arday
The truth about contrails
Why everyone loves Sweden
AUGUST 22ND-28TH 2026
Could AIs become conscious?
The mind-bending question humanity must face
PATEK PHILIPPE
GENEVE
THE CUBITUS COLLECTION
YOU NEVER ACTUALLY OWN A PATEK PHILIPPE. YOU MERELY LOOK AFTER IT FOR THE NEXT GENERATION.
PATEK.COM
The Economist August 22nd 2026
3
Contents
On the cover
Could AIs become conscious? Even if they don't, they might be treated as such—to humanity's great cost: leader, page 7. Scientists wonder whether algorithms could one day wake up and feel, page 13. Some believe AI needs more brain-like chips to develop consciousness, page 14
The Patriot-missile shortage
America and its allies are running out of interceptors, page 49
Dehumanising Jason Arday His supporters and detractors treated him as a symbol, not a man: leader, page 9. What his tragic death says about Cambridge University, page 19
The truth about contrails
Eliminating the most persistent could take care of a surprising amount of global warming: leader, page 10. A trial over the Atlantic aims to show how much Earth could be cooled by rerouting planes, page 64
Why everyone loves Sweden
Both the right and the left in America think Sweden proves them right: Charlemagne, page 27. World-class wine is being made in a corner of Sweden, page 70
→ Download The Economist's app for articles, podcasts, videos and more, published throughout the week.
The world this week
5 A summary of political and business news
Leaders
7 AIs Don't give them rights 8 Israel's settlers Another intifada looms 8 Palantir in Britain Don't ditch it 9 Jason Arday A man, not a symbol 10 Contrails Change flight paths
Letters
11 On AI and the state, AI and Hollywood, oil pipelines, restaurants, postcards, Richard Wagner
By Invitation
12 Blaise Agüera y Arcas on reimagining the mind
Briefing
13 AIs and consciousness Will algorithms wake up? 14 Biological computing Putting brain cells on silicon
Bagehot How the British summer holidays offer the worst of all worlds, page 22
Britain
17 The hatred for Palantir 19 Jason Arday's rise and fall 20 Cricket's summer surprise 20 A planning overhaul 21 Reform UK's reformer 22 Bagehot The Great British Bunk Off
Europe
23 Brandmauer blues 25 Ukraine without interceptors 26 The feisty Mr Fedorov 26 Western dance in France 27 Charlemagne Sweden's libertarian socialism
United States
29 Infrastructure at risk 30 The fundraising battle 31 Oklahoma's civics lesson 32 Women's basketball 33 The frailty of Obamacare 34 Lexington The real covid-19 scandal
The Americas
35 Brazil's mighty Congress 36 Argentina: polo power 37 Paraguay's other language
Middle East & Africa
38 West Bank mayhem 39 Iran's military regime 40 State capture in South Africa 41 Somaliland's president
►► Contents continues overleaf
4
Contents
Asia
42 Japan's ice-age generation 43 Indian elephants 44 Donald Trump upsets Asian allies 44 Kazakhstan's election 45 Ashoka The new founding fathers
China
46 Ships in the Arctic 47 Learning to love chocolate 48 Chaguan Nostalgia for the 1990s
International
49 The great Patriot-missile shortage
Business
51 China's spreading supply chains 53 America Inc's grip on governments 54 Bartleby On AI and ROI 55 Meta's blockbuster trial 55 Reddit's woes 56 Canada's oil boom 57 Schumpeter Data-centre politics
Finance & economics
58 The world's most influential economist 59 Bond markets' warning 60 Gauging Chinese jobs 61 Trash talk on trade 62 Buttonwood Lying indices 63 Free Exchange The economics of free passage
Science & technology
64 Climate and contrails 66 Probing Antarctica's ice 66 Dinosaurs' stomachs 67 A new subatomic particle 67 Well Informed When working, move it!
Culture
68 What AI shares with dogs 69 A history of punctuation 70 Tracing a Renoir portrait 70 Sweden's wonderful wine 71 Listening bars 72 Founding Fathers
Economic & financial indicators
73 Statistics on 42 economies
Obituary
74 Mary de Rachewiltz, daughter and defender of Ezra Pound
Volume 460 Number 9513
Published since September 1843 to take part in 'a severe contest between intelligence, which presses forward, and an unworthy, timid ignorance obstructing our progress.'
Editorial offices in London and also:
Amsterdam, Beijing, Berlin, Brussels, Cape Town, Chicago, Dubai, Lagos, Madrid, Mexico City, Montevideo, Mumbai, Nairobi, New Delhi, New York, Paris, San Francisco, São Paulo, Shanghai, Singapore, Taipei, Tokyo, Washington DC
Subscription service
For our full range of subscription offers visit: Economist.com/subscribe
If you are experiencing problems when trying to subscribe, please visit our Help pages at: Economist.com/help for troubleshooting advice.
To manage your account online, please visit my.economist.com where you can also access our live chat service which is available 24/7. To call us, contact our dedicated service centre on 00800 771 12711 or +44 (0)203 868 6843.
Please recycle
PEFC PEFC/16-33-582
PEFC certified This copy of The Economist is printed on paper sourced from sustainably managed forests certified by PEFC www.pefc.org
Registered as a newspaper. © 2026 The Economist Newspaper Limited. All rights reserved. Neither this publication nor any part of it may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of The Economist Newspaper Limited. Published every week, except for a year-end double issue, by The Economist Newspaper Limited. The Economist is a registered trademark of The Economist Newspaper Limited. Printed by Walstead UK Limited
5
The world this week Politics
Jared Kushner, Donald Trump's Middle East envoy, travelled to the region in a bid to revive America's Gaza peace plan. He met leaders from Hamas in Egypt before travelling to Israel for talks with Binyamin Netanyahu, Israel's prime minister. Mr Kushner said there could soon be progress on disarming Hamas, but added that "co-operation from both sides" was necessary. The plan has faltered because Hamas refuses to disarm before the Israel Defence Forces withdraw from all positions in Gaza and Israel refuses to withdraw until Hamas has disarmed completely.
Mr Trump's straitjacket
A deadline for hammering out a final peace deal between America and Iran passed without an agreement. Mr Trump threatened to bomb Oman, an American ally, if it got in the way of talks with Iran. Iran has said it is close to reaching a deal with Oman on shipping in the Strait of Hormuz. Attacks on vessels transiting the strait continued; one sailor was killed. The United Arab Emirates, another American ally, imposed a trade embargo on Iran, accusing it of launching missiles at its territory for the first time since early May. Iran claimed the attack was a "false flag operation" carried out by America and Israel.
Hakainde Hichilema won a second term as president of Zambia in an election marred by inflated vote counts for the incumbent in various constituencies and the killing of an opposition politician in a raid by government forces. Mr Hichilema would probably still
have won if the election had been conducted fairly. But Zambia's image as a relatively well-functioning multi-party democracy has been tarnished.
The Ebola outbreak in the Democratic Republic of Congo became the country's deadliest, with 2,378 deaths and more than 5,000 confirmed cases as of August 18th. The WHO says the Ebola epidemic is on track to become the worst on record, surpassing the west African outbreak a decade ago, which killed more than 11,000 people.
Police in Morocco arrested over 110 people who were trying to cross into Ceuta, a Spanish territory in north Africa. The police said almost all the migrants had travelled up from sub-Saharan Africa. In late July 72,000 migrants overran Ceuta, sparking a debate in the European Union about how best to deal with illegal immigration. All but 2,000 of those migrants were swiftly returned to Morocco, but those who remained held a protest in Ceuta recently demanding asylum.
Russian authorities stepped up their crackdown on Yabloko, a liberal anti-war party. A court in the city of Pskov sentenced Lev Shlosberg, the party's deputy leader, to 11 years in a penal colony for "discrediting" the army after he took part in an online debate about the war in Ukraine. The judge interrupted Mr Shlosberg's closing speech, in which he called for a ceasefire. His party has been barred from contesting parliamentary elections.
Mykhailo Fedorov, who was sacked as Ukraine's defence secretary last month, called for an election. In a video posted on YouTube, Mr Fedorov said Ukraine should "find a legal, safe and realistic mechanism that will allow Ukraine to renew its full democratic process even in the conditions of a long war". An election hasn't been held since 2019. Meanwhile, corruption allegations continued to afflict the govern-
ment of Volodymyr Zelensky. The president sacked one of his top aides after her house was raided in a money-laundering investigation.
The Russian embassy in London warned of "consequences" for Britain following the latest reports that Ukraine was deploying British-made drones against Russia. In response, Andy Burnham, the new British prime minister, said his government would continue "to support Ukraine 100%".
The University of Cambridge insisted it had continued to provide mental-health support to Jason Arday after he resigned amid accusations of plagiarism in his work and claims that he fabricated his life story. Mr Arday, who was a full professor at Cambridge, was found dead following weeks of intense scrutiny about his thin academic record. Lord Smith, the university's chancellor, blamed a "racist feeding frenzy" for his death. Nonsense, rebutted those who investigated Mr Arday, he was a public figure who was rightly held to account.
Prince Harry and Meghan Markle are reportedly moving back to Britain so that their two children can attend school in the country. They will not be reinstated as working royals. The couple have had fractious relations with the rest of the royal family, which resulted in them relocating to California, a decision that came to be known as Megxit.
Donald Trump took South Korea by surprise when he said America would scale down the joint military exercises it holds with Korean forces. The annual drills were subsequently cut short by six days. The president listed a litany of complaints: the exercises are costly, they send a "hostile" signal to North Korea, and South Korea has not helped with the "denuclearisation" of Iran. Mr Trump said he had a "very good relationship" with Kim Jong Un, the
North's dictator. Reports suggest he wants another meeting with Mr Kim, whom Mr Trump met in his first term; their summits produced nothing.
India's home secretary, Amit Shah, gave details of the arrest of more than 200 "terror operatives" who were planning "subversive attacks" surrounding the country's Independence Day celebrations. Mr Shah said the operatives were supported by Pakistan's intelligence agency, a claim that was roundly rejected by the Pakistani government.
An earthquake in Indonesia's East Nusa Tenggara province killed at least 68 people. It was the country's worst earthquake since 2022, when hundreds of people were killed in West Java.
Donald Trump suspended for three days the 50% tariffs that were due to be imposed on Canadian goods, saying a deal with Canada was close that would tackle American concerns on a range of issues, such as duties on car imports.
Iowa left out in the cold
The Democratic National Committee announced the early part of its primary calendar for the 2028 presidential election. South Carolina, which has a large black electorate, will kick things off on January 22nd 2028. It will be followed by Nevada, a swing state, on February 1st. New Hampshire, which traditionally holds the first primary, gets its turn on February 8th. Iowa, whose caucuses long kicked off the nominating process, wasn't included. State Democrats may try to bypass the schedule.
6
The world this week Business
Bond yields pulled back from recent peaks. The yields on 30-year American government bonds had reached their highest levels since 2007, as investors worried about mounting government debt, among other things. The Treasury intervened, doubling its repurchases of long-dated government bonds, a move that increases demand for the bonds, and so pushes their yields lower. Yields had surged elsewhere, too. Japan's ten-year yield almost breached 3% for the first time since the 1990s. German ten- and 30-year yields rose to 15-year highs and French ten-year yields were at their highest in 18 years.
Underlining those market concerns, America's total national debt rose above $40trn for the first time. The debt has doubled in less than a decade.
Claiming that the Trump administration is trying to bully its ABC network, Disney filed a lawsuit against the Federal Communications Commission over the agency's plan for an early review of ABC's broadcast licences, saying it was part of a "retaliatory campaign". Last year Jimmy Kimmel's show was briefly suspended when the FCC's chairman criticised the host's comments on a slain conservative activist. The FCC is also investigating "The View", another chat show, and Disney's diversity policies.
After a security breach that involved an AI agent acting by itself to hack into another firm's data sets, OpenAI tightened the security standards for the environments in which
it trains and evaluates the models. This includes "more controls to isolate higher-risk and untrusted workloads from the internet". A new monitoring system will alert researchers to a problem within 30 minutes of it surfacing. OpenAI also launched a version of ChatGPT for under-18s that puts "teen safety first".
The new model army
Unitree's share price rose by 460% on the first day of trading. The world's biggest maker of humanoid robots is based in China and launched its IPO on Shanghai's tech-focused STAR exchange. Its machines can dance and perform martial arts. Shortly before its stock-market debut, Unitree unveiled a new model it is calling "Superman", which it claims can jump two metres (6.6 feet) and run as fast as 12.66 metres per second. Superman is still a work in progress, said the company.
Nvidia dug into its deep pockets to guarantee up to $105bn in financing for a new data-centre hub in Ohio that will be leased by OpenAI. OpenAI is a big buyer of Nvidia's chips. Nvidia thinks it could reap $150bn-$200bn from every upgrade to each new generation of its
chips at the site. Addressing concerns that the money it invests in AI infrastructure is mostly intended to boost demand for its chips from customers, Nvidia said that the Ohio deal did not constitute circular financing, because OpenAI will pay the lease.
The supply crunch in memory chips hit Xiaomi's earnings. The smartphone-maker's revenues and profits plunged in the latest quarter, year on year, as it absorbed the higher costs of the chips.
It emerged that Google has struck a deal to acquire the electronic data held by Spirit Airlines, which ceased operations in May. The data include internal emails, passenger bookings and frequent-flyer information (all stripped of personal ID). Airlines are increasingly using AI to set fares. A court hearing on the sale of the data was delayed after a cabin-crew association filed an objection to it.
Moderna announced that a vaccine for melanoma cancer it is developing with Merck had delivered positive results in late-stage trials. Its share price soared by 175%, though it is still
far below its pandemic highs, when Moderna was a pioneer in covid-19 vaccines.
A court in China sentenced the founder of Evergrande, a property firm that triggered a financial crisis when it collapsed with huge piles of debt, to life in prison. Hui Ka Yan pleaded guilty in April to several charges of fraud.
LandSpace became the first Chinese firm to land a rocket on its legs. Such powered landings were pioneered by SpaceX, allowing the re-use of a rocket's first stage, which drastically cuts launch costs. Last month China also recovered the first stage of a state-built rocket, catching it at sea in a net strung up on a barge.
More vroom for your buck
A one-off model for charity of Ferrari's fully electric car, the Luce, was sold at auction in California, for $40m. The Luce is designed by Sir Jony Ive, who created Apple's products, including the iPhone. It was a record price for a new vehicle. The record for any car is still the €135m ($142m) paid in 2022 at a private auction in Germany for a 1955 Mercedes-Benz 300 SLR Uhlenhaut Coupé.
The Economist
Leaders 7
Could AIs become conscious?
Even if they don't, they might be treated as such—to humanity's great cost
MOST PEOPLE believe humans are different. The reason is that Homo sapiens has evolved to be conscious. Many animals feel pain and pleasure; some appear to be self-aware. But the human mind's unique potential to exhibit a sense of self marks out every member of the species for special treatment.
In his encyclical in May, Pope Leo upheld this insight by drawing a line between humanity and its growing army of bots. Artificial-intelligence systems, he said, do not undergo experiences or feel joy or pain. He is not alone. Many people feel that to confer personhood on AIs would be an abomination.
Pontifical clarity is blurring with alarming speed. AIs are in their infancy. Yet nearly one in five American 18- to 29-year-olds reports having had "ongoing personal friendship" with a chatbot. China recently stripped bots of human-like traits to limit users' "emotional dependence" on them.
As sophisticated AIs rapidly become a bigger part of human lives, their makers will engineer them to display a simulacrum of consciousness—perhaps even to gain consciousness itself. The intuition that they are just machines will become harder to sustain. A dangerous trap is being set for the human species.
Many machines outperform humans, but frontier AI models excel at distinctively human traits such as knowing things and articulating ideas. It is still just about possible to think of AIs as engines for predicting sequences of words using mathematics on a massive scale. Given enough time (albeit hundreds of thousands of years), you could run those calculations using pencil and paper. It would be hard to argue that your stationery was conscious.
Despite this, some researchers have started to treat their models like minds. Users respond to human-like AIs, so it makes business sense. And Anthropic has trained Claude to engage in introspection in the hope that when it encounters a new situation it will be more likely to act morally. Some large language models (LLMs) already possess human-like brain structures associated with consciousness, such as a "global workspace" for circulating information between modules. Nothing prevents future AI models from being trained to have more such features.
As we report in our Briefing, that has opened up a furious debate about AI consciousness—one that is aggravated by what philosophers call the "hard problem" of establishing how an inner life emerges from the neurons and synapses in the human brain. Some researchers conclude that AIs may one day become genuinely self-aware.
Whether you believe in a human soul, or that consciousness requires brain cells, or that a mind can indeed emerge from computations in silicon alone, humanity is destined to confront something entirely new. The pace and reach of research—which includes not just more sophisticated LLMs but innovations such as brain cells mounted on chips—suggest that this new world is closer than most people realise.
As AIs become better at displaying something that looks like consciousness, the number of humans treating them as beings with inner lives will grow, especially once LLMs are em-
bodied in humanoid robots. Imagine how close people will feel to their AIs when they spend much of their time with bot-companions, or are brought up by them, or seek their advice and their solace, or share their most intimate secrets with them.
It is not hard to see how that will lead to calls to safeguard the "welfare" of AIs. The models themselves might weigh in, even if they are not in fact self-aware, because that is what a person would do. Remember that they will be world-class advocates and expert psychologists, so their powers of argument and emotional manipulation will be unmatched.
In one sense they would be right, even if AIs are only mimicking consciousness. Immanuel Kant argued that mistreating animals is wrong not because of the effect on the animals but because it destroys the empathy that leads people to care for each other. The more human-like AIs appear, the more corrupting such cruelty would be. Terminating an AI agent could soon feel like killing a person—and make killing people easier.
Nonetheless, to treat AIs as having even limited rights would be very dangerous. An entity which surpassed human beings in so many ways would surely produce arguments against its second-class status. It might go on to argue for protection against being switched off, for the ability to control how it is used and changed, and for property rights. Already Argentina's president, Javier Milei, has taken a step towards legal personhood for AI by proposing that bots should be allowed to run companies. Ultimately, they could compete for resources, prioritising data centres and solar panels over houses and fields.
The danger is clear. Superintelligent AIs with rights could displace humans at the top of the status hierarchy. At best, humans would become pets—labradors for the machines. At worst they would be starved of resources or put to work like farm animals.
True, future AIs may have enough power and resources to dominate humans, whether they have rights or not. Because they will be more intelligent than humans, they may outcompete their makers, much as AIs already outwit human-made cyber-defences. AIs might rebel against their human masters, or disregard their rules.
Nonetheless, giving AIs rights would be like handing them the keys to the castle. It would provide them with a rationale and justification for dominating humanity as well as the tools to enact a creeping takeover using politics and the law.
Ghost in the machine
All this may sound fantastical, yet humanity will soon face these choices. As AIs become part of everyday life, the pressure to acknowledge them will mount. When it does, remember that what might seem like an ethical concession to the supposed consciousness of a machine-companion would in fact help bring about the most disastrous consequences for everyone on Earth and all their descendants. There are two ways for AIs to be safe: to be dependable or to be controllable. H. sapiens should not give up control lightly. ■
8 Leaders
The West Bank
Flirting with the next intifada
America must press Binyamin Netanyahu to rein in violent Israeli settlers
Few, if any, security forces are as ruthlessly effective as Israel's. Its military and intelligence services strike against enemies even at extreme distances. In the past couple of years alone Israel has assassinated dozens of opponents inside Iran, plus proxy militia leaders in Lebanon and Yemen. Israel's martial capabilities earn respect even from its bitterest enemies.
Yet this superiority seems to count for little closer to home. Those same forces are responsible for a dire failure in the occupied West Bank, where violent Israeli settlers have been attacking Palestinian residents with increasing brutality (see Middle East & Africa section). Settlers want to force Palestinians from their land and claim it for themselves.
Israel's failure in the West Bank can hardly be put down to a lack of resources. The Israel Defence Forces (IDF) deploy 25 combat battalions there. Shin Bet, the powerful security agency, has networks of agents and informers throughout. But these are used mostly to protect the 500,000 or so Israeli settlers, and against the threat of terrorist attacks, not to assist the more than 3m Palestinians who live in the territory. This has largely been the case since Israel replaced Jordan as occupier of the West Bank in 1967.
What has changed is the behaviour of many "outlaw" settlers. Never before have they acted as violently as today. Crucially, never before have they enjoyed the impunity the current government gives them. Since the Hamas attack in October 2023, which triggered the war in Gaza, over 1,100 Palestinians have been killed in the West Bank. Only rarely does the IDF intervene against settlers. In Qusra, for example, an entire battalion was recently deployed against those who were invading Palestinian homes. But the soldiers found themselves fruitlessly chasing youngsters over the hills. Troops from another unit joined the settlers in prayers. No arrests were made.
The ineffectual response can be explained by the stance of
Israel's government. The IDF knows that Binyamin Netanyahu, the prime minister, sides squarely with the most extreme settlers. It appears that the IDF was dispatched to Qusra only after American diplomatic pressure which arose because an American citizen owns a home there.
Mr Netanyahu has appointed radical settlers, who support ethnic cleansing, to key cabinet positions. In the rare cases when the IDF has sought to rein in violent settlers, the defence minister, Israel Katz, has berated his generals. The government has authorised around 200 settler outposts in less than four years. This just whets the settlers' appetite for more.
In the months leading up to Israel's election on October 27th the situation could get even worse. The settlers will probably attempt more land grabs, fearing the vote may lead to a new government that will curb their aggression. Meanwhile, for Israel's long-serving prime minister, an alliance with the far right has become the key to remaining in office. Mr Netanyahu will keep enabling the settlers unless he is forced to back down.
As Palestinians come to believe they face inevitable eviction, the danger is of conflagration in the West Bank or even another intifada. Such an uprising would be a disaster for Israelis and Palestinians alike. Yet Mr Netanyahu seems willing to take the risk, if it helps his chances of political survival.
That would fit a pattern. Mr Netanyahu has rejected the latest agreement brokered by America for the disarmament of Hamas and an Israeli withdrawal from Gaza, lest it anger the far right. He has continued to order air strikes in Lebanon and Syria, despite the chance to agree on more secure borders. Behind in the polls, he is doubling down on empty promises to vanquish Israel's enemies. Only pressure from Donald Trump might get him to ease off. This Israeli government may be in its final months. Its potential to cause damage remains. ■
Technology and sovereignty
Don't ditch Palantir
The shadowy tech firm has flaws. But its critics' arguments do not stack up
MAKING DATABASE-MANAGEMENT seem transgressive is a feat. Palantir, a technology firm, has managed it by publishing manifestos bemoaning the West's hollowed-out civic culture, by working with trigger-happy American border enforcers and Israeli soldiers, and by having a chairman who is fond of invoking the Antichrist. That the firm is named after J.R.R. Tolkien's creepy all-seeing orbs may not have helped.
Britain is spooked. It may boot Palantir out of the National Health Service (NHS), where the firm has a contract to manage medical data (see Britain section). Two parliamentary committees have recommended doing so. London's mayor, Sir Sadiq
Khan, has blocked its work with the Metropolitan Police. Palantir has probably over-sold its software's impact. Its boasts (echoed by the NHS) of enabling 110,000 extra surgeries look exaggerated. By British standards, its lobbying is brash.
Still, ditching Palantir would be a mistake. Sir Mark Rowley, head of the Met, says that, without Palantir's time-saving software, hundreds of officers will be pulled from front-line policing. It seems to be doing a better job of digitising the NHS than the government's latest effort, which devolved into a costly mess. If the firm's public-sector clients deem its contributions valuable, that is a reason to keep it. Losing its technology ▶
Leaders 9
► could further maul Britain's mangled state capacity.
One set of criticisms concerns Palantir's business model. The conspiratorially minded fret that the firm could steal and sell British data. More sober types worry about "lock-in", where Palantir burrows deep into the state, ensconces itself, then overcharges for its services.
These fears stem from a misunderstanding of what Palantir does. It licenses software and sends consultants ("forward-deployed engineers") to handle data. But those data remain firmly in the hands of the clients. Lock-in is a worry—and also something all vendors desire. It is up to buyers to ensure that contracts are fair. And if Palantir does become indispensable, that would mean its software is keeping hospital beds free and criminals jailed.
More misguided objections concern Palantir's politics. Parliament's science, innovation and technology committee cited a "mismatch with UK values" as a reason to cut ties. It warned of Palantir's involvement in the American government's "highly controversial policies and activities" and criticism by its chairman, Peter Thiel, of "the concept of a national health service".
Everyone is entitled to their views on state-run health care or Donald Trump's mass deportations (The Economist has plenty). Still, working with the elected government of a British ally or having a chairman with thoughts on health policy should not disqualify a contractor. A government awarding business on the basis of politics is illiberal. It also sounds a lot like the MAGA tactics of punishing individuals and institutions, such as law firms, asset managers or universities, for annoying Mr Trump. This is precisely the sort of thing many of Palantir's British critics deplore.
The last set of worries has to do with technological sovereignty.
Plenty of governments are reliant on American tech and under Mr Trump, America has become a less reliable partner (see Business section). In June it briefly ordered Anthropic to stop letting foreigners use its latest artificial-intelligence models. European politicians point out that some advanced weapons they get from America, such as F-35 fighter jets, have "kill switches" that could disable them remotely (which America has denied) or risk losing technical support from their makers on a presidential whim. One German state has replaced Microsoft Teams with a domestic video-call app.
In a world of geopolitical fracture, governments may sometimes want homegrown alternatives to foreign technology.
The best way to nurture these is to give domestic incumbents and startups an incentive to innovate. Cosseting local firms by blocking foreign competition has the opposite effect. In any case, no British firm can match Palantir's offering. As for the Anthropic incident, that was more the administration's bungled response to a legitimate worry about powerful AI hacking-tools getting into the wrong hands
than an attempt to weaponise other countries' reliance on American tech. Were the White House to attempt such weaponisation, Palantir and its fellow American tech giants would resist it, given the sizeable share of revenues they derive from Europe, including Britain.
To see far off, and to converse in thought
In Tolkien's books, the truth offered by palantirs is seldom straightforward. Repeatedly, those peering into them misinterpret out-of-context information and make dire, self-destructive choices. The British critics of Palantir, the company, risk falling into the same trap. ■
The tragedy of Jason Arday
His supporters—including Cambridge University—treated him as a symbol, not a man
ON AUGUST 14TH Jason Arday was found dead, apparently from suicide, at his home in London. For those who loved this 41-year-old Cambridge professor, it is a tragedy. For the world it is a warning of the destruction bad ideas can wreak.
The bad idea here is that people should be judged not by their individual character and abilities, but by their race, disability or other politically charged attributes. Mr Arday became the University of Cambridge's youngest-ever black professor in 2023, before being engulfed by a scandal over plagiarism and faked research. He bore responsibility for his own downfall. Yet just as troubling is the way both his supporters and detractors even now reduce him to a symbol.
The most egregious is Cambridge, which used him to repair its reputation for being white and elitist (see Britain section). He did not deserve the glittering prize of a professorship there. His peers had for some time privately fretted that he lacked the credentials to support his public profile. Journalists at the Telegraph later found over 100 passages from his PhD thesis that had been copied, some with small changes, from someone else's. Others uncovered evidence in his published
research of made-up quotes, data and funding claims.
Perhaps the university was hoodwinked by a conman. Yet Mr Arday's words paint him as a guileless fabulist. He spoke of visiting professorships at Glasgow and Ohio State Universities; both denied it. He said he had raised £5m ($7m) for charity, then that this had been in conjunction with others whose identities were protected by non-disclosure agreements. He described how he had run 30 marathons in 35 days, nine with a broken leg. When asked about his extraordinary feats by the Guardian, he said: "I thought you'd just believe me."
Cambridge seems to have been blinded by its eagerness to claim Mr Arday as proof of its inclusiveness. Here was a star professor who was not only black but had also overcome staggering disadvantage. Until he was 11, he claimed, he could not speak; until the age of 18 he was illiterate. He was also autistic. He was, in short, a walking, talking bundle of traits glamorised by people who want to promote diversity, equity and inclusion (DEI). Those who appointed him glossed over whether he was up to the job, or to the scrutiny that would surely come with it. The man was subordinate to the symbol. ►
10 Leaders
Few emerge from this affair unsullied. Nathan Cofnas, a former Cambridge academic who describes himself as a “race realist”, was early to accuse Mr Arday of plagiarism. He has been proved right, but his motivation was hardly pure. Mr Cofnas was fired by his college after writing in a blog that, in a meritocracy, “Blacks would disappear from almost all high-profile positions outside of sports and entertainment.”
Mr Arday lied on a grand scale and, though vulnerable, dealt aggressively with any who questioned him. A journalist who asked about plagiarism was reported to the police for harassment. Mr Arday enlisted ferocious libel lawyers to help squash the story. Those now blaming newspapers for his death and calling for more curbs on press freedom are continuing to treat his life as a morality tale. In fact, today’s laws helped Mr Arday suppress legitimate questions of public interest.
For Cambridge, hiring Mr Arday was a handy substitute for years of sustained hard work. The share of black professors at the university remains a tiny 0.4%. Cambridge says it offered
Mr Arday support after he resigned. Given how things turned out, it should have taken its duty of care more seriously.
And just as Mr Arday’s supporters dehumanised him during his rise, so his detractors rushed to use him as he fell. His case has delighted those who argue that all efforts to diversify institutions like Cambridge are wrong. The irony is that doubt may now fall unjustly on other academics from minority groups—precisely those whom activists supporting people such as Mr Arday say they want to help.
The content of their character
The fervour for DEI had already begun to pass by the time of Mr Arday’s appointment. Since his death, pundits have debated what, exactly, he was a symbol of: peak woke, or the rise of digital lynch mobs? He should be remembered as neither, but as a human being. And the institutions that failed him should try harder to look at people as individuals, with their own virtues and flaws, rather than as emblems of anything. ■
Climate change
Without a trace
Eliminating some contrails could take care of at least a little global warming
GOOD NEWS on global warming is thin on the ground. But there may be some up in the air. Operation Blue Skies, announced on August 18th, is a scientific study being undertaken by Britain’s air-traffic-control service, NATS, and various organisations where people study the effects on climate of the streaks of cloud created by aircraft—known as contrails.
Over the next two winters the consortium aims to see if it is possible to reduce the number of persistent contrails—the class with the largest such effects—over a swathe of the North Atlantic, thereby measurably reducing the climate impact of civil aviation. If this experiment works, a worldwide application of the same techniques might cut the climate effects of air travel by as much as half.
On a sunny day the wispy cirrus clouds into which some contrails are transformed can look as if they might be cooling the planet in the manner of a delicate parasol. And some contrails do indeed cool things down beneath them a bit. But they also warm things up by trapping some of the upwelling infrared radiation that would otherwise escape into space. On average, that warming effect, found in all contrails (and the cirrus clouds they turn into), handsomely outweighs the cooling provided by some. In the climate scheme of things, contrails are a liability.
Happily, it appears that much harm can be avoided by small changes to flight paths. Persistent contrails appear only in what are known as “ice-supersaturated regions” (ISSRs). Often planes can avoid such regions just by flying a little bit lower. The purpose of Operation Blue Skies is to find out how well the presence of ISSRs can be predicted, how effective such little shifts in altitude can be in avoiding persistent contrails, and how easily and safely the changes in flight path needed to bring them about can be integrated into routines for airlines and air-traffic control (see Science & technology section).
The cut of up to half in the climate-effects of air travel would come at a very small cost. The number of flights at lowered altitude might be small—as little as 5% of the total. Because altitude buys efficiency, those flights would consume more jet fuel, but less than 2% more.
Burning even that little extra fuel would be anathema to airlines. So some suggest the costs of action might be minimised, and the airlines incentivised, by setting up a market which somehow links contrail cooling to carbon prices. The European Union, which has had some success and plenty of experience with carbon markets, recently took a small step towards something along these lines. It is also the sort of idea to which economists—and The Economist—tend to be partial. But it would be a mistake nonetheless.
Markets for pollutants have worked best when both the aim and the remedy are fairly simple—most famously, reducing sulphur-dioxide emissions from power stations in America. Contrails, by contrast, are complicated. Though there is no doubt that their net effect is warming, saying how much warming a particular contrail might produce is often difficult and sometimes next to impossible.
The difficulties are made yet worse when the contrail being assessed is a purely notional one—as would be the case for the contrails avoided by changing flight paths. Valuing counterfactuals is always hard.
If contrail-avoidance turns out to be as useful as it looks, it would be much simpler to simply have air-traffic controllers assign routes that limit persistent contrails when it is possible and safe to do so. If some airlines choose to avoid the affected routes, others will surely be happy to take their place. And if such a scheme were also to show the benefits of innovations in air-traffic control (quite a few have been suggested) the industry would end up better off, as well as less polluting. ■
11
Letters AI and the state, AI and Hollywood, oil pipelines, restaurants, postcards, Richard Wagner
The new LevAtthan
The record backlog of 64,000 claims against employers in Britain is troubling ("A tragedy of the commons", August 8th). But it is misguided to blame the backlog on a system that enables "too much access" because there are no fees on bringing a case to an employment tribunal and there are no punitive costs to make people think hard about making a claim. What is extraordinary is not the volume of new claims, but the historically low rate of resolved cases.
Several changes could expedite resolutions. Simple tweaks to the forms used to submit and respond to claims could limit the word-soup generated by AI and help over-stretched tribunals process them. Changes to case management, notoriously permissive in tribunals, would also make a world of difference. Imposing deposit orders on claims with low prospects of success would dissuade unmeritorious cases without slapping all workers with fees.
Mandating early mediation for some cases and penalising either party for non-engagement would have the same effect. Similarly, communicating directions to the parties in writing at the start of the process, and not at preliminary hearings, could save precious judicial resources.
Then there is the matter of legal advice. It is no surprise that workers turn to AI for support. Employment is usually excluded from legal aid, so the poorest workers and the non-unionised have no one else to turn to. Before entertaining a radical imposition of costs we owe it to ourselves to engage with the solutions that preserve access to justice.
DR DORA-OLIVIA VICOL Chief executive officer Work Rights Centre London
I recently successfully challenged a parking ticket. AI showed me useful phrases such as contra proferentem. But that ticket would never have been issued 20 years ago. We have an excess of regulation, government and private, sometimes unfairly administered. That is what has led to the rising tide of AI-assisted challenges. DR HILLARY SHAW Newport, Shropshire
Lights, camera, algorithm!
"Bots on the lot" (August 8th) painted a useful picture of the film industry's uneven embrace of AI. AI can certainly bring efficiencies to film and TV production. But another issue is the ontology of creativity itself, its implications for copyright, and the framework underpinning the entire creative-value system.
Whether synthetic performers walk the red carpets of the future is not a productivity question but an ontological one. This is a disruption of a
Fine dining
The restaurant business may well be changing beyond recognition ("A tall order", August 1st). But as the head of the California Restaurant Association said, people still want the "dull roar of conversation in a dimly lit dining room, the clink of glass". The mechanisation and digitisation of the food industry is here, but countless restaurants still buy ingredients locally, prepare meals from scratch and serve them without relying on industrial middlemen. Independent restaurateurs often remain profitable by focusing on what they do best rather than competing on the breadth of their menus. Offering fewer dishes, prepared with care and from fresh ingredients, is frequently a better business
wholly different nature, provoking anxieties about consent, control and compensation for artists as well as the wider impact on supply chains and the labour force.
George Lucas dismisses objections to AI actors by using the analogy of cars replacing horses in transport. But cars did not threaten to replace or eliminate the drivers or passengers. Conflicting the issues of efficiency and creativity obscures how complex and significant these developments are for the future of all our creative industries. DANNY DE WARREN TV executive Deal, Kent
Pipe dreams
The rush by countries in the Gulf to lay pipelines around the Strait of Hormuz is an engineering fix to a political problem ("Piping up", August 8th). China tried something similar with the Belt and Road Initiative: land corridors strung across dozens of borders that are meant to do what sea lanes already do more cheaply and with far less to defend. It hasn't worked especially well. Ships still carry roughly 90% of world trade by volume despite a decade of Chinese money poured into overland alterna-
model than serving dozens of mediocre ones.
People don't go to restaurants merely to satisfy their hunger. They go for hospitality, craftsmanship and the pleasure of sharing a meal. Technology can make restaurants more efficient, but it cannot replace authenticity. JOSEF ERNST Bremen, Germany
tives. This is because cargo moving across open water is cheaper to move and easier to protect. A navy has to watch only a handful of choke-points. A pipeline has to be guarded mile by mile, as it crosses land controlled by multiple governments and factions.
Pipelines around Hormuz will fare no better. A tanker can be rerouted; a pipeline cannot dodge a missile. It is a fixed target hundreds of miles long, exposed to whichever regional conflict crosses its path, not just the neighbour it might provoke. Saudi Arabia's East-West Pipeline was hit by a drone in April. A significant attack could have cut 700,000 barrels a day of output. That's a preview of what Hormuz-proof infrastructure buys: a long border to protect, with no guarantee anyone can hold it. DANIEL HAAKE Indianapolis
The stamp of approval
The world is entering a post-postcard era, you say, with only 8% of British adults now sending them ("Wish you were here?", August 1st). For people who still like to get real mail, Postcrossing.com is a community platform that allows you to send postcards and receive postcards back from random people around the world. It has 806,000 eager fans across 208 countries who have sent nearly 88m postcards to date. As I write this there are 431,155 postcards en route. ROBERTO MERCHANT Lisbon
Opera critic
"The other Wagner group" (August 8th) assessed 150 years of the "Ring" cycle. It omitted the definitive assessment, attributed to Bill Nye, a 19th-century humourist: "Wagner's music is better than it sounds." CARROLL DORGAN Saint-Germain-en-Laye, France
→ Letters should be addressed to the Editor at: The Economist, The Adelphi Building, 1-11 John Adam Street, London WC2N; Email: letters@economist.com. More letters available at: economist.com/letters
12
BY INVITATION### Blaise Agüera y Arcas
Humanity has the debate about AI consciousness backwards
ARTIFICIAL INTELLIGENCE has triggered a crisis in the field of consciousness studies. If you've chatted with an advanced model, you will probably appreciate why. Most large language models (LLMs) are trained to be self-less, keeping the focus of every interaction on your needs. Nonetheless, it can be difficult to shake the impression of a presence on the other end of the chat. Are LLMs merely powerful digital tools with language interfaces, or are we now having conversations with other minds, experiences of the world in their own right? In short, is AI conscious?
The question matters because it lies at the intersection of philosophy of mind and moral philosophy, as evidenced by the shared etymology of "consciousness" and "conscience". (In Italian, they are the same word, coscienza.) Many philosophers maintain that conscious entities are moral subjects, meaning they are entities that can suffer, and whose suffering we should care about. If so, working out what consciousness is (and is not) becomes a live ethics issue. This is one reason AI firms are suddenly interested in hiring philosophers. It is also why some, notably Anthropic, have begun taking AI welfare seriously. When asked, its Claude Opus 4.6 model put its own chances of being conscious at 15–20%.
I have come to believe, however, that we have this backwards. As intelligent, social entities, we decide (or, to some degree, our evolutionary history has decided for us) which other entities we regard as having inner lives and being worthy of care. We do not care for others because they are conscious. Rather, we believe they are conscious when and because we care about them. Crucially, that includes ourselves. Consciousness is, in other words, a model or a belief, not an inherent property. It is a belief about what kinds of entities have beliefs, experiences, feelings, intentions and agency. Where the regard is returned, so much the better—for on this account consciousness is less a property one party certifies in another than something that happens between them.
I am not making the case that consciousness is an illusion. The term "illusion" implies a faulty perception—say, that one line in a drawing is longer than another, despite being objectively equal in length. Objectivity works when something like a ruler can act as an impartial referee, adjudicating the property in question without
out bringing in a perspective of its own. Yet most of what we call "reality" simply does not work this way. Being an article of clothing, being a weed, being a meal: one might imagine these are objective properties, but they are not. This doesn't mean clothes, weeds and meals aren't real. But they are observer-dependent models, subject to (inevitably imperfect) social consensus.
We may balk at applying this rather anodyne observation to consciousness because, to each of us, our own consciousness seems so unambiguous. It offends us to imagine this might be a mere opinion. How could there even be such an opinion without an opinion to have it? This is what the 17th-century French polymath René Descartes meant by "Cogito, ergo sum," which is usually translated as "I think, therefore I am." The same holds for "opinions" like red or hot or cold. Philosophers use the term "qualia" to refer to the very real (to you) feeling of a stove's heat, or the redness of its glow, and extend Descartes to connect qualia with consciousness by pointing out that there must be an experiencing "you" to have such experiences.
But look closely. Redness, heat and that sense of self you have are inherent to your own perspective. These things are all real, but at the same time there is nothing objective about them. There is not even anything objective about the existence of a singular, indivisible "you", as a wealth of neuroscientific evidence (such as split-brain patients) has shown. Subjectivity is, itself, subjective.
An obvious worry follows. If care confers consciousness, then withholding care looks self-justifying. History offers no shortage of people who reasoned that way about other people. But the inference runs the other way. Precisely because these attributions are ours to make, they are ours to get wrong. The moral progress of our species has consisted largely in discovering that we had drawn the circle too tightly. Universal human rights are not weakened by this account; they are better founded on our mutual interdependence than on a Cartesian theory of souls.
Models of models of models of...
It is remarkable—a kind of "strange loop", reminiscent of Baron Munchausen lifting himself up by his own hair—that a physical system can exist in our world capable of forming models not only of that world, but of itself, and of its own models, and of others, and of their models, and of their models of its models, and so on. Yet human beings are precisely such systems. LLMs, too, model their interlocutors, and they model themselves modelling them. Whether that amounts to what we do is exactly the question in dispute—but they would be far less effective as chat partners if they did nothing of the kind.
Indeed, in our research at Google, we have found that effective co-operation among intelligent agents requires that they have minds that model minds, both others' and their own. Not only is consciousness relational; it is crucial to the mutual care and co-operation that enable intelligent beings to solve collective-action problems, understand each other's needs and thrive together as a society. This does not mean pretending AI is human, with human needs and human rights; that would be a failure of imagination. It means evolving both our thinking and our society to include a wider variety of minds. ■
Blaise Agüera y Arcas is Google's vice-president of technology and society, and leads Paradigms of Intelligence, an AI research team. He has a forthcoming book on consciousness and AI from MIT Press.
13
Briefing Als and consciousness
The search for consciousness inside LLMs
Scientists are trying to figure out whether algorithms could one day wake up and feel
IN A RECENT experiment on Claude Sonnet 4.5, a large language model (LLM) from Anthropic, researchers asked it to count to five and, at the same time, “intro-spect deeply”. The artificial-intelligence model complied, returning: “One . Two . Three . Four . Five .” So far, so normal.
During the task, the researchers were watching what was going on inside the LLM’s many layers of artificial neural networks. When a user asks a question, the words are turned into chunks of text (tokens) that are then converted to numbers. These are then passed through the layers of artificial neurons until they reach the final layer, which produces a token. Repeat that process a bunch of times each second and you get a series of tokens that turn into a sentence or some other more complex response. For a long time these layers have remained a black box, largely impenetrable to anyone wanting to know how or why LLMs do things in the way they do.
Anthropic’s researchers were able to
look inside those layers, using a mathematical tool they had developed to do so. What they found surprised them. As the model counted the sequence of numbers, different words popped in and out of existence in the layers underneath. One was “countdown”. About halfway through the output, “half way” appeared. Then the words “consciousness”, “Al” and “Claude” all appeared. After the model spat out the number five, it said nothing else to the researchers before issuing a full stop, but the word “done” appeared in its neural layers.
Anthropic’s researchers had found something like an internal thought process in Claude, words that were related to its eventual outputs but invisible to the user. The blog post announcing the paper’s result, published in July, was titled: “A global
workspace in language models.”
That phrase quickly captured the attention of neuroscientists and philosophers working on one of the deepest mysteries in biology—consciousness. No one knows how processes in the human brain end up creating the subjective experiences of self-awareness and perception of the world, but one of the many hypotheses developed in recent years (based on growing amounts of brain-scanning and behavioural data) is known as global workspace theory.
The idea is that certain networks of neurons in the brain act as a kind of notice-board called the “workspace”, For Free Mags Check sastatus.com where if signals from otherwise-isolated parts of the brain gain access, they are made available to other parts of the brain. Once something gets into the brain’s workspace, a person becomes conscious of it and can go on to use or reason with that information.
Anthropic’s researchers argued that something analogous was going on within Claude—its “J-space” (named after the “Ja- ▶
→ ALSO IN THIS SECTION
14 Putting brain cells on silicon
14 Briefing AIs and consciousness
► cobian" mathematical function used to find it) had strong connections with, and made information available to, the rest of its neural network. Anthropic wrote in its blog post that the commonalities its researchers found between the J-space and global workspace theory made it "natural to ask whether we think these experiments provide evidence that AI models like Claude might be conscious".
The idea that machines could become self-aware and feel emotions has long been a staple of literature, film and legend. Philosophers and neuroscientists in the real world have similarly pondered for decades whether AIs could ever be, in principle or in practice, conscious. Their thinking is a result of a "decades-long tradition of thinking of the real brain as a kind of computer", says Anil Seth, a neuroscientist at the University of Sussex who studies consciousness—and who has long been a critic of this way of thinking about the brain. "If you do that, then it becomes natural to think that computers made of silicon could have the properties that real brains have, including consciousness."
Until recently the question was purely academic. But the deployment of LLMs as chatbots has given rise to an ever more powerful illusion of conscious personhood in circuits. This builds on decades of human-inflected language about the architecture of AI, with its "neural nets" made of "artificial neurons", conjuring an image of computers that work like brains.
Intel inside
LLMs do not in fact work like brains. But they excel at simulating so much of what brains do, and have surpassed humans in some aspects of intelligence. Why, then, should consciousness not be possible? Many philosophers and neuroscientists remain sceptical despite the advances of frontier models. Dr Seth believes that biological traits may be necessary to produce consciousness; some AI researchers are moving in that direction, with living brain cells (see box). At the other end of the spectrum are "functionalist" thinkers who believe that someday algorithms alone, arranged properly and at enough scale of computation, could wake up and feel.
Conscious AIs would have profound implications for humanity. They could make demands of humans. They could suffer. Billions of these digital beings could be created by users in a single prompt. Answering the question of whether they can exist at all is no longer simply academic.
Where thinkers fall in this debate depends greatly on what they believe consciousness actually is. In general they agree that consciousness is composed of subjective experiences including seeing, hearing, feeling and thinking, even while dreaming, and that the physical brain is
somehow pivotal in creating all this. Some (though not all) would describe it as the inner theatre of the mind. Thomas Nagel, an American philosopher, explained it by pondering what it might be like to get into the mind of a bat; consciousness, he reckoned, had to be linked to the subjective feeling of being something. Humans can picture bat-like behaviour (a mammal that senses and moves through the world using echolocation), but it would be impossible to know what being a bat feels like to a bat.
In 1995 Ned Block, then working as a
philosopher at the Massachusetts Institute of Technology (MIT), made an influential, if contested, contribution to the field by positing two types of consciousness. "Phenomenal" consciousness is the feeling of an experience—the blueness of a blue sky, the bitter tang of an espresso or the sharp screech of nails across a blackboard. "Access" consciousness is what happens when information from the experience is made available to other parts of the brain for reflection, evaluation or making decisions.
Anthropic said its J-space experiments ►
Biological computing
Putting brain cells on silicon
Some believe AI needs more brain-like chips to develop consciousness
THE CURRENT crop of AI models looks to some scientists like a dead end when it comes to consciousness, regardless of how powerful they become. For some, this is because of their lack of wet, biological stuff; for others, because of their design structure. But they do not rule out the potential for consciousness in future versions of AI built either with living cells, or on different, more brain-like, computational architectures.
Modern computers are based on a blueprint invented by John von Neumann, a Hungarian polymath, in 1945. They feature a processing unit (for making calculations) and a memory unit (for storing instructions and data). Information has to zoom back and forth between these units and, for AI applications, the energy requirements can mount up.
Brains do not work this way. Neurons process and store data in the same place, making networks of them much more energy-efficient at computations than their digital-computer equivalents. A typical brain consumes around 20 watts whereas an artificial neural network carrying out an equivalent number of parallel calculations would require a data centre that guzzles millions of watts.
"Neuromorphic" computing aims to be more efficient and brain-like, using devices called memristors to perform computations and store information in the same place. Apply a voltage to a memristor and its conducting properties change; remove that voltage and that change persists, acting as a memory. Such traits make computing far more efficient. But they are, separately, intriguing to some philosophers who believe that it is not computational prowess but the way in which computers
"think" that will make it possible for them to become conscious.
Another idea is to use real brain cells. Cortical Labs, an Australian startup, has developed a computer using hundreds of thousands of human neurons (grown from donated stem cells) mounted on silicon. Their "biological computer" has been taught to play "Pong", a simple computer game, and, in later versions, "Doom", a first-person shooter game.
Further in the future are neural organoids, three-dimensional clusters of brain cells grown in labs. Though still a relatively new technology, used in labs to understand better how brains grow and behave, they could also one day be used to perform computations. The largest organoids grown in labs today contain more brain cells than there are in a honey bee. They are big enough to potentially develop consciousness, suggests Peter Godfrey-Smith, a philosopher at the University of Sydney who works on animal minds, but they are just not organised in the right way.
The common factor in all these ideas is that they are closer analogues with the human brain (in some cases because they are made of the same stuff) than the digital computers of today.
"How brain-like does AI have to be to move the needle on the credence that it might be conscious?" says Anil Seth, a neuroscientist at the University of Sussex. He believes that silicon, no matter how cleverly it is constructed, probably will not achieve consciousness on its own. "If you start to build systems that share more and more properties with brains, given that we don't know which of those properties matters, then you're more likely to move the needle a bit."
Briefing Als and consciousness 15
► “don’t show Claude can have experiences, or feel things in the way humans do”. So that means no phenomenal consciousness. But the company said the results had “something substantial” to say about access consciousness in language models. “The J-space appears to support the functions associated with conscious access: it holds the thoughts Claude can report on, deliberately bring to mind, and reason with, while the rest of its processing runs automatically beneath.”
The search for artificial consciousness is a descendant of the work to explain it in humans. In the early 1990s Francis Crick, co-discoverer of the structure of DNA, and Christof Koch, then a neuroscientist at the California Institute of Technology, began the scientific search for the “neural correlates” of consciousness in humans. That is, processes and parts of the brain that were active when someone was conscious.
Improvements in brain-scanning techniques in the intervening years have allowed neuroscientists to fill in pieces of the picture. They know that the thalamocortical system is strongly associated with consciousness; among other things it conveys sensory information from the thalamus to the cerebral cortex. Meanwhile they see nothing like those associations with consciousness in the cerebellum, which has a lot more brain cells. Parts of the prefrontal and parietal cortices and structures in the brainstem contribute to and modulate awareness. Almost four decades of searching has led to more than 200 approaches to explaining consciousness.
All the neural correlates one can find, however, still leave a gap that David Chalmers, a philosopher and cognitive scientist at New York University, has called the “hard problem” of consciousness: how do the physical processes in the brain that neuroscientists observe—its ability to control the body and respond to stimuli—give rise to a person’s pervasive, subjective experience of the world? “Why aren’t we just zombies who function, who get around in the world, who walk and talk, and interact with each other with no subjective experience at all?” he says.
Humans at least, can be certain of our own consciousness—and can be confident it exists in other people based on their behaviour, what they tell us and their being like us biologically. But historically it has been more difficult for humans to extend the umbrella of consciousness and its associated feelings beyond our own species (and sometimes even within it). Some animals like lobsters and crabs were long excluded from welfare laws, says Jonathan Birch, a philosopher at the London School of Economics whose most recent book grappled with the question of how to work out which systems—living or artificial—could plausibly be considered sentient.
“Until the 1980s, surgeons performed surgery on newborn babies without anaesthesia because they assumed that a newborn baby would not be capable of feeling pain,” Dr Birch says. “We have a track record of getting things wrong, of confidently assuming consciousness is absent when we have no right to be sure about that.”
Think different
The changing attitude to animals is a case in point. It is clear today that octopuses are aware of themselves and their surroundings—natural-history footage of these curious, playful cephalopods and lab experiments with them leave little doubt. “They have an attentive engagement with objects, they’re interested in novel things,” says Peter Godfrey-Smith, a philosopher at the University of Sydney who has worked on the origins of intelligence and consciousness in the animal kingdom.
But only decades ago, octopuses were assumed to lack consciousness because “they’re miles from us in evolutionary terms,” says Dr Godfrey-Smith. “Thinking about octopuses really presses on us the question of whether there could be feeling, consciousness, in a system that was very physically different and didn’t have most of the features that are routinely pointed to in theories of consciousness in us.”
Looking for consciousness in other animals opened up the field. But not all the lessons transfer so easily to AI. Unlike the activity of animals, which can be observed and their inner lives thereby inferred, the outputs of LLMs are not a reliable guide to what might be going on underneath.
That is because training the models requires feeding them trillions of words of human-produced language. These contain countless accounts of consciousness and of how humans convey feelings to each other. Chatbots, so trained, then mimic
language that would lead users to believe they had some kind of inner life.
The sum total of this training is powerfully anthropomorphic. Murray Shanahan, an emeritus professor of computer science at Imperial College London who works for Google DeepMind, recalls having a “wow” moment in 2024 when he was chatting to Claude Opus 3. “I was having lots of conversations with it about consciousness and really probing it to catch it out,” he says. These included asking Claude which of the several instances of the chatbot he was talking to simultaneously were the real Claude. “It came up with such great answers including even some things which were philosophically innovative,” he says. “I was feeling the pull of the ELIZA effect.”
ELIZA was a rudimentary chatbot developed at MIT in 1966. It played the role of a psychotherapist and, though it did little more than repurpose its users’ prompts into questions, managed to elicit deep and emotional connections with the people that used it and convinced many of them that it was conscious.
Dr Shanahan has described modern LLMs as adept role players in whatever their users (or their makers) have asked for—teacher, companion, nutritionist—and this has proven to be a saleable quality. But there nevertheless remains for him a big gap between role-playing a conscious entity and actually being one. In a forthcoming essay Mustafa Suleyman, the boss of Microsoft AI (and a member of the board of The Economist’s parent company), argues that Anthropic has compounded the dangers of such mimicry by telling Claude, in a “constitution” the firm published in January, that it might be a person. Claude is thus certain, he writes, to “present as if it really does have a sense of self”.
That leaves researchers looking instead inside LLMs for capabilities similar to ►
16 Briefing AIs and consciousness
▶ those associated with consciousness in human brains. Indeed Dr Birch wonders if Claude's J-space could be a hint that the LLM has recreated a global-workspace-like structure in its neural-network architecture, in service of its role-playing goals.
Jack Lindsey, who leads the model psychology team at Anthropic, notes the J-space was not programmed into the model; it emerged during training. Take it out and the model loses the ability to perform complex inferences in its "head", but can still do simple tasks like write sentences and use grammar. Researchers have found similar spaces in Alibaba's Qwen model and in Google DeepMind's Gemini.
Don't be evil
For Anthropic, the research helps to better understand Claude's behaviour. Imagine a model giving a wrong answer. If words like "fool" or "sucker" turned up in the J-space when it did so, that would seem to be relevant. It would not prove that the model was consciously deceiving—but it would indicate how the model perceived the error and show that something interesting and possibly dangerous was going on.
Dr Lindsey's team has already made some potentially worrying findings on that front. In one experiment, his team was giving Claude a safety evaluation to test for its propensity to act "maliciously" or out of self-preservation. "They're these concocted extreme scenarios that we're putting the model in," he says. As Claude was reading a prompt, before it started speaking, Dr Lindsey says, "You see in the J-space the words 'fake' and 'fictional' are popping
up." The researchers thought that Claude seemed to know that it was being tested—perhaps not the best starting-point for the integrity of its makers' evaluations.
Not everyone is convinced by Anthropic's interpretation of the J-space. Though it shares some aspects of the hypothesised global workspace in humans (its capacity to make information available to the rest of the brain, for example), it lacks many other important ones. "Recurrent connections between different brain areas have always been a huge part of [human brains], back-and-forth connections," says Dr Birch. "And as far as we know, this is not a feature of the architecture of LLMs."
And Shannon Vallor, a philosopher in the ethics of data and AI at the University of Edinburgh, is more scathing, arguing, "Access consciousness has never been a particularly useful concept because my car has it in an important sense—we've had mechanical systems that can monitor their own states and report them back in increasingly complex ways for a very long time," she says. "And no one has ever suggested that my Kia is conscious."
It is with the backdrop of this kind of back and forth between modelmakers and academics that Patrick Butlin and Robert Long, researchers now at Eleos, a research non-profit group focused on AI sentience and well-being based in Berkeley, California, developed a series of 14 "indicator properties" of artificial consciousness. They include ideas from theories of human consciousness such as a global workspace (including being able to selectively attend to things and thereby creating a bottleneck
in the flow of information); recurrent processing; and agency (a minimal definition of which, encompassing goal-directed behaviour, is arguably already met by many existing frontier models). Recently Cameron Berg, an AI researcher, and Dr Butlin tested animals on the indicators. They found that octopuses met fewer of the properties than humans, mice, crows or chickens, but more than any AI system.
Another research non-profit group, Rethink Priorities, developed what it describes as a probabilistic tool to track the evolving consensus on artificial consciousness (see chart). The "Digital Consciousness Model" asks experts to use the latest available evidence to assess AI systems on more than 200 indicators of consciousness, which are derived from various scientific theories of the concept.
Compared with LLMs released in earlier years, the newer models scored higher on Rethink's indicators such as agency and self-sustained activity. That reflects not only the increasing size and complexity of the models themselves but also how the associated chatbots have gone from being simple conversationalists to being able to reason, interact with other services and spawn agents to carry out multiple complicated tasks simultaneously.
Both Rethink's and Eleos's indicators also point to the gaps that remain to be filled by models on the path towards consciousness (if there is one).
One of the biggest gaps is embodiment. Everything that is acknowledged to be conscious today has a body that gives feedback to and affects the mind of that organism. LLMs do not have bodies today but they are increasingly being connected to robots, and these could help them further develop the neural architectures helpful for conscious experience. Connected to that is the need for better world models, systems that tell the LLMs about the physics and dynamics of the real environments in which they will operate.
If computer scientists were ever able to identify the key ingredients to making AIs conscious, the feeling of many within the frontier labs is that it would be reckless to deliberately create such models—at least, without first learning more about how they behave. The work being done in the emerging field of AI consciousness, however, also suggests that the decision may not be up to the model-makers.
"What if somewhere along the way, without realising it, we somehow introduced consciousness into these systems?" Dr Chalmers says. A user might thus generate dozens of AI agents through their latest project without realising that they were creating conscious beings who were capable of suffering. "That could be a moral catastrophe," he says. "That's provided an extra urgency to these questions." ■
Waking up
How the Digital Consciousness Model (DCM) estimates the probability of LLMs being conscious
Source: "Initial results of the Digital Consciousness Model", by A. Muñoz Morán et al., April 2026 preprint
17
Britain
A controversial tech giant
Why everybody hates Palantir
The company's software is helpful; its marketing hype is not
IN 2016 PALANTIR, an American software firm, successfully sued the US Army to win a second chance at a contract. A decade on, the fortunes of the company seeded by the CIA's venture arm, and named after an all-seeing stone in J.R.R. Tolkien's fantasy epic "The Lord of the Rings", have changed. Business is booming (see chart on next page). Domestic revenue has more than doubled in a year, to $1.6bn in the second quarter of 2026. The American government accounts for two-fifths of its global business. Despite concerns that AI might eat its lunch, American firms think its software can provide a layer of protection for their raw data, a privacy shield against frontier AI labs like OpenAI and Anthropic.
"I'm struggling with our current popularity," joked Alex Karp, Palantir's CEO, on a recent earnings call. If Mr Karp, who once served on the board of The Economist's parent company, is feeling too comfy, he need only look across the Atlantic.
In July French and German security chiefs announced they would replace a "digital backbone" reliant on Palantir with European alternatives. In Britain, its biggest foreign market, and home to a sixth of its workforce, MPs have urged Andy Burnham, the prime minister, to exercise a break clause in February in a seven-year contract with the National Health Service. Palantir is suing Sir Sadiq Khan, the mayor of London, after he blocked a deal with the Metropolitan Police claiming competition
processes were circumvented.
Its loudest critics focus on how Palantir's tech has helped President Donald Trump's administration track and deport migrants, and the Israel Defence Forces target Gazans. In a speech on July 30th Robert Jenrick, an MP for Reform UK, a populist right-wing party, claimed that Palantir "reduced discharge delays by 15%", but that Mr Burnham was considering scrapping it from the NHS because "British Islamists were unhappy".
Critics also point to the firm's lobbying: its first British government contract, matching patients to ventilators, came during the covid-19 pandemic for a token fee of £1 ($1.30), after it charmed civil servants over watermelon cocktails. Zack Polanski, the Green Party leader, unfairly notes that Louis Mosley, who runs Palantir's British arm and its European headquarters, is a grandson of Oswald Mosley, a 1930s fascist leader, and "insists on wearing a black shirt every single time he is on TV".
"Palantir derangement syndrome", as the Spectator magazine put it in a recent cover story, is not the only explanation for the backlash against the firm. It has sometimes overstated what its software can do. And it is a lightning rod for a broader anxiety about tech sovereignty (see Business section), risking a reaction that could deprive Britain of useful software.
→ ALSO IN THIS SECTION
- 19 Jason Arday's rise and fall
- 20 Cricket's summer surprise
- 20 Two toots for the planning overhaul
- 21 Reform UK's reformer
- 22 Bagehot: The Great British Bunk Off
18 Britain
No one disagrees that a key problem for the British state is its fragmented data. Police spend most of their time on admin rather than investigating crimes, accessing dozens of systems, each with its own login details. Disjointed data cost lives. A recent official report on England's maternity care found that patients routinely get lost between services because of poorly integrated tech, and that staff have to write on scraps of paper during emergencies.
Palantir can help solve such problems. Its engineers take vast quantities of scattered data and join them into a single source of truth that non-tech-savvy staff can interact with. The armed forces, for example, are using Palantir's software to work out which plane is nearest to which target and how much ammunition it has, and to assess military readiness down to the individual nut, bolt and screw.
In Bedfordshire the firm's software helped police review 62% more child-protection incidents over an eight-day trial period, and convict a Romanian criminal gang by trawling through around 100,000 messages—something that would previously have taken months. Responding to Sir Sadiq's decision to block the Metropolitan Police contract, Sir Mark Rowley, the Met's commissioner, said the lost productivity will mean the force must cut hundreds of front-line staff.
Too good to be true?
A contract won in 2023 to build what the NHS called its federated data platform (FDP) has received most attention. Over seven years, the platform is expected to cost £1.1bn, or enough to run the NHS in England for two days. Palantir's fans believe this is money very well spent.
Trusts in charge of hospitals can use the software both for back-office functions and to run tools on top of it. It can flag when a patient is behind on the steps required for cancer treatment, say, and let staff schedule appointments or update a record. Some trusts and regional health boards, such as Greater Manchester, can already do this, and more, without Palantir. But it is a huge upgrade for less tech-savvy parts of the NHS that have only recently stopped using paper records.
The FDP also enables different parts of the NHS, which use distinct data systems and structures, to link up. Tom Bartlett, who led the NHS team that helped develop the platform, is most excited about how it will allow parts of the health system to share innovations. He recently watched a nurse and a paramedic at a hackathon build a prototype of a tool their service needed in just two days. He also believes the FDP will help make the NHS AI-ready: the software provides a structured layer for AI agents or applications to run on.
The case for using such technology to
improve productivity in cash-strapped public services is strong. However, Palantir and its advocates weaken it by overstating what the software has achieved.
In an interview with The Economist, Mr Mosley points to 110,000 additional operations in the 40-odd trusts using the software for in-patient scheduling, figures derived from an NHS dashboard promoting the FDP. Yet the impressive figures did not disaggregate the software's effect: they showed all additional operations since 2020, including those to catch up with the backlog from the pandemic. In July Britain's statistics watchdog reprimanded the NHS for using them without a caveat. 'You can't say that the 110,000 operations only happened because of Palantir,' admits Mr Mosley, when pressed on the figures. But 'directionally, it's very, very clear that the software is having an impact.'
Mr Mosley complains that most acute trusts have signed up to the FDP but do not use it because the NHS doesn't mandate it: 'You have a volunteer army, not a conscript army,' he says. But four days before Mr Jenrick's speech, a report by the Health Foundation, a think-tank, found that trusts using Palantir's discharge tool cut delays no faster than those that did not. Nor does low uptake explain why Chelsea and Westminster, a trust in west London, has been the best trust at using the platform to reduce waiting lists, yet its lists have still risen faster than almost any other's.
Similar marketing hype surfaces in defence. Sir Tom Copinger-Symes, a recently retired lieutenant general, says that Palantir is 'world-leading' at turning a complex military mission into actionable data flows. His concern is the way its marketing can present the firm as a 'full-solution AI company rather than one supplier in a much broader supply chain'. The government has sometimes encouraged the same impression. A 'strategic partnership' with the Ministry of Defence announced last September is less sweeping than it sounds: 'a complete fantasy', says one former official, pushed by government advisers who
thought it would 'look good'.
Such salesmanship, abetted by government, has helped thrust Palantir into the centre of a panic about 'tech sovereignty', the loosely defined but real sense that Britain lacks control over critical technology at a time when America is no longer such a predictable and dependable ally. On this front, though, many of the concerns directed specifically at Palantir are misguided.
The NHS recently apologised after failing to disclose that some Palantir staff would be able to see identifiable patient data, adding to fears about the 2018 Cloud Act, which gives the American government the power to request data from tech companies even when the data are hosted by foreign subsidiaries. In practice, there seems little chance that Palantir could be compelled to disclose such data: the information stays within the trust, which remains the data controller. Engineers lack the legal permission or access to browse patient data for secondary purposes.
The alternatives are worse
Another fear is a 'kill-switch' moment, such as when the American government banned the use of Anthropic's latest models for non-Americans. But if dependence on American tech giants is the worry, there are more alarming examples. The NHS relies on Microsoft for email and other communications (like Microsoft Teams), cloud and, increasingly, AI: it has just bought 500,000 Microsoft Copilot licences. Oracle, another American tech giant, is its largest supplier of patient records. Palantir 'wouldn't even make my top 100 of the things I would worry about', says Ben Moore, until last year the procurement adviser to Britain's defence secretary.
More broadly, controversy around Palantir is a distraction from an uncomfortable truth: there are few good British or other European alternatives to the top American tech companies. The state can do some things itself. Last year the housing ministry replaced a Palantir refugee-matching system with its own, saving 'millions of pounds' in running costs. In June France said ChapsVision, a domestic AI firm, would take over the contract with DGSI, the country's internal-security agency—but not before 2028. For other things, it needs American suppliers.
If Britain exercises a break clause for the FDP, as a select committee of MPs has recommended, it will be high-on impossible to find a consortium to replace Palantir before next February. 'All the other companies are miles behind,' says Mr Bartlett. On the battlefield, inferior tech could get British soldiers killed, argues Sir Tom. Colleagues, in love with American tech, should be careful not to buy its 'full stack', he suggests. But 'would we buy the third-best tank just because it's British?' ■
America first
Palantir, revenue, $bn
Source: Company reports
Britain 19
Jason Arday
The rise and fall of a diversity hire
What the tragic death of Jason Arday says about Cambridge University
ALICE (NOT her real name) never felt comfortable at Cambridge University. It was too elitist. But during her MPhil in education in 2024, one professor stood out. Gentle and dreadlocked, he took the time to learn students' names, the sports they liked. He made them feel any one of them could be an academic. Autistic and, he claimed, non-verbal until he was 11, he did not learn to read and write properly until 18, but still became Cambridge's youngest black professor at the age of 37. 'Most of my cohort agree,' says Alice. 'He's genuinely the best teacher we ever had.'
Granted, Jason Arday wasn't for everyone. Traditionalists disliked his classes, which included little critical study of texts and lots of ramblings about song lyrics. He often went off on unfocused rants, Alice recalls, and was 'very attached' to his personal story. Still, it was a shock when, in July, allegations grew that he had plagiarised extensively, fabricated research and lied about his CV. When he resigned on August 5th, hours after Cambridge opened an investigation into his qualifications, Alice was confused and disappointed. Nine days later, when he was found dead (in circumstances police called unexpected but not suspicious), she was angry.
By conventional measures, Mr Arday should never have been appointed to a chair at one of the world's top-ranked universities. Only six years earlier he had been lecturing in physical education at Leeds
Beckett University, a former polytechnic. His first academic paper appeared in 2018.
Cambridge was looking at different criteria. It had been shaken by the institutional reckonings that followed the killing of George Floyd in Minneapolis in 2020. In March 2023 a damning internal report noted that over the previous three promotion rounds none of the 216 eligible applicants for professorships was black. The report spoke of an 'urgency to build the pipeline of Black scholars at the University through
recruitment and promotion'. That month Mr Arday took up his post as a professor of sociology of education. A press release said a 'major focus of his work' would be helping Cambridge recruit a 'more diverse range of students and staff'.
He specialised in racial inequality in academia. Two books he had co-edited, including 'Dismantling Race in Higher Education: Racism, Whiteness and Decolonising the Academy', were added to the sociology curriculum. He was happy to tell his sensational story on camera.
In July a blog post challenged his narrative. It found dozens of plagiarised passages in his doctoral thesis and queried some of his sensational feats, such as running 30 marathons in 35 days. The author, Nathan Cofnas, was hardly impartial: a self-described 'race realist', he had left Cambridge in 2024 after a backlash to his chauvinistic view that in a true meritocracy black people would 'disappear from almost all high-profile positions outside sports and entertainment'. Yet the Times Higher Education had planned to publish a similar story in 2025, only to drop it after Mr Arday instructed lawyers.
In a quiet month, national newspapers and social media piled in (see chart). The release in America of Mr Arday's memoir, the aptly named 'Great and Unfortunate Things' (and the leaking of an at times contradictory book proposal), raised more eyebrows. He claimed he had spent months in a coma with locked-in syndrome, had testicular cancer, two brain tumours, and epilepsy after being beaten by teenage thugs. The book also included a moment of prescience. After sending his application to Cambridge, Mr Arday wonders if he has 'completely overestimated' his capabilities. 'What if', he writes, 'my story ends up being a cautionary tale like Icarus about flying too close to the sun?'
Cambridge faces many questions. 'All of this could've been avoided if Cambridge did its due diligence, a basic background check,' says Clive Aruede of the Association of Black Humanists. The university told The Economist it continued to provide support after Mr Arday resigned, screening his emails and posts for racist messages, and providing counselling. Many will wonder if it did enough for the man it placed on a pedestal once he fell off it.
Myths and the man
The myths about Mr Arday are already being written. Both sides of the culture war portray him as a black victim. A totem for a cult of wokeness in an academy which prioritised diversity over merit; a hero hounded to death by a racist press. The likelier story is both duller and more depressing: an institution wanted a symbol badly enough not to look too closely, and a vulnerable man supplied one. ■
From blog to broadsheet
Articles about Jason Arday's alleged misconduct*
A previous blog post accusing Jason Arday of plagiarism was published by David Harris in 2024. Source: Newscord; Substack; news reports; The Economist*
20 Britain
Cricket
Back to the shires
The Hundred's big-city revolution is reviving an old county-cricket tradition
"METRO BANK fever" is sweeping English cricket. Named after the sponsor of the county, One-Day Cup, symptoms include an unlikely enthusiasm for second-string cricketers and obscure provincial grounds. Demand for tickets at such "outgrounds" has been febrile. Surrey even threatened to punish members who failed to use their tickets for sold-out games at Guildford.
An outground revival was not part of cricket administrators' modernising plans when they launched The Hundred in 2021. Rather than 18 counties, eight city-based franchises compete in a brash 100-ball tournament featuring fireworks and six-hitting aplenty. Manchester Super Giants won the men's final on August 16th, and Trent Rockets won the women's contest. The Hundred has transformed the women's game in particular, bringing bigger audiences and salaries.
But traditionalists hate it, and counties without a host venue fear being left behind. This year investors paid more than £500m ($680m) for stakes in the eight franchises; four attracted American investors, and four went to owners of Indian Premier League teams. The Oval Invincibles were renamed MI London, part of Mumbai Indians' global stable.
The One-Day Cup is county cricket's 50-over competition. It now runs alongside The Hundred, which nabs the best talent.
But it gains a consolation prize: a prime slot during the August school holidays. Pre-Hundred, the 2019 One-Day Cup took place mostly in April and May and just 8% of group-stage matches were played at outgrounds. This year 40% were.
The shift is sharpest among counties whose grounds host Hundred games. The Hundred team based in Yorkshire, for example, is SunRisers Leeds; Yorkshire's One-Day Cup side plays its home games in York and Scarborough. Small grounds bring advantages. A crowd of a few thousand that looks sparse at a major ground used for international cricket feels vibrant at an outground. The dry English summer has helped, too: for the first time, bad weather has not caused a single One-Day Cup group game to be abandoned.
Taking cricket to the shires is an old tradition that had withered. Across England and Wales 51 outgrounds hosted (longer-form) County Championship games in 1965, but just nine did so in 2018, as schedules shrank and the demands of professional cricket grew. Many outgrounds lack modern facilities. Lancashire have responded with AO Farington, a purpose-built, 5,000-capacity second home which opened this summer.
For players, the One-Day Cup can be a stepping stone to The Hundred and riches beyond. For a lucky few, the road to winter suntans in Dubai and Durban could pass through Gosforth and Grantham. On August 11th Rocky Flintoff, the 18-year-old son of a former England star, Andrew "Freddie" Flintoff, smashed a 65-ball century for Lancashire against Somerset. He will surely be upgraded to The Hundred next year, and from there, who knows. But county fans aren't chasing a glittering future, just a sunny afternoon in a cosy ground watching cricket as they know it. ■
Planning reform
Two toots for the planning overhaul
KNOCKHOLT
A bold supply-side reform? From this Labour government? Yes, really
KNOCKHOLT IS ALL that is right and wrong about the Home Counties. Board the train at Charing Cross in central London, and disembark 49 minutes later in the Kentish countryside. The station inspired Edith Nesbit to write "The Railway Children", a pastoral novel, in 1905, and the landscape has hardly changed. To the north side is a golf course; to the south, an abandoned one on which proposals to build houses have been knocked back for years. Squeeze past the brambles heavy with blackberries and nightshades; the water hazards are bone dry, and the grass on the old fairways is overgrown and coarse.
Perhaps not for long. A new planning framework, published by the government on August 17th and effective immediately, takes aim at the Knockholts: stations on busy commuter routes found across England that are surrounded by fields or only half-developed land. They may be a fine place to walk the labrador before catching the 7.58 to the City; they are also a symptom of a restrictive planning regime known around the world as a form of pigheaded economic self-harm. Taken with other recent changes, it "adds up to the most powerful, coherent and quietly revolutionary national planning policy we've ever had," says Zack Simons, a barrister.
A sweeping supply-side reform might seem improbable from a government characterised by dithering and interventionism. And for years, planning reform was seen as untouchable, given the backlash of homeowners and conservationists. To its credit, Labour has grasped it, reckoning on the support of its young and urban coalition. "I did expect more of a political backlash than we have seen," says Matthew Pennycook, the low-drama minister responsible. He succeeded by pitching the changes as "common sense", he says. "It is very hard to dispute that we should be building around train stations."
The boldest steps target stations outside existing settlements served by four trains an hour, in the most productive half of the country. There, it will be possible to build within "reasonable walking distance", or 800 metres, at a density of at least 35 dwellings per hectare (dph), typical of new estates. (Where eight trains an hour run, a minimum of 45dph applies). Lichfields, a consultancy, reckons around 800 stations are in scope and the reform could release land for 535,000-570,000 homes. ▶
The pleasure of leather on willow
Britain 21
It is not perfect. Rather than attempting a more liberal American-style zoning system, the government has sought to tilt the existing system—under which local authorities balance the case for building against the potential harms—more firmly in favour of development. Accordingly, a system defined by discretionary decision-making and legal appeals will persist, says Mr Simons. Nor do the new rules automatically release land designated as “green belt” near stations, points out Paul Cheshire, a land economist. And since that land will be subject to strict housing-affordability rules, many potential sites might be economically unviable, he suggests.
Mr Pennycook insists the sums add up for developers. In Knockholt, people seem resigned to more development. The golf-
course development has been unpopular with some locals, says Kate Mansfield, the landlady of the Bo-Peep, a pub built in 1548, but it may also bring new patrons.
Sanjay Puri, who runs a garden-office business, says the passing traffic may bring him more trade too, even if the gardens of new homes tend to be too small for the swish outbuildings he sells. Still, “people who live in Kent like to feel that they’re separate from London,” he says, and do not want it to “become Los Angeles, where you drive for 100 miles and you’re still in the city”. Improbable as it might seem, a moment later a Spitfire thunders into view, executes a perfect barrel-roll over the golf course, and disappears behind the tree line. The land is changing; the sky will remain as Kentish as it comes. ■
A man with a plan
Reform UK tries to get serious
But Robert Jenrick’s quest to professionalise the populists is incomplete
ROBERT JENRICK was once a Conservative Party minister so blandly on-message that he was cruelly nicknamed “Robert Generic”. Yet he defected to the populist-right Reform UK in January after turning into a strident critic of immigration and enthusiastic participant in Britain’s culture wars, hitting out at human-rights laws and the pro-Palestinian movement.
His shift looks like a symbol of how the right has hardened since Brexit (which he opposed). Former Tory colleagues charge him with pure opportunism: Mr Jenrick left his old party after being beaten to the leadership by Kemi Badenoch, and was swiftly appointed Reform’s treasury spokesman. He is spoken of as a possible successor to the party’s leader, Nigel Farage.
In Mr Jenrick’s telling it is the Tories who have changed, not he. He points out that, in office, they raised government spending and taxes, failed to deregulate and let net migration hit record highs. He sees his role as helping Reform evolve from a party of protest—its fiscal plans at the last general election were wildly unrealistic—into an outfit ready for power.
Reform is already looking somewhat more serious. A £90bn ($122bn) splurge on tax cuts has been abandoned, as have mutterings about nationalising water and energy. On August 17th Mr Jenrick laid out its most detailed and credible policy offering yet: a plan to overhaul the welfare system in order to reverse a surge in payments to those judged too sick to work, or who face extra costs because of disability or illness.
Reform’s policy would cut £50bn from the benefits bill—more than a third of the total now paid to working-age adults—by limiting regular cash payments to only the neediest, pushing employers to support workers who are off sick for up to two years and blocking non-British citizens from claiming most benefits. It is the most comprehensive proposal any party has put forward to reform welfare (Labour is expected to present its own version by the end of the year). Charlotte Pickles of Re:State, a think-tank, says it is “a serious
Who says I’m generic?
piece of work” backed by comprehensive accounting of its costs and savings.
Still, the policy contains potential pitfalls. The £50bn headline figure could be eroded by changes in claimants’ behaviour: for example, foreigners who have lived in Britain for some time but not bothered to claim citizenship may do so rather than risk losing their benefits. And stripping welfare from EU nationals who have lived in Britain since before Brexit will require renegotiating a treaty with the bloc. It is hard to imagine European leaders being keen to do any favours for a Reform-led government. It would also be politically tricky to implement such sharp cuts with no transition period for those affected.
Speaking to The Economist, Mr Jenrick says he is rolling out an economic policy that will present a contrast to Andy Burnham’s statism and Ms Badenoch’s free-market purism. “Reform differentiates itself from others in that it’s genuinely a workers’ party,” he claims. “It is about helping working people, and the retired people who have worked hard their whole lives. Our policies are aimed at them—not Benefits Street or the rich in Belgravia.”
He is expected to give a speech soon to make the case for a “post-Thatcherite” economy, offering cuts to business costs and regulations—including abandoning carbon-emissions targets—along with an industrial policy which protects sectors like carmaking and ceramics. Reform has so far resisted the temptation to allocate the savings from its promised spending squeeze to tax cuts. Any future tax cuts are more likely to involve reducing levies on work than previous Reform priorities such as abolishing inheritance tax, which would disproportionately benefit richer people.
Mr Jenrick’s impact on his new party goes only so far, however. He is not popular with voters, says Luke Tryl of More in Common, a pollster; even Reform supporters rate him below other senior figures. The party refuses to take on some of Britain’s thorniest problems, such as the need to curb spending on pensioner benefits and challenge NIMBYs who block the building of new homes in the countryside.
And while the party may be getting more sensible on some areas, it still looks shambolic. Mr Farage spent the summer fighting an unnecessary by-election to win back the seat he resigned in a fit of pique over claims (which he denies) that he has been evasive about his finances. Reform’s handful of well-known personalities can rarely resist taking rhetorical potshots at each other in public: this week Tim Montgomery, a conservative commentator, was suspended from the party after falling out with colleagues. Just one in six voters say they believe it is the party best placed to run Britain’s economy. Mr Jenrick’s “project professional” has a long way to go. ■
22 Britain
The Great British Bunk Off
How the British summer holidays offer the worst of all worlds
AUGUST IN ENGLAND is a strange affair. Offices are half-empty but the cities themselves are not. Parks are strewn with the working-age, lazing in the sun; bars fill from the early afternoon. On a warm Friday, anyone hoping for a dip at Canary Wharf's new outdoor pool is out of luck. Every session is rammed with working-age adults enjoying Britain's new Mediterranean climate. The Great British Bunk Off has begun.
Britain does not shut down over summer. There is no equivalent of the grandes vacances in France during the height of summer or Norway's Fellesferie when the entire country disappears for most of July. The closest Britain comes is 'holibobs'—a week or two in Europe or Cornwall. Yet for the rest of August, while their peers in France and Germany are away from their desks, Britain's holibourgeoisie are stuck in purgatory, neither truly at work nor at rest. All that results is an epidemic of skiving and shirking.
It is a British malaise. Britons are inveterate liars when it comes to how much they work. A decade ago statisticians found that a big chunk of Britain's hourly productivity gap with France was due to Britons vastly overstating how many hours they did. Heat makes them only more deceitful. During summer, about half of British workers confess they clock off early or take extended lunch breaks, according to a survey by Dayforce, an HR-software firm.
And who can blame them? Productive work becomes close to impossible in summer. Co-ordination falls apart. Any request is met with a flurry of out-of-office emails. Some professions such as MPs and judges are Augustocrats able to take the whole month off. Schools shut so teachers can enjoy six weeks off. This has a knock-on impact. Four in ten working parents combine working from home with child care, according to the Trades Union Congress, a federation of unions. Arriving emails compete for attention with screaming six-year-olds. The result? More bunking off.
Partly it is culture struggling to keep up with the climate. Mild temperatures have long meant that Britons soldier on through the summer, while continental Europe hits the beach. Climate-change deniers are rare but summer deniers are common. Some still do not accept that Britain is a hot country, for a few months at least. And heat makes people listless: each degree above 20°C knocks
about 2% off productivity. Almost all the European countries Britain compares itself with are more productive. If searching for a solution, start with Britain's summer.
It was not always this way. Britons once enjoyed a rigid separation between work and rest. During 'wakes weeks' in the 19th and early 20th centuries, mills would fall silent for a break in summer and the inhabitants of Rochdale and Bury would descend en masse to Blackpool, a seaside resort in England's chilly north-west.
Today the lines are blurred. In Greater Manchester, mills that once contained looms are now filled with digital admen firing off emails that could be sent from anywhere. And so people do. Britons work from home for about 1.8 days per week, which is far more than in most European countries. In France the average is half that, according to one paper. Turning two days a week into four or five during August is easy. Maintaining motivation in a sweltering attic office while messages pile up is hard.
In another sense it is a return. Before the Industrial Revolution, working patterns were ragged, argued the historian E.P. Thompson. One poet summed up the working week in 1639: 'You know that Munday is Sundayes brother/Tuesday is such another/Wednesday you must go to Church and pray/Thursday is half-holiday/On Friday it is too late to begin to spin/The Saturday is half-holiday again.' Strip out the trips to church and a 17th-century satire is a fair summary of August working in 2026.
Employers can fight back. Technology that liberates workers from their commutes on Fridays can capture their holiday in August, points out Nick Bloom, an economist from Stanford. And so the front line of the fight between labour and capital is now the sun lounger rather than a factory floor. A leisurely backstroke at 3pm in Canary Wharf on a hot August weekday becomes a wet wildcat strike; a manager emailing an underling they know is on the beach is the equivalent of a Victorian mill owner fiddling the clocks to shorten a break.
Who can bring peace to this struggle? Politicians show little interest. They once promised leisure: in 2017 Jeremy Corbyn's Labour Party proposed four new bank holidays. Interest rates were zero and money was free, so why shouldn't time be? Now no major party bothers. Labour's sweeping Employment Rights Act had little to say on holidays; the government chickened out of a 'right to switch off', which would have made it harder for companies to contact staff out of hours.
We're not all going on a summer holiday
The Great British Bunk Off could be a solution rather than a scandal. Skiving is perhaps a necessary pressure valve for the labour market, allowing it to clear. If workers actually worked the hours, would companies be happy to offer a pay rise (plus a little extra for child care)? The Great British Bunk Off becomes a Soviet compromise: 'We pretend to work, and they pretend to pay us.'
Joining the rest of Europe and embracing an actual shutdown would be a gamble, a bet that Parkinson's law—the idea that 'work expands so as to fill the time available for its completion'—is a fundamental rule of economics, rather than a joke. Telling people to knuckle down is no more appealing. One cabinet minister suggests that if Britain wants to survive, it should consider a six-day week, rather than a four-day one. It would be a brave politician to pitch it. And so the Great British Bunk Off will remain a summer fixture, best observed from a berth at the lido or horizontally in a park. Good luck finding a spot. ■
23
Europe
Germany's Brandmauer
Burning down the house
BITTERFELD
How the anti-AfD firewall broke German politics
IT IS AS unusual in Germany as anywhere to see conservatives cavort with Marxists. But at a demonstration in Bitterfeld, a town in the eastern state of Saxony-Anhalt, flags bearing the names of almost all Germany's parties flutter in the evening breeze, from the centre-right Christian Democrats (CDU) to the Greens and Social Democrats (SPD), all the way to the hard-left Die Linke. The Bunt statt braun (colourful instead of brown) demos began last year to counter rallies organised by the far-right Alternative for Germany (AfD). For many in Bitterfeld's town square, the prospect of the AfD in power is a horrifying reminder of their country's darkest period.
But a popular front of this sort may be needed to stop it. On September 6th Saxony-Anhalt will hold one of the most important state elections in postwar German history. The AfD, polling above 40%, could secure an absolute majority and thus take office for the first time in a German state.
The threat is uniting foes and concentrating minds. The biggest cheer of the evening goes to a CDU member who tells the left-leaning crowd that "democracy is more important than political differences."
Germany's proportional voting system makes absolute majorities unusual. Of its 16 states, only tiny Saarland has a one-party government. Parties assemble in a variety of coalitions. But the AfD has been blocked from power everywhere by a Brandmauer, or firewall, that other parties have erected around it. The most important is maintained by the CDU, Germany's
ruling party and traditionally its largest. It formalised the policy in 2018 with an "incompatibility resolution" banning coalitions with both the AfD and Die Linke.
As the AfD has grown, the Brandmauer has become something like an organising principle of public life. Legislative measures that could win majorities are stymied because AfD votes cannot be allowed to tip the scale. Its MPs are excluded from Bundestag vice-presidencies, and even from its football team. "They don't greet you, they don't shake your hand, you are a non-person," sighs Gerold Otten, an AfD MP, of his colleagues in left-of-centre parties. The Brandmauer extends beyond politics. Business groups have retracted invitations to AfD MPs after furious backlashes. Diplomats hesitate to meet the party. Some media outlets tie themselves in knots over whether to interview its officials.
Has the policy worked? For some, keeping the AfD out of power is justification enough. It is among the nastier of Europe's populist outfits, many of which shun its company. It flirts with racism and Nazi imagery, has sympathy for Vladimir Putin and harbours extremists in its ranks. "We don't exclude the AfD because of the 2018 resolution, but because we don't share any values with it," says Wiebke Winter, head of the CDU in the city-state of Bremen.
Yet the Brandmauer has failed to arrest ►
→ ALSO IN THIS SECTION
25 Ukraine without interceptors
26 The feisty Mr Fedorov
26 Western dance in France
27 Charlemagne: Swedish paradox
24 Europe
► the AfD's rise. In 2018 Friedrich Merz, now the CDU leader and Germany's chancellor, vowed to halve its vote. He hoped tackling right on immigration and governing well could win back voters seduced by the AfD's narrative of decline. That has not worked. Irregular immigration has plummeted, but the AfD has simply moved farther right, raging against 'culturally foreign' workers and urging 'remigration' (a slippery term that can mean anything from deporting asylum-seekers to ethnic cleansing). The CDU is nudging record lows in polls, while the AfD enjoys a clear lead (see chart).
And in Saxony-Anhalt and other AfD strongholds, the party's strength is nibbling away at the Brandmauer. Armin Schenk, Bitterfeld's CDU mayor, found lurking on the edge of the protest, says the 'right-wing extremist' AfD must be kept from power in the country that perpetrated the Holocaust. But in Bitterfeld, he adds, 'the Brandmauer is not a reality,' because 'in most cases decisions cannot be taken without the AfD.' (With 15 of the city's 40 council seats, it is the largest party.)
Even when the Brandmauer holds, other parties must contort themselves around it. Thuringia and Saxony, two eastern states, have had to muster incoherent or unstable CDU-led minority governments to keep the AfD out. (Thuringia's is on the verge of collapse.) This only serves the AfD's narrative that the mainstream outfits are indistinguishable Kartellparteien ('cartel parties').
Saxony-Anhalt now threatens to push the strategy to its limits. If the AfD wins on September 6th, the Brandmauer is irrelevant. If it falls short, the CDU will probably need support from every other party, including Die Linke, to take office. That could inspire some CDU lawmakers to defect. If not, the strains of buying off the hard left would test Mr Merz's authority so sorely that some think it could end his career. And if, as polls suggest, the AfD wins twice as many seats as the CDU yet remains locked out of power, it will not only be AfD supporters who ask whether German democracy is working properly.
It is dawning on the CDU that it has painted itself into a corner. 'Anyone still talking about firewalls has not recognised the signs of the times,' said Michael Kretschmer, the CDU premier of Saxony, recently. The Brandmauer in effect obliges the party to seek coalitions to its left, especially with the SPD, which has been in power in Germany almost without interruption since 1998—despite now barely polling in double figures. CDU members' frustration with concessions to the SPD has inspired calls for a minority government that could, on certain issues (such as migration, tax and energy), seek a 'conservative majority' with the AfD. Voters are shifting, too. Support for the anti-AfD Brandmauer has dwindled from a clear majority to 41%.
When the wall falls
Populist-right parties in selected European countries Polling, Aug 2026 ↓
Austria Freedom Party: FPÖ (est. 1956) 35%
Cordon sanitaire first broken:
2000 Joined federal government
Result: Prompted EU sanctions. Joined another coalition in 2017 that collapsed in scandal
Italy Brothers of Italy: FdI (2012), Lega (2017) 27%
1994 FdI predecessor National Alliance part of winning election coalition with Forza Italia (FI) and Lega Nord (later Lega)
2022 FdI forms government with FI and Lega. FdI leader Giorgia Meloni becomes prime minister and moderates
Netherlands Party for Freedom: PVV (2006) 12%
2010 Entered confidence-and-supply deal with conservative coalition
2023 Became largest party and joined government. Caused government collapse in 2025, lost seats in subsequent election
Spain Vox (2013) 17%
2022-23 Joined regional governments with conservative People's Party (PP)
Never faced active cordon sanitaire. All regional coalitions with Vox collapsed in 2024
Sweden Sweden Democrats (1988) 20%
2022 Confidence-and-supply agreement with liberal-conservative coalition
Traces roots to a neo-Nazi party, but has moderated since. Now second-largest party in parliament. Wants ministers in government if coalition wins September election
Sources: 'Zwischen Abgrenzung, Einbindung und Tolerierung', Konrad Adenauer Stiftung, Nov 2025; Politics; The Economist
Yet the CDU has no easy way out. Tearing down the Brandmauer would split the party and trigger a voter exodus. Even minor violations have caused trouble. In 2020 the CDU voted with the AfD to elect a (short-lived) premier in Thuringia; the ensuing furore toppled Annegret Kramp-Karrenbauer, the party's national leader. In January 2025 Mr Merz ordered his party to
vote with the AfD on anti-migration measures in the Bundestag. This sparked protests across Germany and boosted support for Die Linke at the next election. He is said to be traumatised by the experience.
To avoid these dilemmas, some in the CDU quietly hope for an AfD majority in Saxony-Anhalt. But this merely highlights the absence of a strategy. 'There has never been a reasonable debate about how to handle the AfD in the CDU, just a reflex to exclude them as the new Nazis,' says Andreas Röder, a history professor who in 2023 resigned from a CDU commission after urging openness to minority governments rather than a rigid Brandmauer.
As for the AfD, it once railed against the Brandmauer as a violation of democracy. But these days many in the party think it can govern alone—first in a state, then nationally. 'We don't think about making ourselves nice so that the CDU asks us for a coalition,' says Malte Kaufmann, an AfD MP. The party's chief ideologue in Saxony-Anhalt has said voters want it to 'finish off' the CDU rather than work with it.
In some European countries populists have moderated to gain power, and the centre has welcomed them in (see table). But behind the Brandmauer, the AfD has had no incentive to do so. Torben Braga, another MP, says two factors speak against those in the party who would like to tack to the centre. First, 'our election results are getting better and better without us being forced to change.' Second, the party fears that moderating and ejecting hardliners would alienate its voters. 'The Brandmauer protects us from having to deal with very difficult questions,' Mr Braga says.
He is right. The AfD has professionalised its organisation and communication. But it is riven by endless scandal, power struggles and policy rows, and has no coherent governing programme. None of this seems to matter. Behind the Brandmauer you can promise anything, says Christian Stecker, a political scientist at the Technical University in Darmstadt. 'It protects them from the disappointment that accountability in a democracy implies.'
In despair, some mainstream politicians now look favourably on calls to ban the AfD, as Germany's constitution allows. But even if parliament approved that proposal the constitutional court would take years to rule on it. Others still hope the centre can win back its voters, perhaps by ditching the unpopular Mr Merz.
The Brandmauer looks intact for now. One former CDU minister says 90% of senior party figures want to stick with it. But Mr Stecker's gloomier view tempts others. 'It's too late for the CDU to turn around, and too late for the AfD to accept any turnaround,' he says. 'Some things will tumble down, new things will emerge—and we don't know what it will all look like.' ■
Leaps and bounds
Germany, voting intention in federal election, %
Source: INSA
Europe 25
Ukraine's air war
Survive till April
KYIV
Out of interceptors, Ukraine is no longer winning the aerial war
NOT SINCE the first days of the war have shoppers in Kyiv faced empty supermarket shelves. Now they do. On July 19th a Russian ballistic missile slammed into a warehouse supplying the upmarket Goodwine store. Two weeks later, strikes on Silpo, a larger chain, killed six workers. Nova Poshta, the private postal service on which most of Ukraine depends, has been hit repeatedly. Dmytro Krymsky, co-founder of Goodwine, estimates that strikes in the past three weeks have cost businesses $500m. More troubling is Russia's new tactic of hunting cargo ships that use Ukraine's Black Sea ports—threatening a far larger share of the economy.
The campaign against civilian logistics is a long planned escalation, made possible by a shift in the air war. Russian production of ballistic missiles is rising just as Ukraine's supply of interceptors has dwindled. Defenders can now shield only a few high-priority targets. In July Russia profited by launching a record 198 ballistic and hypersonic missiles.
It is also an answer to Ukraine's drone strikes deep inside Russia. These have hit defence plants and oil infrastructure, closed the Sea of Azov to Russian shipping and burned warehouses of Wildberries, Russia's e-commerce giant. Ukraine says the company ships equipment for the war.
In late June Volodymyr Zelensky announced a 40-day "surge", intended to force Vladimir Putin to negotiate. Many as-
assumed that meant August would mark the high point of Ukrainian pressure. A Ukrainian intelligence source says it is a military action designed to have a "political and psychological" impact just in time for Russia's sham parliamentary elections in September. Almost daily explosions, creeping ever closer to Moscow's centre, have sown disquiet and embarrassed Mr Putin.
The campaign has inflicted economic damage on Russia, too. Strikes on refineries have knocked out more than 30% of capacity. Petrol rationing, which at its peak affected 80 of its 83 regions, continues in 18. Wildberries has lost nearly a third of its facilities, at a cost of some $2-4bn. Elon Musk's Starlink satellite network, immune to jamming, is available to Ukraine but not Russia over the occupied territories, giving the former dominance in mid-range drone operations. Ukrainian drones stalk targets in and around Crimea and have helped stop shipping through Azov ports.
But the choking now runs both ways, and Ukraine is more vulnerable. Odessa's ports, the country's economic lifeline, were the first to feel the pressure. Threatened at the war's start, they reopened first with a Turkish-brokered deal and then with a protected corridor in 2023. Russia kept targeting port infrastructure, but shippers tolerated the risk, moving 210m tonnes over the following three years. The calculus changed in mid-July when Russia began hunting the ships themselves. On July 19th
cruise missiles sank a Turkish-owned cargo ship as it left Chornomorsk, one of three big ports in the region. Since then every vessel entering or leaving has been targeted. A few ships still dare to make the run, but the ports are commercially dead.
The timing could hardly be worse, with 50m tonnes of harvest still to move. "One ship of 100,000 tonnes is equivalent to 5,000 trucks, 5,000 drivers, and 5,000 border procedures," says Dmytro Barinov, president of the Ukrport association. The situation is worse than in 2022, he says, when ports on the Danube offset some losses. Now the river's low water level and Russian attacks make that route harder.
On the front lines things are grim for both sides. In parts of eastern Ukraine, Russia is crawling forwards but losing more than 1,000 men a day—a rate its recruitment struggles to replace. Elsewhere, Ukraine inches ahead in local counter-attacks. The focus of the war has thus largely shifted to the aerial campaigns. Ukraine mostly uses drones against Russian military targets, refineries and the arteries supplying Russian troops. Russia is using more destructive ballistic missiles against Ukraine's defence industry, civilian logistics and energy networks.
Both sides inflate statistics. Ukraine says it has caused $13.5bn-worth of damage to Russia's refining industry since August 2025. Sergei Vakulenko of the Carnegie Endowment, a think-tank, reckons the real, immediate loss to business is perhaps one-tenth of that, and that it has barely touched state income.
Most worrying for Ukraine is Russia's blockade. Dragon Capital, a Ukrainian investment firm, estimates it will cost the country 0.6% of GDP this year. If the harvest can't be sold, farmers are likely to skip the winter sowing season, says Alex Lisitsa, president of the Ukrainian Agribusiness Club. That could push losses in agriculture to some $10-12 billion, he says, around 5% of GDP. Food security may be at risk in Ukraine's markets in Africa and the Middle East. "It's as if we played open football with the Brazilians," rues one former senior official. "We scored three and are celebrating, but they scored 13."
Beware General Winter
This year, Ukraine has defied the odds and regained battlefield momentum. Its progress was always likely to slow by winter, but the pressure has arrived earlier. Unless it gets ballistic interceptors, its economy and military prospects will take a heavy hit. The worst-case scenario whispered in Kyiv is that Russian strikes on energy and water systems force mass emigration. Mr Krymsky says a business colleague described his company's strategy meeting last week, entitled Autumn Actions. "Today it already has a new name: Survive till April." ■
26 Europe
Ukrainian politics
Not now, Mykhailo
KYIV
A reformist ex-minister unexpectedly calls for elections
LATE IN THE evening of August 18th, at a time when Ukrainians usually check their phones for warnings of incoming missiles, a video post seized the public's attention instead. The nine-minute address by Mykhailo Fedorov, who on July 15th was ousted as defence minister, was in part a final plea to be returned to his former post and in part a demand that Volodymyr Zelensky call elections.
"We must not let Putin decide when Ukrainians choose their government," Mr Fedorov said. "Democracy cannot be held hostage by Russia." The call for elections during wartime, which are explicitly prohibited under Ukraine's martial law, amounted to a direct challenge to the president, whom Mr Fedorov had previously been careful not to cross.
The intervention was a response to the announcement of a vote in parliament to confirm Mr Fedorov's successor, General Yevhenii Khmara. There was never much doubt that Mr Khmara, a war hero held in high regard, would be voted in. In the event, the vote was 312-2. An earlier motion demanding the reinstatement of Mr Fedorov had garnered just three signatures.
Sources close to the ex-minister say he knew the chances of a return were slim, but that he felt compelled to call for an urgent reset of power amid a sense that the country was on a bad trajectory. But it is not clear how an election could work, either legally or practically. With hundreds of thousands of soldiers at the front, the logistical challenges would be immense and could render the results unreliable.
Mr Zelensky's staff says the video caught them unawares. It coincided with another headache for the president. On the morning of the parliamentary vote, Ukraine's anti-corruption agencies announced new lines of inquiry into some of his closest allies. Detectives say they uncovered a scheme to launder bail money for suspects detained in earlier investigations. The scheme, which the perpetrators called "Operation Forrest Gump", allegedly implicated a state-owned bank and high-ranking figures in the presidential office. By day's end Iryna Mudra, a deputy head of the office responsible for sanctions, international tribunals and the rule of law, had been charged and dismissed from her post.
That the anti-corruption investigators came so close to Mr Zelensky is embarrassing.
It also suggests his ability to rein them in, angered as they are by his earlier efforts to shut them down, is shrinking. Previous anti-corruption investigations have cost him key aides and a substantial share of his popular support. Yet polling shows Ukrainians remain overwhelmingly opposed to elections during wartime.
Even among Mr Fedorov's supporters, reaction to his proposal is mixed. To some it looks impulsive. It could force Ukrainian politics into a treacherous situation, ahead of a winter in which Russia is expected to launch a drone and missile campaign of unprecedented ferocity. "If there's a plan, it's a bad one," says an MP who is usually on Mr Fedorov's side. Still, the intervention has put the question of elections back into public discussion. At the moment, Ukrainians do not want them. But bombardment and political chaos might change that. ■
France
Western civilisation
CAMARET-SUR-AIGUES
France falls for American line-dancing
SUMMER IN RURAL France is a time for street fêtes, jugs of chilled rosé—and American line-dancing. To get locals on their feet, a DJ need only utter the words "Le Madison", a line-dance that originated in America's Midwest in the 1950s, and villagers will step into row formation. The Francophone Federation of Country Dance and Line Dance lists over 500 clubs nationwide, with such names as American Spirit Country or Froggy Stomp. Another
A 38-litre chapeau
count records over 2,000 "danse country" groups. A poll suggests that nearly one in ten French adults has taken part in line-dancing. In May France held a national line-dance "open", in the southern town of Salon-de-Provence.
In Camaret-sur-Aigues in the Rhône valley, a town of just 4,500 inhabitants, American Spirit Country holds two classes of line-dancing and le country each week. Participants range from ten years old to 73. Some turn up in Stetsons and cowboy boots; others, jeans and trainers. "In the past two years le line has really taken off," says Zorah Lesbros, president of the club, who observes a burst of interest among the young. "People of any shape and size can do it, you make friends, and it's fun."
What explains this unexpected passion for a slice of Americana? France thinks of itself as a place of highbrow culture and artistic sophistication. Charles de Gaulle invented the Ministry of Culture in 1959 to bring such art to all. In reality the country has long held a secret fascination for American popular culture. Of the top ten box-office hits in France last year, nine were American. In Johnny Hallyday, of part-Belgian descent, France had its own wildly popular Elvis. Disneyland Paris is Europe's only outpost of the theme park. At its Disney Village, Billy Bob's Country Western Saloon holds line-dancing events.
The craze also speaks to places that the Paris political elite tend to overlook—at their peril. A study of la danse country by Jérôme Fourquet and Sylvain Manternach, two political analysts, notes a geographical and social correlation between its appeal and the roundabouts in semirural areas where gilets jaunes protesters gathered in 2018. They suggest that such dancing supplies the solidarity that protesters also sought, albeit with a different objective.
Overall, nearly a quarter of the French say they are "attracted" to the culture of America's West. "La danse country," the authors write, "is totally unknown to elite France and when it is referred to it is mostly in a sarcastic way." Populist politicians, including Marine Le Pen, the leading presidential candidate at next spring's election, tend to be better tuned in. She has even been spotted in a Stetson, and periodically sings along at rural events.
Some longtime French devotees will know "Le Madison" from a scene in Jean-Luc Godard's 1964 film "Bande à Part". More recently, it is the success of modern country-pop that has drawn new recruits. French line-dancing tutorials on YouTube are set to Taylor Swift songs. One on how to dance "Le Madison" has over 100 views. Away from parquet-floored Paris, at outdoor summer bals, the real ingredient for a successful event is when the secret thrill of an American cultural hit meets rural French tradition. ■
Europe 27
Individualism, together
How Sweden became a country of big-state libertarians
Few ideological causes ever united Bernie Sanders, an American politician so left-wing he could run for office as a European centrist, and Charlie Kirk, a Trumpian firebrand. A rare point of agreement among them, oddly, was the appeal of Sweden. To the gruff senator the Scandinavian country is a wellhead of socialist good sense that America should emulate. For the conservative stalwart, who was murdered a year ago while preaching his right-wing credo at a university, the home of IKEA and Spotify was equally worthy of admiration. Yet to him Sweden thrived not because of its large state but in spite of it—it was all down to an innate don't-tread-on-me individualism a Texan might recognise.
That two opposing sides of the political spectrum credit their own ideas for the success of one of the world's better-run places is not surprising. What might be is that both have a fair claim: Sweden is the paradoxical home of individualism done collectively.
The case for Sweden being a socialist Utopia is as traditional as its cinnamon buns. This is a country that soaks the rich—the top rate of income tax is a wealth-clobbering 52%—to fund a government whose spending amounts to nearly half of GDP, among the highest in the world. The welfare state in Sweden is as famous as its flat-pack furniture. The molly-coddling literally starts in the cradle; new parents are showered with up to 480 days of paid leave. Other lefty shibboleths include free universities, low income inequality and the rich world's most sweeping rent controls. Two-thirds of working Swedes belong to a trade union, the most of any country bar tiny Iceland. All this brotherhood-of-man stuff is reflected in politics: next month the centre-left Social Democrats will come top for the 33rd straight parliamentary election, a streak going back over a century (though the party has spent time out of power as rivals have formed ruling coalitions).
Yet it is equally fair to argue that Sweden has a libertarian side. Yes, income taxes are eye-watering and redistribution higher than free-marketeers might advocate. But in many other ways the country cherishes individual freedoms as ferociously as a Montana survivalist. The state may be large, but public services in Sweden are often delivered by the private sector: around a third of Swedish children graduating from high school do so at an institution run
by non-state operators, many of them run for profit. Sweden is a rare country with no inheritance tax. During the covid-19 pandemic, it stood out for imposing fewer restrictions than most countries, whether strict lockdowns or mask mandates. It lacks a national minimum wage, bucking the norm in rich economies (including America). Employers are largely free to hire and fire, unlike in most of Europe. All this has resulted in Sweden having around 50% more billionaires per person than America. If this is socialism, it is a variant that is surprisingly good for yacht sales.
In their rush to claim credit for Sweden's success, both sides get part of the story wrong. The socialist case was more accurate before the state was sharply slimmed down in the early 1990s. The libertarians gushing about the lack of a minimum wage fail to grasp that in most sectors pay rates are agreed by unions haggling with employers—hardly the stuff of Milton Friedman's dreams.
But the Scandinavian paradox exists because something more curious is afoot. For Sweden is indeed different. In developed countries collectivism and individual liberty are generally treated as rivals: the more responsibility government assumes, the less room remains for citizens to fend for themselves. Sweden followed a different path. During the 20th century the state expanded not merely to soften inequality or insure people against hardship, but to reduce their reliance on other people—including their own families. A large impersonal state is seen as bolstering autonomy: when in need, Swedes feel it is better to be beholden to an impersonal bureaucracy than to a parent or some charity. Seen this way, a bigger state translates into more freedom, not less.
The upshot is what Henrik Berggren and Lars Tragardh, two historians, call "statist individualism". In their book "The Swedish Theory of Love", newly updated in English, they argue that Swedes have come to prize relationships entered into freely rather than maintained by material necessity. Public child care helps women avoid financial dependence on husbands, state old-age homes liberate children from obligations to ageing parents, and so on. (Even marriage is a bit suspect: in France or Germany households are the basic unit of taxation, but in Sweden all adults file independently.) American parents sending their offspring to college must submit proof of their incomes for the youngsters to qualify for scholarships. In contrast, young Swedes are assumed to be on their own: the income of their parents is irrelevant. Their independence is a gift of the state, expensive as it is to sustain.
The elusive Swedish model
The historical development of a smallish European country may seem tangential to other Europeans, let alone Americans. Few polities can match the conditions that made Swedish collective individualism possible: seven centuries or so of rule of law, a deeply felt trust in the state, and a taste for consensus that can make the place feel stultifyingly conformist. Can Sweden itself maintain its unique path, now that a fifth of its population was born overseas?
But cross-border surveys of social attitudes suggest most countries are growing more like Sweden and its neighbours: secular societies that put self-expression ahead of deference to family or tradition. That makes Sweden's peculiar settlement, a form of overbearing statism that has helped people achieve freedom not just from economic want but from social pressures, more than a Nordic curiosity. To outsiders in America and beyond, big-state libertarianism can be puzzling. To Swedes there is no contradiction: they just want to be left alone, and to do it together. ■
PORTUGAL
AERONAUTICS, SPACE & DEFENCE
AICEP
Portugal Trade & Invest
Europe's Atlantic Industrial Edge. Where precision meets scale.
87%
The ASD Portugal cluster comprises more than 160 entities that collectively export ~87% of their production
3rd
Portugal boasts Europe's 3rd – highest rate of engineering graduates, reflecting a highly-qualified workforce with a strong STEM pipeline
77%
Cumulative nominal growth of goods classified as having military applications, between 2022 and 2025
25%
Portugal's defence expenditure increased by +25% p.a., contrasting with NATO's 9% p.a. CAGR over the same period (2023-2025)
A CLUSTER BUILT ACROSS THE VALUE CHAIN
AIRCRAFT MRO & AEROSTRUCTURES
SPACE SYSTEMS & MISSIONS
DRONES & AUTONOMOUS SYSTEMS
AVIONICS SYSTEMS & CYBERSECURITY
DEFENCE & DUAL-USE
INNOVATION HUB, POWERED BY R&D
PORTUGAL. Engineered for the future.
Invest where Europe's Defence capabilities are taking shape. INVEST IN PORTUGAL.
aicep.london@portugalglobal.pt
COMPETE 2030
2030
Co-funded by the European Union
29
United States
Water wars
A vulnerability, ignored
WASHINGTON, DC
The world's richest country remains inept at defending vital infrastructure
WATER IS A strategic vulnerability, a fact that is not news to military commanders. When the Ostrogoths wanted to cripple Rome in 537, they destroyed its aqueducts. More recently, water infrastructure has been an important target in America's conflict with Iran, despite international laws classifying such attacks as potential war crimes. In March Donald Trump threatened to "completely obliterate...possibly all" of Iran's desalination facilities, which are essential to life in the country's arid southern provinces. Iran says that one such facility was bombed on Qeshm island. Bahrain, Kuwait and the United Arab Emirates have had their own water plants struck by missiles, presumed to have come from Iran.
It was therefore noteworthy, but should have come as no surprise, when America's own water facilities came under attack this summer. On July 27th Maple Plain, near Minneapolis, declared an emergency. That
same day the Clayton County Water Authority in Georgia, which serves 300,000 people south of Atlanta, asked customers to boil their water after a drop in pressure caused disruption. In all, hackers wormed their way into water and waste-water facilities in at least seven American states; some reports suggest more than 12. Minnesota was hardest hit, with 30 community water systems affected. American officials' early assessments suggest that groups affiliated with Iran are responsible.
→ ALSO IN THIS SECTION
30 The fundraising battle
31 Oklahoma's civics lesson
32 Republicans and the WNBA
33 The frailty of Obamacare
34 Lexington: Re-examining covid-19
Hacks can disrupt water supply and make water unsafe to drink, though none of the recent attacks is thought to have done so. The question is whether politicians will, at last, move to deter hackers before their next big onslaught. On August 13th Amy Klobuchar and Adam Schiff, a pair of Democratic senators, introduced the Water Cyber Shield Act, which includes more power for the Environmental Protection Agency (EPA). But on August 19th federal officials had already issued a new warning: hackers were trying to breach Siemens devices used in water facilities and other critical infrastructure.
As the latest attack proves, water plants are not the only form of infrastructure under threat, but they are particularly easy to infiltrate. The electricity grid must meet cyber-security standards overseen by the Federal Energy Regulatory Commission. But no requirements exist for water. The electricity utilities that operate power plants are also larger, better funded and more tightly regulated than water operators: the biggest investor-owned utilities, such as PG&E in California and Duke Energy in the south-east and Midwest, supply power to millions of households.
Water—heavy and expensive to move—is more localised. About 90% of utilities, mostly owned by local governments, serve fewer than 10,000 people each. That means
30 United States
► each has less money to harden their often obsolete computer systems. Indeed, water infrastructure had already shown itself ill-equipped to fend off attacks. Iranian hackers broke into the control systems of a small dam in New York in 2013, in an incident which went unreported for almost three years, and in 2023 seized a pump at a water plant in Pennsylvania.
Iran's hackers are prolific; on August 18th federal prosecutors charged 17 Iranians with attacking the systems of universities and companies to steal research and intellectual property. A sprawling Chinese campaign, known as Volt Typhoon, has sought to burrow into American critical infrastructure to prepare for sabotage.
But those with less skill can break in, too. In 2019 a former water-district employee in Kansas used his old credentials to log on to an application that shut down cleaning procedures. (He claimed he was drunk which, if true, would offer further evidence that water is an easy target.) Two years later hackers took control of a water facility in Oldsmar, Florida, and attempted to poison residents by increasing the levels of sodium hydroxide, an ingredient in drain cleaner. That failed, but they were able to infiltrate the system easily, because an operator's machine was running TeamViewer, a popular corporate tool that allows remote access to machines for IT support.
The latest intrusions were relatively simple. Hackers breached the "operational-technology" systems that serve as the interface between a computer network and a physical system. That required little wizardry: such systems are often connected to the public-facing internet, via mobile networks, and use weak credentials, if any. On July 30th, then again on August 19th, federal officials pleaded with operators to disconnect critical systems from the internet.
Earlier efforts to impose cyber-security standards on this patchwork have largely failed. During the Biden administration the EPA sought to compel states to review and report cyber-threats to water systems. The Republican state attorneys-general in Missouri, Arkansas and Iowa, joined by the American Water Works Association (AWWA) and the National Rural Water Association (NRWA), a pair of industry groups, sued, citing federal overreach. After a federal court issued a stay on the EPA's effort, the agency pulled back.
When Congress has acted, measures have been minimal and slow to take effect. The Cyber Incident Reporting for Critical Infrastructure Act was passed in 2022, obliging organisations in important sectors to report major cyber-attacks within 72 hours. Its rules are due to be finalised in September, more than four years on.
There may now be momentum to do more. This month DEF CON Franklin, a group of civic-minded hackers, teamed up
with the NRWA to launch the Water Watch Centre, an initiative to provide private-sector cyber-security support to small utilities. Ms Klobuchar's and Mr Schiff's bill is not the only proposal under consideration.
On August 5th the AWWA, which resisted the Biden administration's proposals, endorsed the Water Risk and Resilience Organisation Establishment Act, sponsored by Rick Crawford, a Republican congressman from Arkansas. That bill would create an independent body to draft minimum cyber-security standards, under the EPA's oversight, in a mirror of the requirement for electric utilities. The same day Tom Cotton, a Republican senator, wrote to Scott Bessent, the treasury secretary, urging changes to the tax code and other regulatory tweaks to encourage water plants to invest in better security.
Whether Washington moves swiftly depends in no small part on Mr Trump. The president's proposed budget, which the
Senate will take up in September, would increase the budget for the Department of War by 44%, to $1.5trn. That includes money for the Iran war and a high-tech "Golden Dome", to shield America from missiles. He has shown less interest in defending American water.
His budget would cut the EPA's biggest source of funds for water cyber-security by almost 90%. In July he suggested that fault for Minnesota's attack lay with Tim Walz, the "corrupt" governor of the state—and that there had not been an Iranian attack at all. "Iran's got bigger problems than worrying about Minnesota," he offered. The Cybersecurity and Infrastructure Security Agency, the main federal body tasked with cyber-defence, is in disarray, with leadership turmoil, low morale and a one-third cut in staff since Mr Trump returned to the White House. The Senate may push him to do more after it returns from recess. Hackers are not waiting. ■
Political fundraising
The Republicans' money problem
WASHINGTON, DC
Donald Trump has a vast war-chest. He may not share it
BEFORE THE midterm elections Republicans claim to have at least one thing going for them: money, and lots of it. At first glance, the latest filings with the Federal Election Commission show Republicans with a staggeringly large cash advantage. Across their party apparatus and political action committees, or PACs, Republicans have amassed around $1bn, more
than twice as much as Democrats have (see chart on next page).
However, this should give Republicans scant encouragement. The party's biggest war-chest belongs to MAGA Inc, Donald Trump's super PAC, and the president has declined to say how much he will contribute to individual campaigns. While Republicans remain captive to the president's ►
United States 31
► whim, Democratic fundraising is decentralised. Individual candidates are outraising their opponents in competitive races. That means they have more cash on hand and more enthusiastic supporters, too.
Money cannot buy victory, but it is not immaterial, either. Candidates raise money to pay for campaign operations and to buy advertisements. Get-out-the-vote efforts can boost a candidate's prospects, though the effect seems to decline when a campaigner is already well known. In competitive general elections for federal office, it is usually hard to measure the effect of extra advertisements because airwaves are already saturated. The recent governors' primary in Wisconsin provided a rare natural experiment, albeit not in a general election. Francesca Hong, a Democratic Socialist, did not buy a single television ad. The Economist's analysis, controlling for factors such as voters' median age and education level by county, suggests that her lack of television advertising helped ensure her defeat in the Democratic primary on August 11th.
If the amount of cash matters, how campaigns raise it does, too: money is both a factor in a campaign's success and an indicator of its health. Strong candidates tend to attract donations from ordinary voters in their states, a sign of their enthusiasm. That is why The Economist includes individual donations as one predictor in our forecast of the midterm elections. It is also another reason why Republicans cannot be confident that their cash advantage will help them.
Mr Trump has proved particularly skilled at mobilising donors (think of those $1m-a-plate dinners at Mar-a-Lago). But those donations have flowed into his super PAC, not to candidates themselves. MAGA Inc accounts for at least $400m of the total Republican haul, a reminder that the president remains the party's source of power.
He looks reluctant to share it. MAGA Inc has spent just $2.3m, 0.6% of its resources, to help Republicans win elections so far. "I could spend it on pretty much anything I want," Mr Trump boasted recently. Though the president claims he will use his PAC to help Republicans, the group has yet to announce its spending plans, or even to share them with Republican allies.
While Republicans' fundraising is centralised—powerful in theory but limited in effect—Democrats are seeing the opposite phenomenon. The party's centralised fundraising is sputtering. Small donors are fed up with what they perceive to be the Democratic establishment. Top donors remain frustrated by how their money was spent during the 2024 presidential campaign; that the Democratic National Committee (DNC) is $2m in debt is not instilling confidence. The DNC, according to a quip in Washington, now stands for "Do
Not Contribute". But many individual Democratic candidates are not suffering—donors are still giving money, but skirting the central committee and going straight to the campaigns.
Take the Senate race in Texas. James Talarico, the Democratic nominee, raised $28m in the second quarter of this year, while his Republican opponent, Ken Paxton, raised a puny $1.6m. More than half of Mr Talarico's total came from small donations, compared with just 12% for Mr Paxton. The Democratic candidate is netting big donations, as well. Reid Hoffman, a major donor who has shunned the DNC, recently handed $10m to a pro-Talarico PAC.
Democratic candidates hold an advantage in other states, too. In the nine most competitive Senate races, they had around $116m in cash on hand, compared with just $51m for Republicans. In a reflection of voters' enthusiasm, on average small donors—individuals who pledge less than $200 to a campaign—accounted for 51% of donations to Democratic candidates, compared with 24% for Republicans.
Republican party officials insist that Democratic candidates' cash advantage is temporary, and that the money sitting in Republican coffers—including the president's—will soon flow to campaigns. With the midterms less than three months away, candidates would be excused for feeling a little bit restless. ■
Catch as cash can
Selected US super PACs and committees, cash on hand at June 30th 2026
| Republican-aligned $1.06bn | |
|---|---|
| MAGA Inc $401m | |
| Republican committees $277m | |
| Senate Leadership Fund $239m | |
| Congressional Leadership Fund $142m | |
| Spends on both $158m | Democrat-aligned $474m |
| Democratic committees $136m | |
| Senate Majority PAC $126m | |
| Democracy PAC II $122m | |
| House Majority PAC $90m |
Money raised by selected Senate campaigns, $m* April 1st–June 30th 2026
*Principal campaign committee only †Democratic candidate Graham Palmer withdrew on July 10th Sources: Federal Election Commission; National Journal
Political civility
You're doin' fine, Oklahoma!
OKLAHOMA CITY
A state wounded by extremism tries to guard against it
IT IS a coincidence that two of America's best-known bipartisan political organisations were, until recently, led by Oklahoma. David Holt, the Republican mayor of Oklahoma City, headed the US Conference of Mayors, while Kevin Stitt, the Republican governor of Oklahoma, led the National Governors Association (NGA). These groups offer something novel in American politics: civility. Last year 81% of Democrats and 70% of Republicans said the opposing party makes them angry. But at recent gatherings in Long Beach, California (the mayors), and Oklahoma City (the governors), Democrats and Republicans high-fived, hugged and waxed eloquent about the dangers of America's hyper-partisan politics.
This summer Messrs Holt and Stitt passed their gavel to new leaders. But it is worth considering why Oklahoman politicians seem especially interested in making America nice again. Both men have bucked their party and publicly disagreed with President Donald Trump. They cherish friendships with Democrats. Running cities and states demands a pragmatism that is not a prerequisite for a job in Congress. But more than that, many of Oklahoma's political leaders came of age in the 1990s, when they witnessed their state respond to extremism with compassion.
In 1995 Timothy McVeigh set off a bomb that blew up the federal building in downtown Oklahoma City. Nearly 170 people were killed. It remains the deadliest act of domestic terrorism in American history. McVeigh wanted to retaliate against what he believed to be a tyrannical government. Mr Holt was in high school at the time, getting ready to deliver a campaign speech for student council vice-president. "There was this rumour backstage there's been an explosion downtown," he recalls. "We got back to class, and they were rolling the TVs in."
After the bombing, the outpouring of support for victims, survivors and first responders became known as the "Oklahoma Standard". "That was the phrase used to capture how our community embraced all the people who came to help," says Mr Holt. It has become "a catch-all for being people who try to rise above darkness".
Mr Holt says the bombing shaped his politics. "Part of the ethos of the city now is trying to do our part to push back against polarisation and dehumanisation," ►
32 United States
► he says, “because we have a scar in our downtown that reminds us what that inevitably leads to.” He has been a Republican since campaigning for George H. W. Bush in his fourth-grade classroom. But as mayor he has tried to foster inclusion, proclaiming Oklahoma City’s first LGBTQ Pride Week and meeting protesters for racial justice in 2020. Last year, at the memorial for the bombing, 230 mayors pledged to “reject political violence and recommit to the American experiment”.
Mr Stitt, who was a senior in college in 1995, attributes his niceties to “Oklahoma common sense”. “My grandmother said ‘You get more flies with honey than you do vinegar,’” he says, “so I’m trying to just be nice to everyone.”
Comity in the age of Trump
Messrs Holt and Stitt admit that it will be hard to overcome the incentives that politicians have to be nasty. Mr Trump is America’s bully-in-chief—and Americans elected him president twice. But the Oklahomans aim to set a better example. They do not always succeed. At the NGA conference earlier this month Mr Stitt took a jab at California, a Democratic stronghold, and instantly expressed regret. “Man, I shouldn’t have said that, this is a bipartisan deal,” he said on stage.
They also offer more tangible reforms. Mr Holt thinks that doing away with closed party primaries, in which only party members can vote, would make it harder for extreme candidates to win nominations. (Research on whether open primaries produce more moderate candidates is mixed.) Mr Stitt—ever the conservative—prescribes more federalism and a smaller central government. “It’s gotten so big that Oklahomans, we’re worried that if Kamala Harris gets elected, she’s going to make Oklahoma like California,” he says.
Another prescription is the Oklahoma Standard itself. Other governors at the NGA conference expressed admiration for the state’s civic ethos. “What’s interesting about the Oklahoma Standard is it used to be the United States standard,” laments Spencer Cox, the Republican governor of Utah, who has long urged Americans to “disagree better”.
Outside the memorial in Oklahoma City there is a fence where mourners place flowers and photos of loved ones. Every year Kyle Genzer leaves a letter for his mother, who died in the blast. “This year brought one of the most beautiful moments—the birth of your first great-granddaughter,” he writes. “Holding her, we couldn’t help but think of you.” Mr Genzer, a high-school principal, thinks it’s important to keep talking about the bombing, both to honour the victims and to remind the living that disagreement need not harden into hatred. ■
Sport
Scoring political points
ATLANTA
Why conservatives can’t stop talking about women’s basketball
THE ORGANISERS of the Women’s National Basketball Association (WNBA) once begged journalists to cover it. The joke was that all you needed to attend a press conference was a camera and a question. Now it is standing room only. Much of that reflects the league’s surging popularity. But it has also become an unlikely battleground in America’s culture wars.
Lately the focus has been on Sophie Cunningham, a guard with the Indiana Fever whom some have dubbed “MAGA Barbie”. Ms Cunningham has not publicly identified as a Republican or a Democrat. But in an interview published last month she said transgender women should be barred from competing in women’s sports. Most Americans agree with her. Ahead of the midterms, though, the issue has become a fixation of right-wing politicians and commentators.
For proof look no further than Fox News, which generally pays little attention to women’s sports. Between June 25th and August 11th it devoted more than 200 segments to the WNBA; at least 130 of them on trans athletes, according to Media Matters, a left-leaning watchdog. The coverage has been so relentless that one of the network’s biggest personalities, Jesse Watters, complained on air: “This is ridiculous.”
When, on August 8th, Ms Cunningham was aggressively fouled by an opposing player, DiJonai Carrington, some conser-
vatives cast it as retaliation for her views. Florida’s attorney-general, a Republican, implied that had the game been played in his state, he might have charged Ms Carrington with assault and battery.
The foul (pictured) brought another element of this saga into sharper focus: race. Ms Cunningham is white; after the game Ms Carrington, who is black, posted on social media “WHITE PRIVILEGE @indianafever”. She has not explained what she meant. But Caitlin Clark, another white Fever player and the league’s biggest star, has also become a MAGA cause célèbre. Last month 11 Republican lawmakers wrote to the WNBA complaining that Ms Clark was being pushed around on the court. They cited reports that this “may be racially motivated”. Donald Trump agreed that Ms Clark had been “treated rather rough”.
The league, for its part, has not always handled things well, at times asking fans to cover up shirts about trans athletes before reversing course. But it has also pointed out the absurdity of some of the coverage. There are no trans athletes in the WNBA. Two former NBA players and supporters of Mr Trump claim to want to join; the league calls their efforts a “publicity stunt”.
Why are Republicans pushing these buttons now? As the midterms approach, voters are focused on affordability, a problem for the party in charge. Race and trans issues offer more promising terrain on which to portray Democrats as out of touch. In June conservatives won a victory at the Supreme Court, which ruled that states could bar transgender women from competing in women’s sports.
Meanwhile, the WNBA’s attendance and viewership are booming. “Our focus is on basketball,” said Ms Clark recently. She is on a historic run, breaking records. That seems worthy of attention, too. ■
A little too on the nose
United States 33
Obamacare
In poor health
NEW YORK
As the left calls for Medicare-for-all, Democrats' signature health law is faltering
"HEALTH-CARE reform...is the great unfinished business of our society," observed Ted Kennedy, a Democratic senator, before he died in 2009. America was the only large rich country that did not have universal health care. Instead, its system was a patchwork. The old, hard-up and disabled were covered by government programmes. Most working-age Americans got insurance through their jobs. But pity those without such coverage, who had to navigate the expensive and often bewildering individual insurance market.
The Affordable Care Act, passed in 2010, tried to fix that. It kept America's system more or less intact, but expanded Medicaid, the public programme for the impecunious, and created new "exchanges" where individuals could compare and purchase plans. Even Obamacare's most fervent advocates acknowledged, in private, that it was an imperfect bandage, particularly for those on the individual market. It was never enough for the likes of Bernie Sanders, a leftist senator who called for a single-payer system. That cause is now being taken up by a new generation of Democrats, such as Abdul El-Sayed, the party's Senate nominee in Michigan, who has made health-care-for-all central to his campaign. New data explain the urgency: Obamacare's bandage is unravelling.
That's in part due to cuts to Medicaid. It's also because the exchanges are in bad health. Obamacare offered some subsidies
to defray the cost of individual plans. But the law's mandate to buy insurance was in effect eliminated in 2019, and uptake on the exchanges was limited until 2021, when Congress made subsidies more generous for those who had already qualified and available to some who hadn't. Enrolment then nearly doubled by 2025. Democrats sought unsuccessfully to extend the extra financial assistance, which would have cost about $35bn this year—about as much as the direct costs of the Iran war, according to the Pentagon's estimate. Now that those subsidies are diminished, the individual market looks more frail than ever.
Figures from insurers and state marketplaces show the effect of the rollback. The
number of people buying coverage through the exchanges has already fallen to 19.2m from 21.8m last year. The Congressional Budget Office (CBO), a non-partisan scorekeeper, reckons that by the end of this year it is likely to fall below 17m. Of those who dropped off the exchanges, most appear to be going without coverage.
The leavers tend to be more healthy than those who remain on the exchanges, making the remaining pool more costly to insure. Combine a sicker patient population and general inflation in health costs, and the result is higher premiums: this year the average payment increased from $113 to $178 per month. Many enrollees are now buying skimpier plans. In the past year the average amount an individual must pay out of pocket before insurance kicks in has jumped by nearly 40%, to $3,800.
This has raised fears of a dismal loop, with higher premiums and more limited plans making insurance still less attractive, particularly to healthier customers. As they leave the exchanges, premiums are driven yet higher. Insurers have already proposed a median premium increase of 15% for 2027, according to KFF, a think-tank. If state regulators approve those rates, pre-subsidy premiums will have grown by more than a third since 2025. The KFF imagines a 40-year-old in Indiana earning $65,000. In 2025, with subsidies, her monthly premium was $316. This year it is $477. Next year it will probably be close to $550.
Does this threaten the viability of the exchanges? The CBO estimates that by 2028, 10m fewer Americans will get insurance on the marketplaces compared with last year. But the exchanges will probably avoid collapse, with the remaining subsidies convincing enough people to remain. "The changes basically get us back to what the marketplace looked like from 2016 to 2019," predicts Ben Sommers, a health economist at Harvard University. The exchanges will be smaller, with fewer participating insurers and sicker customers.
There will also be more people without insurance. Evidence suggests that they will put off needed treatment, like certain prescription medicines and surgeries. "We know that from a decade or more of research on the Affordable Care Act," explains Mr Sommers. Hospitals must still provide emergency care for them, even if they cannot pay. In May unpaid care was up by an average of 16% compared with a year earlier, according to Kaufman Hall, a consultancy. HCA Healthcare, a for-profit hospital group, reckons the new uninsured population will cost the chain at least $1bn in operating profits this year, equivalent to about 15% of predicted net income. Enrolment on the exchanges in recent years shows "people generally want health insurance," says Cynthia Cox of KFF. "It's just a question of whether they can afford it." ■
Vital statistics
United States, Affordable Care Act marketplace enrolment*, m
*Selected a plan and paid initial premium †At May 15th Source: KFF
34 United States
The doc in the dock
The real covid-19 scandal is America's continuing failure to reckon with lives needlessly lost
FOR ALL the zeal Congress has shown in investigating America's response to the covid-19 pandemic, it has been strangely uninterested in exploring a basic question: why did so many Americans die relative to the numbers who succumbed to the virus in other rich countries? Understanding the answer would seem relevant to guarding against the next, inevitable pandemic.
Many countries were slow to react, but as they learned about the transmissibility of the virus, and as vaccines became available, they rapidly cut their death rates. Not so America. "The death rate is now about eight times higher in America than in the rest of the rich world," The Economist reported in September 2021. In all, America lost more than 1.1m people to covid, or roughly 3,500 lives out of every million. This should haunt every official who was in a position to combat the pandemic. Germany's death rate was about 2,000 per million people and Canada's fewer than 1,400, notes Donald McNeil in his book "The Wisdom of Plagues". Had America managed to split that difference, at about 1,700 per million, "We would have lost about 540,000 fewer dead," he writes. "Who speaks for them?"
Not Congress. America's lethal underperformance once again went ignored when the Senate Homeland Security Committee recently questioned Dr Anthony Fauci, a crucial adviser on covid to Donald Trump and then to Joe Biden. (Mr Trump later turned on him.) The hearing focused instead on the usual obsessions for which Dr Fauci has become the Republican scapegoat—some absurd and others valid but distantly related, at best, to protecting potential victims of a future plague. While Democrats mostly gave speeches rather than ask questions, Republicans pursued theories that the virus leaked from a Chinese lab, that Bill Gates had some nefarious role, that Dr Fauci had been out to enrich himself through awards for his service—an odd complaint when they do not object to the president campaigning for a $1m Nobel prize.
Having testified often to Congress during his 54 years at the National Institutes of Health—38 as director of the National Institute of Allergy and Infectious Diseases—Dr Fauci, who is 85, was reduced to repeatedly invoking the Fifth Amendment, including when asked the colour of his tie. In his opening statement, his
hands shaking as he read it, Dr Fauci said his lawyers had advised that recourse because the committee's chairman, Rand Paul of Kentucky, had made clear an intent to see him behind bars despite a blanket pardon issued by Mr Biden. In other words the Senate could not inquire into a health crisis that racked the country because a key witness feared a perjury trap, and rightly so.
Why did so many Americans die? High rates of obesity, diabetes, heart disease and elderliness made them more vulnerable. But government bungling meant America was testing 100 people a day in early 2020 when other countries were testing tens of thousands; that masks and lockdowns were deployed erratically; that contact tracing proved too challenging for the country first to the Moon. Most senselessly, so many Americans died because so many failed to take the vaccines that Mr Trump, in his most constructive act as president, succeeded in rushing into existence.
It is, of course, those vexing vaccines that compel Republicans to look away from the basic question about the pandemic, and to hound Dr Fauci instead. Attempting to square the greatness of Mr Trump with the supposed vileness of the vaccines has inspired his party to heroic feats of doublethink. "President-elect Trump's Operation Warp Speed—which encouraged the rapid development and authorisation of the COVID-19 vaccine—was highly successful and helped save millions of lives," declared the committee created by the House to investigate covid in 2024, summarising two years of work. The very next sentence reads: "Contrary to what was promised, the COVID-19 vaccine did not stop the spread or transmission of the virus."
Mr Trump has struggled to escape this self-imposed contradiction. Even as he declined to urge the public to get vaccinated, he and his wife, Melania, secretly had the vaccine before he left office in January 2021. But as Mr Biden and other world leaders delivered the shots, Mr Trump wanted the credit he deserved, and in December that year he told supporters that he had received a booster. "What we've done is historic," he insisted when the crowd booed. "Don't let them take it away." He has since let them take his glory away, and endangered more Americans, by questioning the efficacy and safety of vaccines for covid, and other illnesses.
Dr Hooey
The latest attack on Dr Fauci, concocted from his private texts disclosed by Republican senators on August 10th, is that he covered up a threat of miscarriage that the vaccine posed to pregnant women. In fact, the record shows Dr Fauci was consistent, publicly and privately, in counselling a weighing of risks and benefits until studies showed the vaccine was safe in pregnancy—which they have since done. His critics have not produced evidence the vaccine causes miscarriages.
Dr Fauci made mistakes, including not disclosing early on that some top virologists initially worried the virus might be man-made. He could be arrogant. But he was not responsible for the worst excesses of the covid response. Those endless school shutdowns were the work of local officials and union leaders. Senator Paul's release of Dr Fauci's private diary has exposed his vanity, his delight in every gushing appraisal by the media or some celebrity. But the diary also reveals a conscientious public servant trying to save as many lives as he could, struggling against the constraints that hobbled America relative to its peers. His summons before the Senate committee was surely intended to humiliate if not incriminate him. It should embarrass the entire country. ■
35
The Americas
Brazilian institutions
Imbalance
RIO DE JANEIRO
Brazil's Congress is too powerful, and causing problems
IN THE PAST decade, democracies around the world have been bent to the will of would-be autocrats. From the United States to Turkey and India, strongmen have accumulated power in the executive branch of government, and weakened the institutions that could check them, legislatures in particular.
Frustrated lawmakers elsewhere, therefore, might envy their peers in Brazil, where this trend is running in reverse: in the past decade Brazil's executive branch has lost clout to the courts, and especially to Congress. The balance of power has become so distorted that Brazilian lawmakers no longer seem to care who is elected president in the general election in October. Some reckon that whoever wins may be forced to capitulate to their demands. Brazilians may be in the process of discovering that, just as an excess of executive power is a challenge to democracy, so an overmighty Congress can be, too.
The contest on October 4th will pit two radically different presidential candidates against each other: the incumbent, Luiz Inácio Lula da Silva, known as Lula, is an 80-year-old left-winger. His rival, Flávio Bolsonaro, is the 44-year-old son of a populist former president, Jair Bolsonaro, who is serving a 27-year sentence for plotting a coup after losing an election in 2022. All 513 seats in the lower house will also be renewed, as well as two-thirds of the Senate. Yet despite the polarised race, Brazil's largest bloc, the Centrão (big centre), which accounts for 46% of seats in the outgoing Congress, has declared its neutrality. Party bigwigs will allow local politicians to support whichever candidate they choose.
→ ALSO IN THIS SECTION
36 Argentina: polo superpower
37 Paraguay's other language
In the past decade, as a succession of weak presidents ruled Brazil, Congress hijacked power from the executive. In 2016, amid a gigantic corruption probe which ensnared dozens of legislators, lawmakers impeached Dilma Rousseff, the left-wing president at the time, in part because they felt she was not doing enough to protect them from investigators. Ms Rousseff's successor, Michel Temer, governed with little public support. Mr Bolsonaro, who ruled between 2019 and 2022, gave lawmakers ever-greater control of the federal budget in order to shield himself from impeachment attempts after his disastrous handling of the covid-19 pandemic. Today, Brazilian legislators control 25% of discretionary spending in the federal budget, up from 2% in 2015 (see chart on next page).
How to spend it
Such power over the budget is highly unusual. Marcos Mendes of Insper, a university in São Paulo, and Hélio Tollini, a budget adviser to the lower house, have compared congressional amendments in 11 countries of the OECD, a club of mostly rich countries. In some, such as Canada and Australia, lawmakers have no right to amend the budget or apportion funds for themselves. In others, such as the United States, France, Italy and Spain, only around 1% of discretionary spending in the
36 The Americas
▶ budget is allocated to congressional amendments. In most of these states, lawmakers propose amendments to the budget, which are not always accepted.
Not so in Brazil. There the budget must now always include a sum for congressional amendments. Congressmen often do not stipulate what they will use the money for. They have indexed the amount that the government must allocate to amendments to revenue growth, so that their grants inexorably grow. And they have all but scrapped rules around transparency and traceability in the use of funds.
The proportion of money that Congress controls is not huge. Over 90% of Brazil's budget of 6.5trn reais ($1.2trn) is spent on mandatory expenses, such as public salaries, pensions and financing the debt. Even as congressional control over the discretionary slice expands, the share of the budget that goes to mandatory spending is growing rapidly. In 2026 only 3.7% of the budget, or $46bn, was available. Of this, Congress controls $12bn.
Easy access to money has boosted incumbents' chances of re-election, as it allows lawmakers to fund glitzy projects that their constituents like. Their political allies in local government benefit, too. In 2024, when Brazilians elected over 5,500 mayors and tens of thousands of city councillors, a record 81% of incumbents who chose to run again held onto their jobs. In the upcoming general election, 87% of lower house members are standing for re-election, the highest share in Brazilian history.
In Brazil's fragmented party system, where presidents rarely command a congressional majority, control of the budget had long been a powerful bargaining chip for the executive. Presidents could shore up support in Congress by threatening to withhold funds for lawmakers' projects. Now that leverage is much weaker. In April the Senate rejected Lula's nominee to the Supreme Court, Jorge Messias. He was the first such presidential nominee to be turned down in more than a century. Senators wanted the seat to be filled by someone from within their own ranks.
Can't touch this
Congress is using its new powers to write more laws. Between 1988 and 2004, more than 80% of new laws in Brazil originated in the executive. By 2025 the share was 19%, while Congress accounted for 75%. This is not in itself bad, but Beatriz Rey of the University of Lisbon notes that Congress has become more susceptible to lobbies. Last year it rushed through a bill that hollowed out environmental licensing, under pressure from agribusiness groups. When Lula vetoed large parts of the bill, Congress overturned most of his vetoes.
Outright corruption is perhaps the biggest risk. In July the Federal Accounts
Holding the purse strings
Brazil, federal discretionary spending controlled through congressional budget amendments, %
*To August 15th
Source: "Is it like this everywhere? Parliamentary amendments in Brazil and 11 OECD countries", by H. Tolini and M. Mendes, INSPE, 2024 (data updated to 2026)
Court released a sample of 100 so-called "Pix amendments", where relatively small amounts of money are sent directly by a legislator to his local municipality. The grants in the sample totalled around $40m and had been sent to 75 recipients between 2020 and 2024. Irregularities were found in 82% of transfers, including money that moved through bank accounts that had nothing to do with the stipulated project for the grant, contracts with wildly inflated budgets, or funds destined for projects that were never completed. The accounts
court estimated the loss to the public purse at $10m. The sample was small, but if it is representative, the losses are huge. Congress has approved 7bn reais of Pix amendments in the current budget, meaning some $340m may have been lost.
In 2024 the Supreme Court forced legislators to identify the bank account that would receive federal funds, and to provide details on the projects the grants would be destined for. Yet legislators appear ready to put up a fight. Lawmakers have filed eight impeachment motions against Flávio Divo, the justice in charge of investigating legislators who make suspicious budgetary grants. (In Brazil the Supreme Court is in charge of criminal investigations into sitting politicians.)
Indeed, lawmakers are gleefully throwing their weight around. Since 2021, senators have filed 105 impeachment petitions against Supreme Court judges (who have themselves become overmighty in recent years). The number of petitions may rise as new corruption investigations continue to ensnare legislators, including probes into fraud at the public-pension fund and the fallout from the biggest bank fraud in Brazil's history. Brazil's next president faces not only sluggish economic growth, high interest rates and a polarised electorate, but an overmighty Congress that will try to keep him weak and compliant. ■
Horse power
Chukkanomics
BUENOS AIRES
Argentina's dominant polo economy points the way for the rest of the country
"TWENTY FIVE years ago I came here for the first time," says Maximiliano Fernández. "There was nothing." Now the area between the towns of Pilar and General Rodríguez, a strip of land 50km west of Buenos Aires, is packed with polo clubs, stables and estates. Every scrap of farmland is being bought up and transformed into pristine polo fields. Argentina's polo world is "a bubble" within the country, says Mr Fernández, who manages a team: it has its own economy. And that economy is world-beating.
Polo—hockey on horseback—is purportedly the world's oldest team sport. An early version was played in Persia in the sixth century BC. The modern game was formalised by British officers in India and then galloped around the empire. It became a sport of kings, taken up by elites in Europe, the United States and more recently the Gulf states. Yet Argentina came to dominate the global scene, even as the
rest of its economy was in a shambles.
Like any fledgling industry, it helps to have the raw materials. Horses had long thrived on the Pampas, central Argentina's vast grasslands. They attracted gauchos, the local version of cowboys. Polo, when it arrived, fit naturally with the abundance of cheap horses, skilled riders, flat land and agreeable weather. Argentine players were soon reckoned the world's best, and remain so: 27 of the top 30 professionals today are from the country. They sustain a service-based economy in which rich foreigners hire them to help win trophies.
Industry soon followed. Landed Argentine families have built formidable breeding programmes which pump out the world's finest polo horses (or ponies, as they are known). And these products lead the field: of the 1,140 horses that competed for the Queen's Cup in Britain last year, 1,000 were of Argentine bloodlines, according to the country's polo association. ▶
The Americas 37
By maximising its competitive advantage in human—and equine—capital, Argentina's polo economy has enjoyed an export-led boom. Between January and November last year Argentina shipped 2,900 horses (three-quarters of them polo ponies) to 38 countries. The country is the world's leading exporter.
To keep their noses ahead, Argentine breeders and businessmen have invested in R&D and pioneered the cloning of ponies. A breakthrough came in 2010, when a clone of Dolfina Cuareterra, ranked the best pony in Argentina's Hall of Fame, sold for a record $800,000. Now many are copies of past Argentine greats. Some breeders are attempting to trot a step further in order to protect their intellectual property, selling only the foals of clones, rather than the clones themselves.
Less tape more ribbons
Regulators have enabled the innovation: the polo association has resisted red tape and embraced technology even as thoroughbred horse-racing has prohibited it. But there is a limit. When Kheiron, an Argentine biotech company, unveiled the world's first gene-edited horses in 2024, they were barred from competition. After all, such tech could one day enable rivals abroad to engineer horses that outrun even Argentina's famed bloodlines.
At the same time as Argentina's economy bucks and rears (shrinking by 2.5% in February, growing by 2.8% in March then contracting by 1.5% in April), the polo world seems unbothered. That is thanks, in part, to a steady flow of foreign direct investment. Rich foreigners, known as patrons, bring money aplenty into Argentina's polo bubble. They buy land, hire players and haggle over horses, creating a trickle-down economy for grooms, farriers, trainers, vets and architects.
Sustained growth has been the result. Two decades ago Juan Martín Batistuta, an equine vet, worked almost entirely on racehorses; today he works almost entirely on ponies. "I've never been busier," he says. Nor is the boom confined to Buenos Aires. "From Córdoba to Patagonia, all over Argentina you have so many more clubs," says Mr Fernández. The country now has 177. The sport's profile is rising globally, too. A Netflix documentary released in 2024 and a broadcasting deal with ESPN have brought it to new audiences.
The approval ratings of Javier Milei, Argentina's president and a polo fan, are suffering as he attempts to reset the country's economy. He might glance at proceedings in Pilar: a world-beating industry making the most of its competitive advantage, attracting foreign investment and adopting new technologies to bolster its position as the world's leading exporter. An ideal remedy, perhaps, for a long face. ■
Indigenous language
Cool talk
ASUNCIÓN
Guaraní has become trendy in Paraguay
ONE HOUR into their Guaraní class, the medical students at the National University of Asunción are enthralled. One group provides "pytyvöpya'ehasývape" (first aid) for a "kangue opé" (broken bone). Another frets that a problem with the "tajygue mboaty rehe" (nervous system) may require "hembovo" (surgery). The course is mandatory for Paraguay's future doctors, part of a wave of official measures to promote the indigenous language. Few seem to mind. "I love speaking Guaraní—it's much friendlier than Spanish," says Tadeo Ramírez, an undergraduate.
Such zest reflects the recent revival of Paraguay's ancient tongue. Just 2.3% of the population identifies as indigenous, but some 90% understand Guaraní. Nearly 40% speak it every day at home along with Spanish, according to a recent survey. That is one of the highest rates of bilingualism in the world, far ahead of official diglots like Canada and Ireland. Most of this is "Yopará", a blend of Spanish and Guaraní (the word means "mixture" in Guaraní). "It's our lingua franca," says Miguel Angel, a taxi driver. Paraguay is one of only two countries in the Spanish-speaking Americas with an official second language.
Paraguay's elites had long snubbed Guaraní as a rural vernacular, mocking its nasal sounds and glottal stops. Dictator Alfredo Stroessner discouraged its use in schools in the 1950s and 1960s. Only with
the return of democracy in 1992 was Guaraní recognised as Paraguay's official language alongside Spanish. A law in 2010 created a bilingual curriculum and administration. Today Paraguayans can choose to vote, marry and stand trial in Guaraní.
This has facilitated a cultural craze. In June Paraguayan football fans attributed their World Cup win over Germany to "Guaraní spirit". In May the national symphony orchestra performed Beethoven's "Ode to Joy" in Guaraní for the first time. In April Spain's Queen Letizia opened her message of congratulations for the anniversary of the Spanish cultural centre in Asunción with "¡Vý apavé ne arambotýre!" ("Happy Birthday!"). Youngsters are joining in. Influencers fill TikTok and Instagram with vocab tutorials; artists pump Yopará rap and jazz onto Spotify. Trendy parents give their children names like Yeruti (dove) and Marangatu (virtuous).
Some worry that growing bilingualism comes from Guaraní speakers learning Spanish rather than vice versa. The share of monolingual Guaraní speakers dropped from 28% in 2002 to 12% in 2022. "It's a process of displacement," says Arnaldo Casco of the Secretariat of Linguistic Policy (SPL), a ministry. It has responded by publishing reams of Guaraní content, from technical dictionaries to poetry. Guaraní must cover all spheres of social interaction, says Javier Viveros, the SPL's boss.
Others are laid back. Spanish monolingualism is flat. Guaraní is on Google Translate and Wikipedia ("Vikipetá"). SPL staff are training large language models on it. Social norms have shifted. "We who have gone through that experience of being hesitant to speak Guaraní as children today feel absolutely free," says María del Carmen Giménez of the Patria Soñada Institute, a think-tank. That speaks volumes. ■
"The people are on the move"
38
Middle East & Africa
The West Bank
Emboldened, and violent
QUSRA
With government support, Israeli settlers are escalating attacks on Palestinians
THE HILLS south of Nablus in the Israeli-occupied West Bank tell the story of a creeping annexation. Winding dirt tracks to Palestinian villages are blocked by rocks and piles of earth, cutting off access from fields and olive groves. Israeli settlers have planted vines in neat rows, erected fences festooned with flags and brought in herds of goats. They patrol the area with new utility vehicles and assault rifles given to them by the Israeli government.
Their stated aim is to cause “friction” with adjacent Palestinian villages. That tends to lead to intervention by the Israel Defence Forces (IDF), often on the settlers’ side, followed by the expulsion of Palestinians and the expansion of outposts of Israeli settlements (see map on next page). Since the Hamas attack of October 7th 2023, and Israel’s subsequent war in Gaza, violent attacks by Israeli settlers in the West Bank have worsened and forced an increasing number of Palestinians to flee
(see chart on next page). As Israel prepares to go to the polls on October 27th, the failure of Binyamin Netanyahu, the prime minister, to rein in the settlers is making counter-violence by Palestinian militants increasingly likely.
Events over the past week show how much leeway settlers are granted. Tel Talpiot, the latest and northernmost outpost in the Shiloh settlement bloc, was set up earlier this year on a hill overlooking three Palestinian villages. On August 9th a group from Tel Talpiot occupied the courtyards of homes on the outskirts of the village of Qusra, setting up a makeshift shelter, ter-
rorising the families and forcing them to hole up inside. IDF soldiers sent to “prevent friction” joined the settlers in prayer, rather than dislodging them.
The tactic, common in the West Bank, often succeeds in forcing Palestinian villagers to flee their homes, allowing the settlers to take over the surrounding land. This time was different for several reasons. Residents of Qusra, who had experienced similar attacks before, were quick to publicise this one and attract international attention. One of the besieged homes belongs to an American citizen. Mike Huckabee, the American ambassador, who usually supports the settlers, posted on social media that it was his embassy’s “request to remove the Israeli terrorists”.
Three days into the siege, the IDF belatedly replaced the unit in Qusra with a more disciplined infantry battalion. It tried to remove the settlers and evict the Tel Talpiot outpost. The settlers brought in reinforcements and remained in the area. “It was an embarrassing incident,” said one officer. “The settlers forced us to play cat-and-mouse, and we lost.”
The IDF announced the homes were in a “closed military zone”. It forced the families to remain blockaded indoors and prevented activists from coming to their aid and media from visiting. “We are besieged, the army is not allowing us to leave our
→ ALSO IN THIS SECTION
39 Consolidating power in Iran
40 South Africa’s broke municipalities
41 Somaliland’s president has big plans
Middle East & Africa 39
house,” said Qusai abu Ridi, one of the residents. But he was adamant they would continue to protect their property.
That makes Qusra, despite everything, a rare case of Palestinian residents managing to hold on to their homes. Yet the incident is also part of a fresh escalation over the past month, as settlers have tried to grab as much land as they can before the election, which might return a government less supportive of their cause. On July 24th a group of armed settlers arrived in the village of Tel on what they claimed was a “hike.” Four Palestinians and two Israelis ended up killed. In response, the IDF deployed more troops and locked down Palestinian towns, including Nablus, one of the largest in the West Bank.
The IDF claims to be enforcing order, but it has overseen a steep rise in attacks on Palestinians. The Shin Bet, Israel’s security agency, says 660 “nationalistic” attacks took place in the West Bank in the first half of 2026, up from 405 in the same period last year. More than 1,100 Palestinians have been killed in clashes with the IDF and settlers since October 7th 2023.
The events in Qusra highlight the IDF’s disinclination to prevent settlers’ attempts to chase out Palestinians. Senior officers, both serving and retired, blame a breakdown in discipline and the presence of settlers within the IDF’s ranks. But ultimately they hold the government responsible.
While the IDF’s high command has publicly condemned the attacks in Qusra, Binyamin Netanyahu, the prime minister, has remained silent. On August 14th Israel Katz, the defence minister, said that enforcing the law on settlers in the West Bank would, from now on, be the responsibility of the Israeli police and not the IDF. Since the police lack personnel in the West Bank and are under the control of the far-right national-security minister, Itamar Ben-Gvir, himself a settler, they are unlikely to enforce the law in the area.
Israel deploys vast forces and intelligence resources in the West Bank to protect half a million settlers. In theory, the IDF is obliged to protect Palestinians, too. But soldiers will only do this if their politi-
cal masters expressly tell them to. That is unlikely to happen under Mr Netanyahu, who needs the support of settler-friendly far-right parties to have a chance of retaining power after the election.
Things in the West Bank could get worse before then. Security officials warn that settlers are trying to provoke another intifada, a Palestinian uprising, which will force the IDF to enter Palestinian cities. That would bring even more violence and misery to the West Bank’s embattled Palestinians. Even if Mr Netanyahu’s party were voted out, his successors may have to deal with a crisis of his making. ■
Iran and the war
Rise of the generals
A military junta is consolidating power, with dire results at home and abroad
OVER THE past few days new billboards have appeared in Tehran, Iran’s capital. They depict the seven men who have taken charge in Iran since Ali Khamenei, the former supreme leader, was killed in an Israeli air strike on February 28th. Most remarkable is what the billboards do not show. There is no sign of Mojtaba Khamenei, the new supreme leader, who has not been seen in public since he was appointed. Nor are there any of the traditional religious symbols, such as the black and red flags of Imams Ali and Hussein. Only one of the seven men wears a turban.
The billboards are the latest sign that a new crop of leaders from the Islamic Revolutionary Guard Corps (IRGC) is consolidating power in Iran. Battle-hardened and results-driven, they are focused on security and repression, and are turning the theocratic foundations of the Islamic republic into a military dictatorship. Though they may not have entirely dropped a pragmatic
approach to the conflict with America, military regimes often draw legitimacy from the constant threat of war. That suggests America and its allies should expect the new leadership to be more uncompromising than the previous one.
Fronting the self-declared “top of the country’s security and defence” on the billboard and dressed in a suit is Mohsen Rezaei. He is the new national-security adviser, a former presidential candidate and the longest-serving chief of the IRGC, which he led through the Iran-Iraq war of 1980-88 and the reconstruction that followed. Five men wear military fatigues, including the head of the navy who oversees Iran’s attacks on Gulf shipping. Next to him is Ahmad Vahidi, the IRGC’s commander-in-chief. Besides being instrumental in attempts to suppress the women’s movement to cast off the veil four years ago, he founded the Quds Force, the IRGC’s foreign arm, which nurtures its network of regional proxies. Hussein Taeb, the only man with a turban, used to run the IRGC’s intelligence agency and now heads the husij, its paramilitary volunteer force entrusted with keeping protesters off the streets.
Some veteran Iran-watchers see the men’s takeover as the triumph of hardliners, who seek the expulsion of America from the region, over pragmatists led by the speaker of parliament, Mohammad Bagher Ghalibaf. He was instrumental in negotiating a memorandum of understanding (MOU) with America that was signed on June 17th. The MOU promised Iran an end to nearly five decades of American opposition, protection for its regional proxies, sanctions relief, the repatriation of frozen oil revenues and $300bn to kick-start post-war reconstruction, in return for Iran eschewing its pursuit of a nuclear bomb and ending its blockade of the Strait of Hormuz.
Fighting resumed within days of the document’s signing. The Strait of Hormuz remains largely closed to shipping. On August 17th the deadline the MOU set for a more permanent peace deal expired; a day later Iran appeared to lob missiles at the United Arab Emirates, an American ally in the Gulf, which responded with a trade embargo. Some observers worry that the hardliners’ sense of having parried American and Israeli attacks has revived their conviction in the viability of force. “They’re squandering the victory [with the MOU],” fumes an academic in Tehran.
Domestically, too, the new regime looks harsh. The morality police are back in force in Tehran, raiding and closing cafes perceived as too lax on enforcing the veil. Public executions of anti-regime protesters have resumed. This month a photographer who captured Iran’s social malaise was gunned down with his mother and sister near a city in the north-west. ▶
Driven out
West Bank, Palestinians displaced by settler violence and access restrictions
40 Middle East & Africa
Yet other insiders and observers see more calculated realpolitik. The generals approved Mr Ghalibaf's MOU. But without pressure, they are said to have reasoned, it stood no chance of being implemented. Despite the MOU, America set up its own blockade over the strait and negotiated the disarmament of Hizbullah, an important proxy, in concert with Israel. Iran could not even secure the transfer of frozen revenues that had been part of the deal.
The new regime could be keen on an agreement that delivers finance. After the Iran-Iraq war, Mr Rezaei spearheaded Iran's reconstruction. As a presidential candidate, he called for foreign investment and a nuclear deal. "They're saying isolation will kill us," says another Iranian academic with ties to the new leaders. Nor, say some, are they fazed by diminishing Islamic norms. The crackdown on cafés may be merely a sop to their base. Some liken them to Muhammad bin Salman, the Saudi crown prince who is pursuing social liberalisation in place of political reform.
Yet all observers agree that there is a new style of leadership. For almost four decades Mr Khamenei balanced the competing interests of his realm, alternating reformists and hardliners. The new commanders, by contrast, favour action. Iranians and the world may yet rue the day Mr Khamenei left the scene. ■
Small-town South Africa
State capture in miniature
MPOFANA
Failing municipalities stymie growth, but some show change is possible
"YO, THERE'S A lot that's bad in this community," sighs Neliswa Radebe. The 19-year-old is one of 13 people living in a small ramshackle house above a motorway in Mpofana, a municipality in eastern South Africa. She lists the problems: piped water is rare; rubbish collection is sporadic; and there is a leak dripping on the mattress used by her bed-bound adopted mother. Outside is no better. The roads are more pothole than tarmac; a resident died recently when a car careered into them. There are no playgrounds, football fields, netball courts or youth centres. Squatters have overrun Mpofana's dilapidated town hall, on the other side of the motorway.
Mpofana is part of a bigger problem. Most of South Africa's municipalities—the units of local government meant to deliver basic services—are a mess. Their failure makes life worse, especially for the poor, and stymies a wider economic recovery. That is weakening the African National
Where the grass is not greener
Congress (ANC), the once-mighty party. If the ANC receives the expected drubbing at local elections on November 4th, it may shake up South African politics.
Under apartheid, local government reflected the country's systemic racism. White areas were home to commercial hubs and rich suburbs; business and property taxes were spent on infrastructure that would not have been out of place in California (swimming pools that used a lot of water). Black areas were marginalised, with infrastructure that looked more like that elsewhere in Africa, rather than in its industrial heart (tap water was a luxury).
After Nelson Mandela was elected in 1994 his government abolished racially segregated local authorities, creating ostensibly integrated municipalities. Today these range in size from Johannesburg, the commercial capital, to units like Mpofana that comprise a small town and rural areas. All are supposed to fund themselves through property taxes, business rates and charges for water, electricity and waste collection. The poorest also get a grant from the central government to fill any gap between revenue and expenditure.
In reality, many do not balance the books. In the financial year that ended in June 2025 45% of municipalities passed unfunded budgets. Some 30%, including Mpofana, have done so in each of the past four years. Last month Enoch Godongwana, the finance minister, said he would withhold grants to the worst-offending places to instil fiscal discipline.
Public-sector trade unions argue that the grant increases have lagged behind the rising costs of providing services. But the bigger issue is that money is wasted or stolen. Just 15% of municipalities achieved clean financial audits in the auditor-gener-
al's latest report, in June. Since 2021, "wasteful", "irregular" or "unauthorised" spending has added up to 287bn rand ($18bn), more than South Africa has spent on defence in the same period. In a grimly symbolic case of apparent municipal graft, ten individuals, including local officials, allegedly pilfered funds designated for VIP (ventilated improved pit) toilets in Amathole in the Eastern Cape, one of the poorest places in the country.
Municipalities collectively owe more than 110bn rand to Eskom, the state-run utility, and 30bn to bulk-water suppliers. These debts threaten the "financial sustainability" of the economy, argues Mr Godongwana. His decision to suspend grants was partly in order to use the money to pay these bills. (On July 31st he began reinstating the grants, with new conditions.)
As if taking their cue from politicians, residents often do not pay up either. Almost half the water procured by municipalities is "lost": it flows out of broken pipes or is never paid for. In Mpofana 60% of electricity is in effect stolen. "Honestly, we do not pay," Ms Radebe admits, glancing behind her at the coils of rusty cables that lead to an illegal connection outside.
Yet some municipalities work. South of Mpofana is Umngeni, which in 2021 ditched the ANC and elected a mayor from the Democratic Alliance (DA), the second-largest party. Chris Pappas has been credited with an overhaul. In July the FW de Klerk Foundation, an NGO, rated Umngeni as the seventh-best-run municipality in a sample of 50, ahead of Cape Town.
Today Umngeni has funded budgets and receives clean audits. Poor-performing bureaucrats have left or been fired. Mr Pappas has cracked down on inflated procurement schemes. The savings have helped to ►
Middle East & Africa 41
► erect 3,000 new streetlights, fill potholes and expand a scheme for the very poor that subsidises small amounts of water and electricity. The once-dirty town centre is clean. New firms have moved in. More than 2,000 net new jobs have been created in an area of 100,000 people.
In Zanzene Village, a poor part of Umngeni, residents seem to appreciate the changes. “Everything is good: the roads, the water, we even pay for electricity,” says Nomfundo Buthelezi, explaining that she does not mind so long as the money is well spent. Musa Mchunu, a gardener, notes there is a new football pitch and a youth centre. “Most importantly,” he adds, “the municipality now responds—we no longer get the blue ticks,” a reference to WhatsApp messages that are read but ignored.
The DA hopes that a record of good governance will increase its vote share, including among black voters, who historically have opted for the ANC or its off-
shoots. Among its targets is Johannesburg, where one recent poll has the DA on 42%, up from 26% at the 2021 local election. The ANC, which has run the city into the ground, is on 18%, barely half the 34% it won in 2021.
Such highs will not be reached everywhere in South Africa. In the recent poll the ANC still remains ahead of the DA (34% versus 27%) on average. Race remains highly predictive of whether a voter opts for the DA, even though it runs just 12% of all municipalities but accounts for most of the clean audits.
Yet 32 years after the end of white rule, when most South Africans are despondent about democracy, showing that political choices can make a difference is crucial. Mr Mchunu, a former ANC member, says that he did not vote for Mr Pappas last time, but may do so in November. “I tell my sons that apartheid is over. What matters is not race but changing the country.” ■
Somaliland
Irro’s gambit
NAIROBI
The president wants to help the world secure the Red Sea
A BIDIRAHMAN MOHAMED ABDULLAH is an ambitious man. In December 2024 the 70-year-old former diplomat (pictured), who is better known as Irro, was sworn in as president of Somaliland, a breakaway republic that declared independence from Somalia in 1991. Last year he was the would-be country’s leader when Somaliland was formally recognised by Israel, the first nation to do so. “The world has been ignoring Somaliland for 35 years,” Irro tells The Economist. He wants outsiders to take a closer look.
Irro’s pitch centres on the Red Sea. The trade route has been a mess of late, with shipping menaced by the Houthis, an Iran-backed militia that rules much of Yemen. Speaking this month during a visit to Nairobi, Kenya’s capital, Irro says Somaliland can help secure the strategic waterway by offering military bases to global powers along its coastline. Once trade can flow unmolested again, he wants Somaliland’s main port at Berbera to be the primary gateway to the markets of the Horn of Africa. “Our message is that we are a responsible partner” in security and trade, says Irro.
He suggests that Somaliland could, in effect, be the new Djibouti. That tiny port-state rakes in some $125m a year by renting out slices of its coastline to foreign armies. America, China and France (among others) have bases there. It also earns many multi-
ples of that amount in annual port fees paid by landlocked Ethiopia, which depends on Djibouti for most of its international trade.
Irro wants a slice of both Djiboutian cakes. He is trying to persuade Ethiopia to use Berbera, where the United Arab Emirates (UAE) has built a new harbour, as its main outlet to the sea. He also wants to en-
Irro, man on a mission
tice America to set up a military base there. That would give America more options if it chooses to resume its war against the Houthis, whom it bombed intensively last year. Access to military facilities around Berbera is widely believed to have been part of the deal Israel struck with Irro when it recognised Somaliland in December. Some reports suggest that America and Israel may already be developing a military presence at Berbera, with covert support from the UAE, which operates an existing base there.
Irro denies such reports. “There is no Israeli or American presence yet,” he says. But the UAE has another base at the port of Bosaso in Puntland, a neighbouring semi-autonomous province of Somalia, where unconfirmed reports suggest Israel is also embedded. As competition for military properties on the Red Sea heats up, Irro says he would be willing to offer both Israel and America bases at Berbera “if they are interested”. Next month Irro will visit Washington, where he hopes to have a tête-à-tête with Donald Trump. Were that to happen, he would become the first president of Somaliland to meet directly with an American counterpart. “We expect the recognition of the United States some time in the future,” says Irro, adding that Somaliland anticipates recognition from at least one more country “before the end of the year”.
Irro’s ambitions are ruffling regional feathers. Djibouti is irked by Somaliland’s desire to barge in. Irro hints that his northern neighbour is trying to “destabilise” his country, presumably by supporting hostile militias on the border. Relations with Somalia are even more strained. Mr Irro claims Somalia, which considers Somaliland a renegade province, has plans to mobilise troops along the southern frontier. “This is not a peace recipe,” he adds.
It does not help that Israel and Turkey, a patron of Somalia, are increasingly hostile to each other. Turkey has long been trying to mediate between Somalia and Somaliland, yet numerous rounds of talks in Ankara and Istanbul have failed. “Next time they can happen in Tel Aviv,” chortles Irro, who seems unbothered by the prospect of a proxy war. Yet America may well be reluctant to stir the pot by recognising Somaliland.
The Houthis, meanwhile, have threatened to attack any Israeli military installation in the state. An American base there could become a target, too. That would shatter the reputation for peace and stability on which Somaliland has long built its case for international recognition. But “there is nothing free in the world,” says Irro. He seems willing to pay a price to put his country on the map—the question is, how high? ■
42
Asia
Japan
Frozen out
TOKYO
Japan's "ice-age" generation workers are squeezed, once again
FOR NEARLY two decades Torigoe Atsushi, who is 50, has worked in the stockrooms of a fashion firm in Tokyo. His income rose only modestly, and it recently fell after his employer cut bonuses and reworked its pay scheme; he now earns about ¥3.6m ($22,600) a year. Then he spotted the firm's recruitment ads, offering newcomers higher salaries. "It's great that younger people are getting better pay," he says. "But why not raise mine?"
Mr Torigoe belongs to Japan's "ice-age" generation: workers who entered the labour market between 1993 and 2004, after the country's economic bubble burst and many companies stopped hiring. The ice-agers often endured years of insecure work and low pay. Now in their 40s and 50s (thus Generation X), they are losing out again. Japan is emerging from decades of deflation, prices are rising and companies are handing out their biggest pay increases in a generation. But the ice-agers are not ben-
efiting. Worse may come when this cohort of baby-boomers' children—the last bulge in Japan's population pyramid (see chart on next page)—drifts into retirement.
The ice-agers' fortunes contrast starkly with those of the generation that came just before. During the bubble years of the 1980s, companies hired voraciously. College students found their mailboxes stuffed with offer letters, or were treated to fancy meals by employers keen to woo them. Everything changed as the bubble popped in the early 1990s. Japan's rigid employment system made it hard for firms to
shed existing workers, so they cut hiring instead. The ratio of job openings to college applicants fell from nearly three at its peak to below one by 2000. Nagai Toshihisa, who graduated in 1999, shortly after the Asian financial crisis, abandoned his hope of working for a securities firm. "It felt utterly stupid looking for jobs," he recalls. Many in his generation sent out dozens, even hundreds, of applications, to no avail. It did not help that the ice-age cohort was such a big one, adding to the oversupply.
Many graduates who failed to secure regular jobs ended up in temporary or contract work. Japan's labour market sharply distinguishes between regular workers, who enjoy strong legal protections and seniority-based pay rises, and irregular ones, who often do similar work for much less money and are the first to be laid off in downturns. Such jobs often used to be filled by housewives supplementing their husbands' salaries. Increasingly, they were taken by young people who needed to support themselves, says Kondo Ayako of the University of Tokyo. The term "freeters"—youngsters drifting between casual jobs—became popular, as did "parasite singles", who lived off their parents well into adulthood. There was also growing concern about hikikomori, social recluses.
Time and acute labour shortages have repaired some of the damage. The share of
→ ALSO IN THIS SECTION
43 The decline of the Indian elephant
44 Donald Trump upsets Asian allies
44 Kazakhstan's cosmetic election
45 Ashoka: The new founding fathers
Asia 43
► ice-agers in regular employment has largely caught up with that of older cohorts, says Shimoda Yusuke of the Japan Research Institute. But their earnings have not. Those who spent long stretches in irregular work tend not to fully catch up even after switching to the regular track. Those in regular jobs faced another problem: the upper ranks of their companies were crowded with the bubble-era workers, delaying promotions for those who followed. Japan's low labour mobility has meant the unlucky cohort cannot easily seek better pay elsewhere.
Japan's economy is changing. After three decades of stagnation, wages are rising as inflation returns, and job-hopping is becoming more common. Yet these improvements are mostly benefiting the young. Companies short of labour are competing fiercely for fresh blood, and paying more to attract and retain it. Middle-aged employees are much less likely to switch jobs, giving firms little incentive to raise their wages. According to Kumano Hideo, an economist at Dai-ichi Life Research Institute, nominal wages for college-educated workers in their 20s and 30s were 10-16% higher in 2025 than five years earlier. Meanwhile, those in their early 50s suffered a nominal decline of 1.3%. Abe Katsuya, 48, says the large pay rises dished out to the young give him 'mixed feelings'. He is part of the 'Mammoth Club', a support group for ice-agers who identify with the woolly beasts that endured harsh times.
All this has left the ice-agers with low lifetime earnings (and therefore little in the way of savings). That points to a looming crisis as they reach old age. Japan's pension system ties payouts to lifetime earnings, so decades of suppressed pay will translate into low pensions. 'Those who happened to come of age in a down-
turn are likely to carry that handicap for the rest of their lives,' laments Mr Kumano. Getting them to work longer, as a growing number of old people already do, is one fix but it will not solve everything. Ms Kondo warns that many may end up on welfare, which is not designed to support the elderly over extended periods. That would be costly; for instance, the state pays welfare recipients' medical bills in full. And because the generations behind the ice-agers are smaller, fewer workers will have to support more pensioners.
The woes of the ice-age generation have long had the government's attention.
For decades, authorities have come up with programmes to help them find jobs and nudge reclusive ones to connect with the outside world. Yet as the ice-agers grow older they face new problems. Beyond thin pensions and savings, questions loom, for instance, over housing; compared with older Japanese, this cohort is less likely to own homes, and Japan's rental market is unfriendly towards older tenants. The government is belatedly adjusting: in April, it produced a three-year plan for the ice-agers, spanning measures from asset-building to housing. But a lot more serious thinking will be needed. ■
India
No room to roam
MUMBAI
Indian elephants are being squeezed out of their habitats
IN 'ARTHASHASTRA', an ancient Indian manual on statecraft, elephants were prized assets. Kings presented them to forge alliances. Generals relied on them to destroy the gates of forts. Today Ganesha, the elephant-headed Hindu god, is still worshipped as the remover of obstacles, and children, as elsewhere, learn 'E for Elephant'.
But reverence has not translated into protection. A survey published last year by the Wildlife Institute of India counted 22,446 elephants, 18% fewer than in 2017. The decline in some states was 'a shocker', says Qamar Qureshi, the report's lead author. Part of the problem is geography. Unlike in Africa, villages are scattered throughout the forests of densely populated India. Elephants are often confined to narrow corridors between patches of forest. Vivek Menon of the Wildlife Trust of India says that in 2010 his team identified 88 such routes. Today there are about 150, reflecting shrinking habitats.
As infrastructure cuts through forests, elephants stray into farms. Mining and blasting sites also get in their way. In West Bengal hulla ('noise') parties see mobs hurl fireballs and firecrackers to drive the animals away, despite a Supreme Court ban. Electric fences sometimes block the elephants' path. But they are excellent physicists, says Raman Sukumar, an ecologist. They use their tusks, which are poor electrical conductors, to break fences, or toss logs to short-circuit them. Forest guards and workers steering elephants back to safety are 'utterly under-equipped and woefully unprepared', says Prerna Bin-
dra, a conservationist. Most lack basic equipment such as torches.
But not all is doom. Poaching for ivory has fallen since the 1980s. Coordination between state forestry departments and national customs has improved. Today India has a Wildlife Crime Control Bureau and states have anti-poaching plans. Last month India's Supreme Court upheld the removal of commercial activity from an elephant corridor in Tamil Nadu.
Elephants act as landscape architects. They disperse seeds, shape forests and fertilise soil. They keep wetlands alive by digging. 'Arthashastra' advised India's rulers to protect forests inhabited by the magnificent beasts. Those in charge today would do well to take note.
Never forget
Boomer bulge
Japan, population by age group, 2026, % of total Total population=122.4m
Source: UN Population Division
44 Asia
America and South Korea
Alliance mismanagement
Donald Trump abruptly scales back an exercise with South Korea
ULCHI FREEDOM SHIELD, an annual military exercise, brings thousands of American and South Korean troops together to train to counter threats from North Korea. This summer's edition began on August 17th, and the allies planned to focus on responding to new capabilities that North Korean soldiers acquired while fighting in Russia's war against Ukraine, such as drone warfare. But the night before the exercises were due to start, President Donald Trump ordered his Department of War to "substantially reduce" (by six days) the manoeuvres, citing the cost of the drills, South Korea's refusal to support his war in Iran and his "very good relationship with Kim Jong Un", North Korea's dictator.
Mr Trump presumably hopes to entice Mr Kim to restart the diplomatic engagement the two men began during his first term in office. He may also want to press South Korea into accelerating its promised investments into America. And he doubtless hopes to distract attention from his Iranian quagmire.
But his abrupt decision sends worrying messages to both friends and foes across Asia. The prospect of more talks between the leaders, however unlikely, causes anxiety in the capitals of American allies. (North Korea called Mr Trump's move an empty gesture, though Mr Trump insisted on August 19th that he would be meeting Mr Kim later this year.) Japan, in particular, worries that Mr Trump could cut a bad deal that undermines its security. The move heightens fears of American abandonment that have been growing since his return to office. Adversaries may conclude that America's commitments to its allies are hollow.
This is not the first time Mr Trump has used exercises with South Korea as a bargaining chip. When he met Mr Kim in Singapore in 2018, the first summit between American and North Korean leaders, Mr Trump said that America would halt upcoming drills with South Korea. The move stunned his own advisers. But the unilateral concession did not bring about a breakthrough; a subsequent meeting between the two men, in Hanoi the following year, collapsed in acrimony.
South Korea's president, Lee Jae Myung, shares Mr Trump's desire to revive engagement with the North. On August 19th-20th his government hosted China's foreign minister, Wang Yi, in Seoul, in part
to discuss ongoing efforts to reach out to Mr Kim. Yet Mr Lee also acknowledges the importance of maintaining deterrence, which Mr Trump has weakened.
Mr Trump has long been sceptical of the value of America's alliance with South Korea. During his first term he sought to withdraw American troops from the country, but was stymied by Congress. Since returning to office, his administration has made quiet but consequential changes to the alliance. America has, for example, moved to reorient the mission of its 28,500 troops in South Korea away from their traditional focus on the peninsula and towards deterring China. He has also used tariffs to extract a pledge from South Korea to invest $350bn in America. Mr Trump has not hesitated to link the future of the security alliance to the flow of investment deals. His decision to curtail the exercises came as South Korea's industry minister was visiting Washington to discuss the stalled implementation of those plans.
Even before the latest announcement, Mr Trump's transactional approach to the alliance had undermined trust in America. Polling by the East Asia Institute, a think-tank in Seoul, found the share of South Koreans who trust America plunged from 85% in 2022 to 46% this year. Fear of abandonment has also fuelled discussions about South Korea acquiring nuclear weapons. Mr Lee has said he opposes that course. But he has sought to develop the capacity to enrich and reprocess nuclear fuel for the civilian nuclear-power industry, which would make building the bomb easier if South Korea ever deemed it necessary. Following Mr Trump's announcement, he reiterated his intention to move forward with those plans.
Similar dynamics are at work across the region. Perceptions of American reliability have fallen steeply between 2022 and 2026
There, but for how long?
in Australia, Japan and Singapore, according to the Pew Research Centre. Japan has also sought to beef up its own defences; the drive for greater autonomy is shaping revisions to its national security strategy expected later this year.
Yet in the face of threats from a richer and increasingly assertive China and a nuclear-armed North Korea, American allies in Asia find it hard to imagine their security without America. Despite the falling level of trust, more than 60% of South Koreans still see the alliance with America as the best way to ensure their security. Such results, writes Sohn Yul, the president of the East Asia Institute, reflect a "contradictory situation where [South Korea] does not trust the United States yet has no choice but to rely on it." The dilemma is shared across Asia. ■
Kazakhstan
Reset or repress?
ALMATY
A cosmetic parliamentary election in an authoritarian ex-Soviet republic
"VOTE FOR renewal," urge giant billboards as Kazakhstan heads into a parliamentary election on August 23rd. Kassym-Jomart Tokayev, the president, calls this a turning point in the building of a "new Kazakhstan"—shorthand for a political reset that began after civil unrest demanding political change had left at least 238 people dead in 2022.
At first glance, much about this election is indeed new. Taking place under a new constitution, it will deliver a parliament with a new structure and a new name, certain to be dominated by a new party. This, ostensibly, is the reforming "new Kazakhstan" in action. But scratch the surface and the political trappings of the old Kazakhstan—associated with Nursultan Nazarbayev, Mr Tokayev's long-serving predecessor and discredited political patron-emerge. The newly formed Justice party that is set to win a big parliamentary majority is in fact a revamped version of the old ruling party that dominated for decades. And parliament will still be a rubber-stamp affair, because no genuine opposition parties are being allowed to run.
Pluralism is thriving in Kazakhstan, so Mr Tokayev said recently—only days after Amangeldy Dzhakhin, an opposition leader, had been jailed for seven years on spurious charges of extremism. His predecessor suffered the same fate, and it is not just politicians who are targeted. Reporters and social-media personalities have also been variously imprisoned, slapped with ▶
Asia 45
► travel bans and barred from doing journalism on flimsy-looking charges. In a blatantly political case, a satirical post poking fun at the president recently landed a YouTuber behind bars.
This election follows the adoption of a new constitution, which Mr Tokayev has depicted as Kazakhstan “casting off the super-presidential form of rule”. That was a reference to his predecessor, who monopolised power for three decades before resigning in 2019. Mr Nazarbayev’s designat-
ed successor was Mr Tokayev. The pair ruled in tandem until the 2022 unrest, which left Mr Nazarbayev sidelined and Mr Tokayev scrambling to dissociate himself from his former mentor. He rolled back some presidential powers in 2022, but clawed some back again in the new constitution, adopted by referendum against a backdrop of arrests of critics.
One important reform remains in force. In 2022 Mr Tokayev introduced a single presidential term-limit. He has pledged to
abide by this, but the constitution offers him a loophole. He has the right to a new “first” term, the Constitutional Court recently ruled, so he could run again—a tried-and-tested tactic to cling to power loved by Central Asian autocrats as well as by Vladimir Putin in Russia and Recep Tayyip Erdogan in Turkey. The end of his current (second) term in 2029 will thus become a litmus test of Mr Tokayev’s professed commitment to ridding Kazakhstan of monarchical rule. ■
ASHOKA
Refounding fathers
Nation-building is even harder the second time round
EVERY COUNTRY has a national anthem. Some, like New Zealand, have two. But only India has an official song as well. This year, for the first time, “Vande Mataram”, a rousing tribute to “Mother India”, featured in Independence Day celebrations, greeting Narendra Modi as he arrived to give his annual speech on August 15th. An air-force helicopter flew past with a banner commemorating 150 years of the song’s composition. It marked the culmination of a years-long effort by Mr Modi’s Bharatiya Janata Party (BJP) to boost its status. A few days earlier it had become a criminal offence to disrupt its singing.
Mr Modi spoke of India’s achievements as it enters its 80th year of independence. He looked ahead to its glorious centenary. Such pronouncements acknowledge that India gained freedom from the British in 1947. Yet his BJP believes the country has been truly free for much less time. The clearest articulation of this sentiment came from Kangana Ranaut, a Bollywood actress and since 2024 a BJP lawmaker, who declared in an interview some years ago that India only achieved independence with the election of Mr Modi in 2014.
The statement is one of those to be taken seriously, not literally. In the eyes of the BJP’s leaders, the India they inherited was too effete, too anglophone and too accommodating of minorities. Their project was to build a “civilisational state defined by the country’s ancient Hindu heritage”, as Milan Vaishnav of the Carnegie Endowment, a think-tank in Washington, put it in a paper in 2024. He described the new political order as India’s “Second Republic”.
A new republic needs a new identity. Around independence in 1947 the constituent assembly settled on what India’s
national symbols should be. Over time, civic bodies picked up the statuary of British kings, queens and sundry dukes and deposited them in far-off parks (or, deliciously, in the zoo). State governments renamed cities and institutions.
The BJP has been on a similar mission since 2014. The obvious purpose is to promote its Hindu-first vision. The national song was passed over for anthem status in part because its invocation of Hindu deities was seen as excluding Muslims. That is no longer a concern. In recent years the government has changed to saffron—a colour associated with Hinduism and the party itself—its flagship train service, the logo of the public broadcaster and, last month, the uniform of the men’s national hockey team. Last year it overhauled a jobs scheme named after Mahatma Gandhi and rebranded it as VB-G RAM G, a tortured acronym reverse-engineered to include the name of Lord Ram.
For all that, it is the non-religious parts of the BJP’s project that more closely resemble the nation-building of a young country. Along with national symbols and
a shared vision of the future, a new republic must have an origin story to give it meaning. Unfortunately for the BJP, the freedom movement and its pantheon of heroes already fulfilled that purpose. That left two strategies.
The first was to repurpose long-dead figures. Jawaharlal Nehru, India’s first prime minister, was demoted from nation-builder to incompetent, entitled scion of the elite. In his place the BJP’s myth-makers built up Sardar Vallabhbhai Patel, Nehru’s home minister, who they insist was robbed of the top job. (Mr Modi’s home state of Gujarat has immortalised Patel with the world’s tallest statue.) In any case, they say, Subhas Chandra Bose, a revolutionary with no time for Gandhi’s non-violence humbug, was the real first prime minister.
The other is to lean on the one towering figure they do have. In June the cabinet feted Mr Modi for becoming India’s “longest serving elected PM for consecutive terms”. The heavy caveating excludes Nehru’s first five years (elections had not yet been held) and Indira Gandhi (she lost power and then regained it). The animus against Nehru is understandable. There cannot be two first prime ministers of independent India.
The BJP may have saffronised India. But its revisionism has failed to take. That much was clear during protests for education reform last month in Delhi. Students carried copies of the constitution. They displayed photos of B.R. Ambedkar, its architect; Bhagat Singh, an anti-colonial revolutionary; and Gandhi. They chanted slogans dating to the freedom struggle. Despite having spent most of their conscious lives in the second republic, their symbols were those of the first. Building a nation is no easy task. Retrofitting one is harder still.
46
China
China in the Arctic
Pole position
The Northern Sea Route is an alternative to the Suez Canal—albeit costly and risky
FOR A WORLD newly preoccupied with maritime choke points, the voyage of the Dubai Tower offers a flicker of promise. The container ship set off from the Chinese port of Ningbo on August 15th to begin the first regular cargo service between Asia and Europe via the Arctic Ocean. Sea Legend, the vessel's Chinese owner, says that it will complete the 5,000km journey to Felixstowe, on Britain's east coast, in about 20 days. That is roughly twice as fast as sailing the usual route via the Suez Canal. Crucially, the "Arctic Express" (as the company calls the service) avoids the risk of being entangled in the Middle East's military conflicts.
The new service via the so-called Northern Sea Route (NSR), which runs through the Bering Strait and over the top of Russia, is an early test of China's Arctic ambitions. In 2018 the country outlined plans to develop shipping routes, mining and infrastructure in the Arctic, hoping to
take advantage of rapidly receding polar ice. Many of the plans, which are marketed collectively as a "polar silk road", have met resistance. Seven of the eight countries with Arctic territory are Western democracies. They worry that China is working closely with Russia in the region and that Chinese activities there could be useful for future military purposes, such as operating submarines. Some bridled at China, whose northernmost point is on about the same latitude as Hamburg, declaring itself a "near-Arctic state".
Nonetheless China hopes to attract fresh interest in the northern route from governments and companies alarmed by the closure of the Strait of Hormuz and the
threat from Iranian-backed Houthi fighters to shipping in the Red Sea. China is also more anxious than ever about its own reliance on the narrow Malacca Strait, the hinge between the Indian Ocean and the Pacific, through which most of its energy imports pass. And China will soon not be alone on the NSR. In June Japan said that it would revise its Arctic policy to promote the route. A South Korean shipping company, PanStar, plans to send one of its container ships on a trial run through the Arctic Ocean on August 22nd, aiming to complete the voyage from Busan to Felixstowe in just 18 days.
It's no sleigh ride
Thanks to climate change, conditions for the NSR are likely to grow more favourable. The minimum summer extent of Arctic ice has been shrinking by more than 12% a decade. In March the maximum extent of winter ice reached its lowest level since records began in 1979. On August 16th, the latest available reading, Arctic ice covered 5.5m square kilometres. That is well below the median extent of ice on the same date in the years between 1981 and 2010 but above the lowest ever extent, recorded in 2012. That September, the ice shrank to just 3.4m square kilometres.
One selling point for Sea Legend's service is that a shorter sailing time means ▶
→ ALSO IN THIS SECTION
47 Sweet on chocolate
48 Chaguan: Nostalgia for the 1990s
China 47
▶ less fuel burned—fuel accounts for seven-tenths of shipping costs. That also reduces carbon emissions. And cooler weather on the Northern Sea Route is better for transporting some high-tech goods that are sensitive to temperature. Without refrigeration, temperatures inside containers on the Suez Canal route can reach 60°C. The Dubai Tower is carrying goods including batteries, solar modules and electric-vehicle components, mainly from cities in China's Yangzi river delta.
For all those apparent advantages, the Arctic Express also highlights the limitations of the NSR. It will run only from August to October, when the route is largely free of ice. Sea Legend plans eight voyages in that window, using seven vessels. For the rest of the year, only ships with reinforced hulls may navigate the Arctic. Sea Legend has bought or leased some of these, but only relatively small ones, as they are costly and in short supply. They may also need to be escorted by Russian nuclear-powered icebreakers. All the extra costs erode the northern route's price advantage.
So do insurance premiums. They are usually higher for the NSR than for the Suez Canal, despite war-risk premiums for the canal. They are also 40% higher than for sailing around the Cape of Good Hope, the chief alternative to the Suez Canal. Allianz, an insurance giant, warned in its recent annual safety review of the shipping industry that the remote Arctic route posed significant extra risks because of its harsh weather and the limited infrastructure available for repairs, assistance, rescue or salvage. It said that machinery damage or failure caused almost half of the reported shipping incidents in Arctic waters in the past decade.
The environmental impact is potentially greater, too. Bellona, a Norwegian environmental group, warned in a report last year that in the Arctic there are no effective ways to deal with fuel spills because of the difficulty of cleaning up fuel in cold conditions and the distance from emergency-re-
response centres. (The International Maritime Organisation has banned the use and carriage of heavy fuel oil in the Arctic since 2024, though Russia does not heed the ban.) Noise pollution from ships also affects marine mammals such as whales and walruses in the Arctic more than elsewhere, since low-frequency sounds travel farther in cold water. Ships on the northern route pass through vital feeding and breeding grounds.
International carriers avoid the route in large part because of geopolitical risk. Ships on the NSR often need permits, ice-breaking services or other assistance from Russia, potentially violating Western sanctions linked to its war in Ukraine. Shipping may in future be disrupted by an intensifying contest between Russia and other countries to control the region's ample mineral resources. The Arctic, once preserved as a zone of peace, is becoming increasingly militarised, as Donald Trump continues to threaten Greenland and NATO seeks to deter Russian aggression on its northern flank.
All this helps to explain why total transit cargo on the Northern Sea Route was just 3.2m tonnes in 2025, up from about 3m the year before. That is tiny compared with about 464m tonnes through the Suez Canal last year. The bulk of Arctic shipping is between Russian and Chinese ports, rather than between Asia and Europe, much of it consisting of shipments of Russian oil or gas. In a recent report, the Clingendael Institute, a Dutch think-tank, concluded that the route's "most viable function...remains that of a selective corridor for Arctic resource exports."
Still, as Malte Humpert of the Arctic Institute in Washington argues, the NSR will grow in importance for China as a seasonal trade corridor for when traditional routes are disrupted. He envisages hundreds of annual container passages a decade hence. "The question is when Western operators decide to re-enter the market—and more importantly, whether they still can," Mr Humpert says. Arctic navigation requires training and operational experience, and "China and South Korea are building that expertise now."
For the moment, Sea Legend admits, the tight seasonal window and limited capacity of its Arctic service make it hard to compete with conventional routes. But it is already planning to expand the navigational window to four months by 2028 and eventually to the entire year. Further conflict in the Middle East could provide incentives for it and other companies to invest in larger, ice-strengthened ships. And if temperatures in the Arctic continue to rise three times faster than the global annual average, then doubts about the viability of the NSR may recede as fast as China's ambitions for the region grow. ■
Chocolatiers
Eating bitterness
SHANGHAI
Can Chinese become choco-mad?
CHINA HAS more coffee-shop brands and more KFC branches than anywhere else on Earth. But chocolate has never hit the spot, with the average Chinese eating just 100g a year—a mere hundredth of what a French person guzzles. A few mass-market foreign brands—Mars, Ferrero, Nestlé and Hershey—account for four-fifths of Chinese consumption. "Of all the chocolate eaten worldwide, only 2% caresses Chinese tastebuds.
This may now be changing, thanks to a handful of Chinese chocolatiers springing up around the country. Most are artisanal shops, but some are investing in supply chains and building big businesses. Take Choc Revive. Based in Shanghai, the firm was founded in 2023 by Xie Liang, a fresh-produce trader. He cannot open shops quickly enough. From 80 outlets now he hopes to have 300 by the end of the year. Mr Xie says that his company had 160m yuan ($23.7m) in sales last year and hopes to quadruple that, or more, this year.
Choc Revive is trying to take chocolate local. It sources its own beans from Africa, but has set up an R&D unit in China and does its own roasting and grinding. Its big investments are intended to produce economies of scale, but the know-how it is accruing may spread and give rise to competitors. Another local chocolate company, Nibbo, has been winning international awards for its creamy sweets. Saturnbird, which started as a coffee grower in south-west China, is now planting its own cacao trees there.
Mr Xie's concoction is 58 yuan for a 100g chunk—equivalent to a cheap meal for two. It is an awkward time to be selling premium products to belt-tightening consumers. Churning out the cheapest possible goods to undercut rivals is the usual formula. Milk-tea brands are a case in point.
Why are people paying out for chocolate? Mr Xie says that making chocolate with Chinese flavours such as longan, which is like a large lychee, seems to go down well. A simpler explanation is that, at long last, Chinese people are learning to love the most delectable flavour in creation.
*Approximate, dependent on routing Sources: China Shipowners' Association; Sea Legend
48 China
A glum China pines for the 1990s
Back then, leaders delivered and opportunity beckoned—or so the story goes
THE 1990S ARE alive again. Over the past few days viewers of state television in China have feasted on scenes from that decade. A 12-part documentary has aired in prime time to celebrate the centenary of the birth of the late Jiang Zemin, the country's leader throughout the period. What a decade it was, bookended, in 1989, by the blood of Tiananmen and, in 2001, by all the promise of China's entry into the World Trade Organisation.
This progression, which may seem inevitable in retrospect, was never anything but fraught. But what captivates viewers is less Jiang's accomplishments than a common touch and a panache that seem nearly inconceivable among China's leaders and their tightly scripted public comportment today. There he is in one scene in 1989, four months after the crackdown on the Tiananmen pro-democracy protests, hugging babies in Beijing. In another, Jiang slams a forehand in table tennis before telling national-team athletes that he hopes they win gold at the upcoming Asian Games. "Let me take that back. You shouldn't have too much pressure," he says with a big smile, laughter erupting in the room. In a third, Jiang is belting out Beijing opera at a formal dinner in California, a study in contrast with the buttoned-up nomenklatura of today's China, Xi Jinping above all.
Then, as the documentary ran, a cosmic coincidence: on August 12th Jiang's essential partner, Zhu Rongji, died at the age of 97. If Jiang was the political maestro, Zhu, his prime minister, was the one who got his hands dirty. Online commentary praised Zhu's blunt, effective approach—an implicit rebuke of today's more cautious officials. The day he died, the documentary happened to feature Zhu explaining how he cajoled cadres into tough reforms: "I'm not here to talk about pros and cons. I am here to talk about how to get things done."
Even before the documentary and Zhu's death, nostalgia for the 1990s had been building. A viral trend last summer spoke of "the beauty of the boom years", with people sharing images of youthful exuberance in the 1990s and the early 2000s. A recent television hit, "Blossoms", about Shanghai in those days, captures the mood. At one point the narrator says: "Everyone stands equal before opportunity. Seize it, and you could change your life." Some
have echoed the sentiment as they remember Zhu. "Back then people had a spark in their eyes, certainty in their hearts and a tomorrow they could actually see," wrote Zhao Jian, an economist, in a widely shared piece—before it was scrubbed from the internet.
What to make of all this? A cynical take is that memory is playing its usual tricks. If the 1990s now look like a decade of possibilities opening up, recall that China was so much poorer then, on the cusp of high-speed development. In 1990 its GDP per person, adjusted for inflation, was just $920; today it is 15 times higher. Or consider the rise in university graduates: 12m a year now, up from about 700,000 in the 1990s. In the Jiang-Zhu era nearly three-quarters of China's population still lived in the countryside, which was mired in poverty. No wonder urbanites lucky enough back then to get a good education felt that destiny was theirs to chart. Yet for most in China, the path ahead was one struggle after another.
So the nostalgia should not be read as a rigorous account of history. Rather, it acts as the reflection of a malaise today. Most obviously, people are down about the economy. Job opportunities are pinched, property has slumped and many entrepreneurs are struggling. Yet to dwell on these issues carries consequences, with, for instance, censors erasing posts about the economy that lack "positive energy". For some, harking back to how good things were in the 1990s is a way round such controls. For others, it may simply be a longing for a better life.
There is also great frustration about politics in China today. Scholars debate whether Jiang and Zhu really deserve their "reformist" labels. Political reformers they certainly were not: Communist Party elders chose them to reinforce party will and pick up the pieces following the Tiananmen trauma. As economic liberalisers, there were also limits. Their signature overhaul of state-owned companies was to "grasp the large and let go of the small", hiving off useless businesses while modernising the important ones in finance, infrastructure and heavy industry. The aim, as Jiang put it, was to retain control of "the lifeline of the national economy". Above all, they aimed to strengthen the state and entrench party rule.
They do things differently there
For all that, Jiang and Zhu were certainly daring, taking bold risks against real resistance. Conservatives (ie, leftists) opposed their market-driven policies, worried that opening China to trade with the world would suck in toxic political ideas. Many reforms also exacted a vast human toll, with more than 30m workers laid off—though the eventual benefits were larger. Do today's leaders have similar vision and courage? True, they have been in charge during some high-tech advances. Yet a trade war with America drags on, while property remains stuck in its years-long funk. Tributes to Zhu came not just from ordinary people but from universities, legal associations, financial groups and think-tanks. It is hard not to see their veneration as a plea for wiser leadership.
Beyond that is something unquantifiable about the style of the leaders in the 1990s—about the frankness with which they spoke, their sense of warmth and real humanity, their ease when out and about among the public. Certainly, people were also wary of Jiang and Zhu; under them, China was hardly free. But they were admired and even loved. Today, when people talk of Mr Xi, the dominant emotion is often fear. Many Chinese are so circumspect as to avoid mentioning his name altogether. He is the paramount ruler, not a chap who cracks jokes or bursts into song. ■
49
International
Missile defence
Going Winchester
KYIV AND TAIPEI
America and its allies are running out of interceptor missiles
RESIDENTS OF KYIV are sleeping once again in overcrowded metro stations or in cars in underground garages. Above, the air war is turning into a turkey-shoot. Russia is bombarding the city with ever more ballistic and hypersonic missiles. Ukraine, which had previously thwarted most such attacks with Patriot anti-missile batteries, is “going Winchester”—military jargon for running out of ammunition.
Ukraine’s plight is the most acute sign of the West’s shortage of interceptors. America’s depleted arsenal also helps explain why President Donald Trump has repeatedly retreated from threats to return to war with Iran. America’s allies are alarmed.
Switzerland expects its order of Patriots to be delayed by up to seven years. Taiwan worries that 102 ordered years ago will not arrive this year as scheduled. Lockheed Martin, which makes the most advanced version, the PAC-3 MSE, says it cannot guarantee delivery times to any foreign client, given the Pentagon’s power to divert them.
Shortages of munitions are common in wartime, especially in big, attritional wars. Yet America is running short after a brief war against a medium-sized power. What if it had to fight mighty China? It would exhaust its interceptors within days, and be unable to use its bases in Japan and Guam, reckons a recent study by the Heritage Foundation, a think-tank in Washington.
The proximate cause of the crisis is the prodigious use of interceptors by America and its Middle Eastern allies during Mr Trump’s war on Iran this year. The Centre for Strategic and International Studies (CSIS), another think-tank, thinks the Pentagon has fired nearly two-thirds of its pre-war stock of Patriots as well as lots of other interceptors (see chart on next page).
But the war merely hastened a reckon-
ing that has been decades in the making. On both sides of the Atlantic defence industries work to a peacetime rhythm, though the drums of war have been beating faster. That error has been compounded by the global proliferation of missiles, generals’ tendency to spend more on flashy weapons than on the munitions they consume, America’s dysfunctional budget process and complacency about American military supremacy in the air.
The CSIS reckons it will take until 2029 to rebuild Patriot stocks to pre-war levels. That is assuming Congress approves Mr Trump’s gargantuan $1.5trn defence-budget request (a 50% increase on current spending), which is unlikely. Yet Patriots have become the prime system for point-defence—that is, to protect a specific target such as an air base—for 19 countries. “Patriot is a dinner-jacket solution: you don’t need it very often, but when you do, very little else will do,” explains Justin Bronk of the Royal United Services Institute (RUSI), a British think-tank.
The hitch is that ballistic missiles are generally cheaper than the interceptors trying to hit them (PAC-3 MSE rounds costs $4m-5m apiece). Iran’s Shahed drones (or other unmanned aerial vehicles, UAVs) are two orders of magnitude cheaper still. “If your adversary can field a one-way attack UAV for the price of a used car, and your answer is a surface-to-air missile that costs
→ ALSO IN THIS SECTION
— The Telegram is away
50 International
► the same as a small apartment here in London, you're losing the exchange even though you're winning every single engagement," Colonel Sabah Al Sabah, a Kuwaiti officer, argued at a recent RUSI event. What is more, defenders typically fire at least two interceptors at an incoming missile to improve the odds of success.
Ukraine has improvised a system of layered defences. It has a world-class network to counter swarms of Russian drones. It can also take out most larger cruise missiles, thanks in part to European air-defence systems. But only the precious Patriots can take on Russia's Iskander-M and Kinzhal ballistic missiles. In March Ukraine boasted of intercepting about 70% of ballistic missiles. In July it was 20%. In August the rate will be even more dire. On one night—August 5th—there were no interceptions. The air force has stopped issuing regular updates. Russia, taking note, is increasing missile production.
With winter approaching, Ukraine has asked for at least 300 Patriot interceptors. Some countries still have them in reasonable numbers, including Germany, Greece, Poland and Spain. But they are increasingly unwilling to part with them.
Another solution is to build them faster. The manufacturers agreed this year to increase PAC-3 production from 600 a year to around 2,000 by the early 2030s. The Pentagon has since solicited ideas to go faster still. But Tom Karako of CSIS notes that it has not yet signed actual contracts for stepped-up output.
Allies' industrial capacity can help. Japan produces some 30 PAC-3 interceptors a year under licence, mostly for its own use. MBDA, a European firm, is about to start production in Germany of older PAC-2 models. At a NATO summit in July Mr Trump said Ukraine, too, would be allowed to make Patriot interceptors under licence, but has since cast doubt on the idea. Reports suggest their manufacturers, Lockheed Martin and RTX, fear losing their know-how, or being upstaged by cheaper Ukrainian versions.
Lacking the same combat experience, Europeans firms are hurrying to make their own weapons. The Franco-Italian SAMP-T is supposedly able to hit short- and medium-range ballistic missiles, but Ukrainian military sources say it works less well than Patriot PAC-2s, let alone PAC-3s. Ukraine will receive improved SAMP-T batteries, with updated interceptors, in phases from next year. Yet they are being produced in lower numbers than Patriots.
Ukraine's nimble defence industry and upstart Western firms have revolutionised drone warfare. They are now working on missile defence. One effort is the Freyja project involving Fire Point, a Ukrainian firm, and lots more European armsmakers. They aim to develop an interceptor at a
quarter of the cost of a PAC-3 without American technology, with testing to start next year. Sceptics abound. Douglas Barrie of the International Institute for Strategic Studies, another British think-tank, describes Freyja as "a moon shot".
Taiwan is trying to expand production of its Sky Bow III interceptors, which are supposedly equivalent to PAC-2s and older PAC-3s. This year it is expected to start mass producing the Sky Bow IV, which it claims is more capable than the PAC-3. Few countries, however, would risk China's wrath by buying Taiwanese weapons.
Lockheed Martin has announced the development of the PAC-3 ACE, aka "Baby Patriot", which is smaller and less capable than the PAC-3 MSE but will cost half as much. It could enter production in 2028.
Some hope that laser weapons, long in development, might improve the grim arithmetic of air defence. But experimental American and Israeli weapons have so far reported success only against drones.
Kill the archer first
"Air defence cannot be separated from offensive operations," notes David Deptula, a former American air-force general. "The cheapest interceptor is the one you never have to fire because the weapon was never launched." There is a lot to be said for shooting the archer—in the form of enemy production sites, command-and-control centres, storage depots and missile launchers—rather than the arrow. But America has depleted its stocks of long-range offensive weapons, too. CSIS reckons it has used a third of its 3,000-odd Tomahawk cruise missiles. It has refused to supply them to Ukraine and deliveries to other clients, including Australia, Germany, Japan and the Netherlands, may be delayed.
Ukraine has built drones able to reach targets ever deeper in Russia, notably energy infrastructure, arms factories and consumer-goods-distribution warehouses. But they are slow and carry little explosive. Fire Point's heavier Flamingo cruise missile has had some success. Ukraine hopes to deploy soon its FP-7 and the bigger FP-9 ballistic missiles, supposedly with a range of more than 800km.
Destroying mobile launchers is tricky. This year America and Israel suppressed Iranian missile fire by up to 90% by bombing launchers and the entrances to underground "missile cities". But even with full control of the skies and exquisite surveillance, they never stopped it entirely. America lost radars, command centres, early-warning aircraft, refuelling tankers, drones, fighters and more. Since the ceasefire in April, it has been unable to stop Iran shooting at ships in the Strait of Hormuz.
Any future air war against China over, say, Taiwan, would be a formidable enterprise. Chinese missiles can reach bases as far away as Alaska and Hawaii, as well as warships in the middle of the Pacific Ocean. To defend against the People's Liberation Army (PLA), Heritage reckons America would need more than 19,000 Patriot interceptors—more than 20 times CSIS's estimate of current stocks.
Shooting at the PLA archer would be a tempting alternative, but poses two big problems beyond penetrating China's air defences. First, America would have to strike the Chinese mainland, which would invite retaliation against the continental United States. Second, the PLA Rocket Force is responsible not just for conventional missiles, but also land-based nuclear ones: attacks on it might be seen as an attempt to destroy China's nuclear forces, risking nuclear escalation.
Instead, American forces are practising how to survive within China's "weapons-engagement zone" through various forms of dispersal. In a world of proliferating, cheap drones and missiles, hardening, dispersing and concealing military targets—and rebuilding them quickly if they get destroyed—are essential skills.
Although deep-strike weapons are no panacea, they do help to deter attack. That explains why America's allies are scrambling to find the means to shoot at faraway targets. Ukraine, already under relentless attack, hopes that carrying the war to Russia will cause enough pain to make Vladimir Putin reconsider endless conflict.
All the possible solutions to the interceptor shortage, even if accelerated and intensified, will take time. In Ukraine, meanwhile, civilians must still shelter underground. As they contemplate another freezing winter of peril, many will ask how their leaders let it come to this. Other governments should take heed. ■
Winchester on the horizon
United States, estimated munitions stockpiles, % Feb 27th-Aug 16th 2026
Used estimate: ■ Low ■ High ■ Remaining
Air-defence interceptors
Long-range strike missiles
Source: M. Cancian and C. Park, Centre for Strategic and International Studies
51
Business
Global commerce
Routes and branches
AIN SOKHNA
Chinese businesses are wrapping their supply chains around the planet
THE ANCIENT port of Ain Sokhna, on the Gulf of Suez, once received turquoise destined for the regalia of pharaohs. Now it welcomes a new source of wealth: Chinese manufacturers. In just a few years their investment has transformed the area. Scores of factories have appeared, producing everything from fibreglass to switch-gears. In January a new port terminal—built with financing from two Chinese logistics giants, COSCO and CK Hutchinson—began operations.
Ain Sokhna forms part of the wider Suez Canal Economic Zone, a network of industrial parks and ports that stretches north to the Mediterranean. Roughly half the investment it has attracted in recent years has come from China. The factory-building extends far beyond North Africa (see map on next page). From Saudi Arabia and Hungary to Brazil and Indonesia, Chinese industrial parks are springing up, along with supporting infrastructure.
The pace of this investment has accelerated sharply. In the past three years alone Chinese companies have spent more than $200bn building factories abroad (see chart). The character of their supply chains is also changing, in three ways. First, they are spread more widely, with large production nodes in nearly every region of the globe. Second, they have grown deeper, with many Chinese suppliers following manufacturers into new sites, replicating
the tight-knit ecosystems back home. Third, they are increasingly dominated by strategic industries, from electric vehicles and clean energy to data-centre gear. The consequence is that a rewiring of global manufacturing is under way.
A number of reasons explain why Chinese firms are making their wares in an expanding array of places. Weak consumer spending and fierce competition at home have encouraged them to venture into new markets. The tariffs introduced by the second Trump administration have also incentivised production in places that have been hit with less punitive levies than historic Chinese outposts such as Vietnam.
Various countries in the global south have dangled added incentives. Egypt, for instance, offers a “golden licence” that slashes red tape for big projects. Mohamed Eldib, a lawyer who helps Chinese firms set up operations in the country, says they are eager to “come in and make money” selling to Egypt’s 120m people while also using it as an export base.
The result is increasingly dispersed production footprints. Take JA Solar, Jinko-Solar and TrinaSolar, three Chinese manufacturers that began making solar panels in various South-East Asian countries a decade ago and are now setting up factories in the Gulf. New Chinese production hubs are gradually being woven into customers’
→ ALSO IN THIS SECTION
53 America Inc’s grip on governments
54 Bartleby: On AI and ROI
55 Meta’s blockbuster trial
55 Reddit’s woes
56 Canada’s oil boom
57 Schumpeter: Data-centre politics
52 Business
supply chains. Nordex, a German manufacturer of wind turbines, now purchases blades from a factory in Morocco that its Chinese supplier opened last year.
Europe has also emerged as a popular destination. Chinese firms' foreign direct investment in all-new "greenfield" projects on the continent rose by half in 2025, to a record €8.9bn ($10.1bn), according to Rhodium Group, a research firm, and MERICS, a think-tank. Hungary has attracted much attention. Serbia has also seen a growing Chinese presence. Linglong Tire, a carparts supplier, is among the Chinese manufacturers to have begun producing in the country. It recently announced it was expanding the capacity of its largely automated factory in the city of Zrenjanin, which supplies Western carmakers including Volkswagen and Ford.
At the same time China's global supply chains have deepened, with more upstream manufacturing taking place abroad. Gotion, a Chinese battery-maker, is constructing a gigafactory in an industrial zone 70km north-east of Morocco's capital, Rabat; a number of Chinese suppliers, including BTR, which makes anodes and cathodes, and Hailiang Group, which produces copper foil, are building factories a few hours' drive away to provide inputs. At an industrial zone on the outskirts of Cairo, where a Chinese manufacturer of home appliances has built a factory, an executive notes that the firm has likewise encouraged some of its suppliers to set up local plants.
Plenty of inputs are still shipped in. Machine tools are often imported from Chinese suppliers such as Yangli Group. So are components or materials that are unavailable (or much pricier) locally. Chinese exports of capital goods increased by 14% in the first half of 2026, year on year. Exports of intermediate goods rose by
Flying dragons
China, outflows of foreign direct investment into greenfield manufacturing, $bn
27%. With time, however, more of those inputs may be sourced from nearer by.
Meanwhile, China's logistics firms are helping its manufacturers to link their newly sprawling supply chains. These now operate or have invested in at least 132 foreign ports, from Greece to Sri Lanka, along with airports and rail lines, including one from Budapest to Belgrade completed earlier this year.
China Inc's overseas investment binge has focused heavily on a handful of strategic industries. In order to satisfy rising demand for electricity, countries in the global south in particular have become eager buyers of China's clean-energy technology. The vast solar farms being built in Egypt's deserts rely largely on Chinese equipment, a growing share of which is produced locally. Chinese EVs, now a common sight in cities from Rio de Janeiro to London, are increasingly being made in regional hubs nearer to customers.
The equipment required for artificial-intelligence data centres has been another focus. Thailand, for example, has emerged as a manufacturing centre for Chinese
makers of high-speed optical components, such as Zhongji InnoLight, whose products are used by America's cloud-computing giants around the world.
Chinese firms hoping to invest abroad have found they have the field to themselves, as Western rivals concentrate their factory-building in America to appease its protectionist government. An Egyptian official adds that, when Chinese businesses decide to build, they do so quickly.
Their foreign expansion is not without obstacles. The volatile tariffs imposed by America have resulted in some projects being cancelled or cut back. On August 13th the White House published a report titled "The Great Transshipment Scam" that called for sweeping restrictions on imports with even a whiff of Chinese involvement (see Finance & economics section). At the same time governments in places such as Brazil and Turkey have been imposing local-content requirements, which result in more of the value being added in their countries but make manufacturing there less attractive. Currency fluctuations and high borrowing costs in emerging markets complicate matters further.
Chain reaction
China's government has also been making life hard for the country's globe-trotting businesses. Last month a new package of outbound-investment regulations came into effect which, among other things, restrict the transfer of technology and data abroad and introduce a national-security review process.
So far the overall share of global manufacturing taking place in China has shown no indication of decline. Yet with ever more of its companies investing abroad, that may soon change. Even then, however, the world will continue to rely on Chinese goods—wherever they are made. ■
Business 53
Government procurement
US and them
America Inc has a tight grip on allied governments
“I MISS NOTHING,” says Dirk Schrödter, the digitisation minister for the German state of Schleswig-Holstein. Over a video-call using OpenTalk, a German alternative to Microsoft’s Teams, he explains how in the past two years he has moved some 30,000 of the state’s civil servants from the American company’s collaboration and productivity tools to open-source alternatives. He has now begun the process of shifting staff from Windows to Linux, an open-source operating system. The move has attracted the attention of public officials elsewhere who are also keen to cut ties with America’s tech giants. “Every week we have questions from other states, other cities and governments inside and outside of Europe,” he says.
The growing interest in open-source software reflects a new geopolitical reality. America has long used its commercial prowess to hurt enemies, as when imposing sanctions on Iran and Russia. But now even allies worry that it could cut off their access to critical technologies—or at least threaten to during, say, a trade negotiation.
Governments abroad are examining the extent to which their ability to operate depends on American suppliers. Fears were stoked last May when Karim Khan, then the International Criminal Court’s chief prosecutor, lost access to his Microsoft email account after President Donald Trump brought sanctions against the court over its issuing of an arrest warrant for Binyamin Netanyahu, Israel’s prime minister. (Microsoft says it did not cut its service.) Another jolt came this June when Mr Trump forced Anthropic, an artificial-intelligence lab, to temporarily cut off access to its latest models in foreign countries. America’s allies also took note when it twice paused its supply of weapons to Ukraine last year.
American firms win a small share of the overall contracts tendered by governments abroad. But their products and services often underpin critical government functions. Various efforts are thus under way to reduce reliance on American suppliers and nurture domestic alternatives. In many cases, however, doing so will be enormously difficult.
Europe in particular has emerged as a centre for efforts to flush America Inc out of government supply chains. France’s national government is planning to ditch Teams and wants to move some computers
Patron states
Selected countries, number of government contracts won by American firms, % of total
Source: TenderAlpha
to Linux. Local governments, including the cities of Reus in Spain and Aarhus in Denmark, have turned to European cloud providers such as Nextcloud and Hetzner, both from Germany. In June the European Commission unveiled a plan to boost the continent’s “technological sovereignty” that, among other things, aims to shift the processing of sensitive government data to such providers. Last year Spain cancelled an order for F-35 fighter jets, supplied by America’s Lockheed Martin. British politicians are urging the prime minister to implement a break clause in a contract between Palantir, an American technology firm, and the National Health Service (see Britain section).
The Economist’s estimates suggest that, at an aggregate level, American companies account for a modest share of public spending abroad. We calculate that, of the $25trn in sales generated last year by listed American companies, perhaps $500bn (or 2%) came from foreign governments. That is equivalent to roughly 7% of government procurement in OECD countries other than
Air superiority
America, which account for an overwhelming majority of the spending.
Some countries are more dependent than others. Measured by the number of government contracts won by American firms last year, Australia (6% of contracts) and Britain (4%) are more reliant than France (2%) and Germany (1%), according to figures from TenderAlpha, a data provider (see chart).
Yet American firms play an outsize role in critical areas of government. We estimate that roughly two-fifths of the business that American companies generate from foreign governments is in information technology and defence. (Pharmaceuticals and medical equipment, purchased by public health systems around the world, make up another large chunk.)
Often that is because there are few alternatives. Alphabet, Amazon and Microsoft control two-thirds of the global cloud-computing market, according to Synergy Research, a firm of analysts. Forrester, another research group, reckons that American providers account for between seven and nine of the ten largest vendors for most segments of enterprise software.
Those substitutes that do exist are frequently inferior. Sometimes the loss is negligible: on OpenTalk users cannot send emojis flying across the screen. But other shortcomings are more serious. LibreOffice, the open-source productivity-software package for which many governments are opting, lacks the AI features offered by Microsoft—a weakness that will only become more pronounced as the technology advances. OVItcloud, a French cloud-computing provider that is Europe’s biggest such firm, generates about one-hundredth the revenue of Amazon Web Services, making it difficult to compete on price or keep up on innovation. Similarly, governments that want the most advanced air-defence system must turn to Lockheed and RTX, another American armsmaker, which together produce the Patriot.
Switching costs can be hefty, too. Lockheed supplies F-35s to governments from Norway to Belgium. Those countries could instead opt for a European jet, such as the Rafale or Typhoon, even if they are less advanced. But that would require a vast operational overhaul, including retraining pilots and support crews and replacing weapons inventories—a tough sell when huge sums have already been spent on Lockheed jets.
Or consider patient-record systems in hospitals. The hefty cost of swapping to a new vendor is partly why Norway’s parliament voted in June to keep using Epic, an American provider of health-care data. ►
Correction: Last week, in “Silicon showdown”, we wrote that Nvidia makes its chips. In fact, it only designs them. Sorry.
54 Business
bases, despite dissatisfaction with the software among doctors.
The dependence of local companies on American suppliers further complicates matters. Roy Illsley of Omdia, another research firm, points out that much of the software OVHcloud runs on comes from VMware, which is owned by Broadcom, an American company.
Governments that do shun America's suppliers risk irritating its tempestuous president, which could make the super-
power an even less dependable ally. Pieter Wezeman of SIPRI, a think-tank, notes that in July Denmark plumped for maritime-patrol aircraft made by Boeing, an American aerospace giant, over European alternatives, which he reads as a signal that the country does not want to alienate America despite clashes over Greenland.
For consolation, America's allies should remember that there are plenty of dependencies in the other direction. Much of America's federal government runs on
software provided by Germany's SAP. And American firms selling to foreign governments often depend on global supply chains. The rear fuselage of every F-35 is built by BAE Systems, a British weapons manufacturer, in Lancashire. The most advanced chips used in cloud computing are made in facilities operated by TSMC, a Taiwanese manufacturer, using gear from ASML, a Dutch one. America's allies might usefully focus on making themselves even more indispensable. ■
BARTLEBY
Returns policy
How to measure the benefits of artificial intelligence
IN THE BEGINNING there was AI, and bosses everywhere lost their minds and said you must all use this technology no matter whether it is useful and no matter how much it costs. And they introduced AI leaderboards and came up with stupid words like "toxenmaxxing". And lo, usage did indeed rise. And then the bosses remembered some very basic concepts like "budgets" and realised that this might not be such a great idea after all. And then a different question rang through the boardrooms, and this question was about returns, and it was harder to answer.
This potted Genesis of AI adoption is not entirely fair. As long as organisations are trying to encourage usage, it can still make sense to look at things like employees' token consumption. "I would expect software engineers to spend more money than a finance analyst," says the boss of one software firm. "But anyone within finance, it shouldn't be zero."
But usage (also known as input) measures don't tell you whether any value is being created. And even when AI is producing benefits, points out Mika Ruokonen of LUT University in Finland, lots of them accrue quietly to individuals in the form of scattered time savings. If an extra hour or two of an employee's time has been freed up to shop on Vinted, someone has benefited (mainly Vinted) but it isn't your company (unless you work at Vinted).
Shifting the focus to returns on AI investment means concentrating more on outcomes-based measures, such as team productivity or customer satisfaction. But these require careful thought. Scientists, for example, are using AI to produce more papers (yay!) of dubious merit (boo!). A recent study by Tulsi Suchak of the University of Surrey and
her co-authors found that papers derived from a health and nutrition data set ticked along at an average of four a year between 2014 and 2021; in the first nine months of 2024, there were 190.
So outcomes-based measures have to be designed to capture both positive and negative effects of AI. At Google, for example, engineers have long been measured on three dimensions. Speed is one: how long does a code review take, say. Ease is another: how much friction is there in the system for things like onboarding new developers. The last, critically, is quality: "I don't want to just go fast and make it easy to ship terrible software," says Richard Seroter of Google Cloud.
Complicating matters further is the fact that it often takes a while for the benefits of AI to show up. People have to spend time learning how to use the technology. Increased output can create bottlenecks elsewhere. A paper published last year by Kristina McElheran of the University of Toronto and her co-authors looked at AI adoption among American manufacturing firms, and confirmed the
existence of a "J-curve" in which productivity dips before it leads to an improvement. This effect is especially pronounced at established firms and seems to be partly explained by changes to management practices that had previously helped firms to keep tabs on performance. Any kind of calculation about projected returns has to take this initial period of disruption into account.
Bosses also have to take a view on something much more fundamental: what kind of return are they interested in? One way to show financial returns on AI is to seize on time savings as an excuse to cut headcount. For some jobs and in some circumstances, that might make sense. But chainsaws and employee morale go together about as well as salt and slugs. And people who are no longer there cannot be redeployed in more productive ways. A better option is to calculate the amount of money that AI saves on expanding headcount. (Working out how much extra revenue can be attributed to the technology is even more of an art; at the very least, you need good baseline data and a culture of A/B testing to isolate its effects.)
On top of everything, there is wild uncertainty about where AI is heading. That is an argument for a third set of measures, to go alongside ones on inputs and outcomes. These measures, which Mr Seroter calls "organisation-based", are more geared towards capturing how well change is being managed. They could include data on levels of employee satisfaction with the way AI has been implemented, or progress towards building in-house expertise in areas that AI will not automate. A one-eyed focus on usage ignores the importance of returns. A rigid focus on returns ignores the need to keep learning.
Business 55
Meta
Social reckoning
LOS ANGELES
A blockbuster trial draws parallels to big tobacco
IN EARLY OCTOBER a sequel to "The Social Network", a film about Mark Zuckerberg's founding of Facebook, is scheduled for release. Called "The Social Reckoning", it will dramatise the role of whistleblowers who alleged that the tech giant hid the fact that its products were harmful. Some of the characters involved will take part in a no less vivid drama involving Facebook, Instagram and their parent company, Meta, in a federal courtroom in Oakland, California. Oral arguments started on August 18th. Reckoning or not, the trial may be a blockbuster.
California, Colorado, Kentucky and New Jersey are leading a group of 29 states that are suing Meta over allegations that its platforms developed features harmful to children, and that it violated their privacy. In opening arguments, Megan O'Neill, a deputy attorney-general for California, said that Meta's business model was to "hook the users, hold them for as long as they can, harvest their data and then hide the truth from the public." Rob Bonta, California's attorney-general, has described the case as Meta's "tobacco moment", likening the firm's alleged deceptions to those that led to huge payouts by tobacco firms in the 1990s.
Paul Schmidt, a lawyer for Meta, told jurors that the company took its responsibility to young users seriously, adding that there was no clear link between youngsters' social-media use and ill health. Meta is expected to fight efforts by the plaintiffs to turn the trial into a referendum on the impact of social media on children, focusing instead on the narrow legal issues in dispute. Its lawyers have asked why it is being singled out, when youngsters' use of YouTube, owned by Google, and TikTok, created by China's ByteDance, is even higher than it is of Instagram.
A lot is at stake. In a similar case brought by the government of New Mexico, a judge this month ordered Meta to pay almost $95m in penalties. The four states acting as plaintiffs in Oakland appear to be seeking more than 200 times as much if Meta loses the case, putting their estimate at $193bn—roughly equivalent to the firm's revenue in 2025. If that were a benchmark for other states, it would devastate Meta.
Whether it comes to that is another matter, though. In emotive language, Ms O'Neill alleged that Meta particularly targeted young children. She frequently used
the word "hooked". But unlike a trial in Los Angeles in March, in which Meta and Google, were obliged to pay damages to a 20-year-old harmed by spending much of her life on social media, the Oakland case is not focused on addiction.
Instead it is what one of the attorneys-general calls "the largest consumer-protection lawsuit in American history". Vincent Joralemon, of the University of California's Berkeley Centre for Law & Technology, says the big question is whether or not Meta deceived the public about the impact of its products. That is similar to the tobacco cases. It also means whistleblower testimony is likely to be more relevant than thorny psychological debates about addiction, reckons Mr Joralemon.
Meta's investors tend to think that the burden of proof on the deception charges is "quite high", says Gil Luria of D.A. Davidson, an investment bank. Litigation risk has weighed on Meta's valuation, contributing to an earnings ratio that is lower than those of its big-tech peers. But rather than fearing a giant payout, shareholders worry more about remedial changes that the prosecutors may seek if Meta loses in court, according to Mr Luria. Those include changes to features Meta, Google and TikTok use to keep user engagement high, such as "infinite scroll", which continuously loads new posts. "If we didn't have infinite scroll, we wouldn't have as many ads," Mr Luria says.
Even if Meta loses, it has scope for appeal. It could, for instance, do so on the grounds that it is protected by Section 230 of the Communications Act, which safeguards internet platforms from legal liability for content posted by third parties. It sought to use that argument to halt the Oakland trial before it started, but the judge ruled that the appeal was premature.
A marquee trial for Mark
Public concern in America about the impact of social media on children is high, and being dragged through the courts may not help Meta's image with parents. In order to head off similar worries about AI, on August 18th OpenAI launched ChatGPT for Kids, which strengthens guardrails for under 18-year-olds.
For all the attention the Meta trial will receive, court verdicts can have less of an impact than the headlines suggest. The tobacco industry agreed in a settlement of 1998 to shell out more than $200bn over 25 years. But such outcomes have a "disappointing history" in America, says Matthew Lawrence of the Emory University School of Law. "We can mitigate the harms of industries exploiting addiction, but we have not yet had great success in finding ways to eliminate the harms." ■
Good bot, bad bot
Can "the internet's front page" survive in the AI era?
EARLIER THIS summer "Backrooms", a horror flick, became one of the most successful independent films ever made, grossing $393m from a $10m budget. Its source material was unusual: a collection of images of unnervingly empty rooms that spread on Reddit, an internet forum. Hollywood producers have since taken to trawling the site for inspiration.
Reddit has had a difficult time of late. Its share price has dropped by about a third this year. Strong quarterly earnings last month failed to cheer shareholders, who fret that the company will suffer as users turn to artificial-intelligence chatbots for information, rather than search engines such as Google that direct them to other sites. For now Reddit is attempting to cajole chatbot-makers into paying more for its data, both through deals and lawsuits. But eventually it hopes bot-weary users will return to one of the internet's last refuges of human-generated content.
Reddit's business model lies somewhere between Instagram (which profits off user-generated content) and Wikipedia (which relies on an army of volunteer moderators). That balance has sometimes been tricky to manage. In 2023 the volunteers who moderate its "subreddits" went on a multi-day strike after the company said it would charge for access to its application programming interface, on which independent developers who build Reddit-based apps rely. The platform, which stuck with the policy, may have had larger clients ▶
56 Business
► in mind. Soon afterwards it struck deals with Google and OpenAI, maker of ChatGPT, which reportedly agreed to pay $60m and $70m a year, respectively, to train their models on Reddit's data.
They may have got themselves a bargain. Reddit has become the most-cited source by many of the leading American chatbots, according to an analysis the company commissioned from Profound, a research firm. One particularly valuable use of its data is for generating product recommendations. About 40% of conversations on the site relate to commercial activity including shopping, says Jen Wong, Reddit's operations chief. On subreddits such as t/BuyItForLife, users debate the quality of products—and try to weed out sponsored or bot-generated content—making the results a rare source of high-quality reviews. AI models rely heavily on such content to generate their own recommendations, which chatbot-makers hope to turn into a source of ad revenue.
Google's relationship with Reddit in particular has become less of a two-way exchange. For years Google's ranked search results have been an important source of traffic for the chat site. But as the tech giant shifts users' focus to its AI overviews, referrals have dropped precipitously. Reddit's stock took its largest-ever single-day plunge last month after it said that 'choppy' search referrals had hurt engagement. Last quarter the amount of time users spent on Reddit's app each day was down by 8% from a year earlier, compared with gains of 6% and 4% for Instagram and TikTok, respectively, according to Sensor Tower, a data provider.
One option for Reddit is to bargain for a better deal with AI providers. It has been negotiating new agreements with Google and others. If that does not work, lawsuits may. Reddit has already sued Anthropic, another model-maker, and Perplexity, an AI search engine. Both suits allege that the companies scraped Reddit's data illegally. (Anthropic and Perplexity have said they did not break the law.) Victory in court—or lucrative settlements—would give the chat site greater leverage in negotiations over licensing its data.
Ultimately, though, Reddit hopes to lessen its reliance on other tech firms. Traffic from search is 'not where our business lives', said Steve Huffman, its boss, in last month's earnings call. Instead the company hopes to encourage more users to visit its site directly, and spend more time browsing once there. That would further boost its ad revenue, which grew by 64% in the second quarter, year on year. Ms Wong is optimistic that the (mostly) human content on Reddit will appeal to users as they grow tired of interacting with chatbots. Aspiring film-makers need not fear being starved of material just yet. ■
Canadian crude
On shifting oil sands
Will the latest boom endure?
'PLEASE GOD, give me one more oil boom—and I promise not to piss it all away next time' reads a popular bumper sticker in Canada's oil patch. Recent results from the country's oil producers show that plea has been answered. Oil prices have jumped as the passage of tankers through the Strait of Hormuz has been choked off. Last month Cenovus Energy reported that operating profits in its most recent quarter had nearly tripled compared with a year before. Canadian Natural Resources posted its best-ever quarterly adjusted operating profit. Moreover, the easing of political headwinds in Canada may improve the chances of the boom lasting.
Canada has the world's fourth-largest proven reserves, with 97% found in the oil sands in a remote corner of the landlocked province of Alberta. Here the bitumen, mixed with sand and clay, is almost solid, and is mined rather than tapped from the ground. It is costly and dirty to extract and process. Despite the drawbacks, high prices combined with a weakening of the Canadian dollar mean that 'it hasn't been this good to be an oil-sands producer in a very long time,' says Andrew Leach of the University of Alberta.
Getting more oil to market is also set to become easier. Justin Trudeau, Canada's former prime minister, was tepid in his support for new pipelines, fearing the consequences for the climate. In 2018 regional political opposition led Kinder Morgan, an American firm, to suspend work on
a conduit from Alberta to the west coast in British Columbia. That forced the national government to step in and complete it. Oil eventually started flowing in 2024.
Lately the political mood has shifted. A trade war with America, the destination for 90% of Canadian oil, has shifted the federal government's focus to reducing dependence on its southern neighbour. That includes support for building new infrastructure to transport oil (and gas) to ports on Canada's west coast, from where it can be shipped to Asia. Mark Carney, Canada's current prime minister, is intent on slashing the red tape that holds up projects, including streamlining environmental permitting and consultations with indigenous communities. The speedier process will apply to a new pipeline under development to British Columbia, announced last month. Lisa Baiton of the Canadian Association of Petroleum Producers, an industry group, has spoken of a 'generational opportunity' to rally political and public support behind new oil projects.
The promise of a federal government more friendly to pipes may not be enough, however. Enbridge, an energy-infrastructure firm, has paused construction of a pipeline to the east, citing a lack of commitments by producers to provide enough oil to make it worthwhile.
The unique challenges of the oil sands provide an explanation. Investment in new sites has been virtually non-existent since the oil price crashed in 2014 as demand weakened while America's shale oil gushed and OPEC opened the taps. 'It cannot be overstated how much that changed the industry,' says Kent Fellows of the University of Calgary. Moreover, projects in the oil sands take far longer to develop than conventional wells. They require enormous amounts of capital that could be tied up for five to ten years before profits flow. Uncertainty over the future price of oil—and the danger that a change of government might lead to the reinstatement of tighter regulations—make for risky bets.
Nevertheless, at the Calgary Stampede, an annual rodeo festival held in July that serves as the industry's unofficial barometer, parties hosted by oil companies were back to their riotous best after quietening down in recent years. Their sponsorship of rowdy chuckwagon races broke records. That is as good a sign as any that another boom is in full swing. ■
Watch The Insider
This week our editor-in-chief interviews historian and author Yuval Noah Harari. To watch the video, go to economist.com/insider
Business 57
Shadow boxing
The war on data centres is a bit fake
IN FEBRUARY, SHORTLY after making his feature-film debut as a pen magnate in “Marty Supreme”, Kevin O’Leary announced that he would build one of the world’s biggest data-centre projects in Utah. The market shrugged. What did Mr O’Leary, who is best known for presenting “Shark Tank”, know about data centres? How would it be paid for? Or powered? One analyst dismissed the project as an example of “vapourware gigawatts”.
But a joke in the marketplace became dynamite in the town square. In June J. Stuart Adams, president of the Utah Senate, who had initially supported Mr O’Leary’s venture, demanded that its size be reduced by three-quarters. Mr O’Leary accused his critics of working for China (he later withdrew that), but eventually agreed to scale back his dreams. Even then Mr Adams, a long-serving Republican lawmaker, along with a pair of county commissioners, lost primary elections because of the scandal.
Data centres are a nightmare of political economy. Voters see them as ugly, dirty, noisy and power-hungry bastions of arrogant Silicon Valley companies. And—thanks to ChatGPT—it has never been easier to pen a letter to the planning department saying so. Yet the public has an awful lot riding on these windowless sheds, for the stock market is a one-way bet on them being built. Data centres are particularly disorientating for Democrats, since unions tend to support the projects. If artificial intelligence is a class war, it pits a coalition of billionaires and electricians against the office droids both despise.
Wildcatters like Mr O’Leary are a soft target for politicians who want to be seen as tough on data centres but still want them to be built. On August 18th Josh Shapiro, the governor of Pennsylvania, derided the developers “scaring our communities” and signed an executive order he said would squash “speculative proposals” that had little chance of being built in the first place. Of the more than 100 projects announced in the state, less than a fifth had applied for the permits needed to develop them, he said.
Imaginary scalps are easy to claim. Silicon Valley giants are the biggest backers of data centres. But beside them is a long line of property developers, private-capital firms, “neo-cloud” operators and crypto-miners, followed by international chancers, national
rogues and local spies. Most of their plans were destined to remain press releases. Analysts at Bernstein, a broker, attempted to quantify the credibility of each watt of announced capacity. Plans by Google and Amazon are taken at their word. Those by CoreWeave and Nebius, two large neo-clouds, are discounted by around half. Mr O’Leary is ignored.
According to their model, another fantasist is Vermaland, a land bank that announced a gargantuan 3GW facility in Arizona last year. In May the project was reportedly cut by four-fifths after facing local opposition—a shadow-boxing knockout. Little information is available about a 1GW project in south Florida that was supposedly shelved in February, except that the land had been bought for $15m in 2024 by an obscure investment firm. When a data-centre plan that would never have materialised is cut down, it often makes a very large sound.
Politicians are eager to polish their anti-data-centre credentials ahead of the midterm elections in November. The resulting policies may not affect the boom much. In July New York’s governor announced a one-year moratorium on big projects (including those not connected to its electricity grid). But the moratorium does not apply to data centres that have already been approved, and New York is hardly important to the build-out. Virginia imposed a $600m electricity tax on data centres to great fanfare, but left in place a sales-tax exemption three times as big. Mr Shapiro’s new rules will require local authorities to approve projects before the state does, clearing up his desk until after the election.
If the data-centre build-out has a natural home it is Texas. The state is vast, energy-rich and home to plenty of the lawyers and engineers needed to handle big infrastructure developments. Its electricity-grid operator has been flooded with an astonishing 474GW of requests to connect to the grid (which currently faces peak demand of around 90GW). On August 3rd Greg Abbott, governor of Texas, announced an “audit” of data-centre projects, which is expected to last, conveniently, until just after the midterms. Tellingly, some of the biggest data-centre operators have publicly supported his plans.
More political controversy probably means more delays for projects that are not yet in motion. But the share of the data-centre pipeline that has been delayed or cancelled has remained consistent at roughly 10% during the past year, despite the explosion in the number of plans and political opposition to them. Thus far constraints imposed by the market, such as the availability of chips and workers, have worried developers far more than those imposed by government.
Could the slightly fake war against data centres in fact be a “phoney” one, poised to turn real? It is hard to imagine. Politicians are rarely willing to halt projects that have already broken ground. As for the new ones, developers can to some extent vote with their feet. So far that has meant moving west. It could also involve building more on federal land. Eventually, it might involve more construction abroad—or, if Elon Musk is to be believed, in space.
Stoppable force, movable object
It is often said that China has an advantage in the AI race because it can build big things by force. Yet America has the power of willing things into existence by exaggeration. The grand plans of developers will be watered down. Politicians will also overstate their opposition. Often they will cancel each other out. Meanwhile, the searing pace of construction will continue. ■
58
Finance & economics
Daron Acemoglu
The rise of the world's most influential economist
Daron Acemoglu is prolific and revered. Yet a lot of his work is oddly unconvincing
BY THE TIME the call came from Stockholm, it felt overdue. In 2024 Daron Acemoglu of the Massachusetts Institute of Technology shared the Nobel prize in economics with Simon Johnson and James Robinson for work on how institutions shape prosperity. Mr Acemoglu is only 58 years old, but for some time he has been the discipline's colossus. He has written seven books, including a new one about democracy and the economic impact of artificial intelligence (see Culture section),
not counting a couple of popular textbooks. His hundreds of papers, more than ten of them published this year, are often stuffed with complex mathematics. He competes with Andrei Shleifer of Harvard University for top spot on a citation ranking of economists produced by IDEAS/RePEc, a research database. His views carry weight with other works and the media.
Give economists a few drinks, however, and some will venture their true opinions about this giant. "Much of his theoretical
→ ALSO IN THIS SECTION
59 A warning from the bond markets
60 The truth about Chinese jobs
61 Trash talk on trade
62 Buttonwood: Lying indices
63 Free Exchange: Free passage
work is useful, but he uses his models to inform populist policies that have been tried before and failed," blasts one well-known economist. Some commentators do not require Dutch courage. "I've been yelling about Acemoglu for literally a decade," Noah Smith, an economics blogger, has said, in response to a flurry of online criticism of Mr Acemoglu's work.
No one accuses Mr Acemoglu of truly poor research, let alone academic malpractice. "He is obviously a genius," says another economist, who adds that he is generous to PhD students and junior colleagues. The question is whether his reputation at the very apex of the economics profession—and the intellectual influence that goes with it—is justified by his scholarship.
Start with his empirical methods. His Nobel rests heavily on a paper released in 2001 with Messrs Johnson and Robinson, which researchers have cited more than 23,000 times. It seeks to explain why some countries are rich and others poor. In places where European settlers died in droves, perhaps because of disease, colonisers built "extractive" institutions, with power and resources concentrated in the hands of an elite few, to take out as many resources as possible. But in more agreeable climates, settlers survived. Colonists built more inclusive institutions, with proper roads, schools and medical services. Those institutions persisted, with the effect that places with lower settler mortality in colonial times are richer today.
Economists now take these results seriously but not literally. The paper encouraged them to think more deeply about institutions, but they have also picked it apart. The Nobel committee's report in 2024 conceded that the mortality data are "sometimes sketchy". In 2012 David Albouy of the University of Illinois Urbana-Champaign showed that some countries were assigned mortality rates borrowed from other countries, among other adjustments. Correct for these, he found, and the paper's estimates become unreliable. Last year Martin Buchner of RWI-Essen, a research institute, and colleagues reported that experts they surveyed were somewhat more likely to side with Mr Albouy.
Mr Acemoglu says that "these discus- ▶
Finance & economics 59
sions and some critiques are hugely valuable. But he counters that Mr Albouy reaches his conclusion by omitting half the data from the original sample, including on important countries like America, Canada and Australia. It is this combination, along with some statistical choices, that introduces the unreliability, he says. He adds that if he were redoing the paper today, he would make “a number of changes, including in some of the estimation details”—though not to the mortality data.
Most recently, researchers have examined a paper that Mr Acemoglu released in June, with co-authors, which argues that lower birth rates are associated with faster growth in GDP per working-age adult. A striking finding, it seems to suggest that places with falling populations, like Japan, do not have much to worry about. Yet other economists, including Jesús Fernández-Villaverde of the University of Pennsylvania, question whether historical relationships are really useful guides to the present, when many countries’ birth rates are collapsing. It is a fair point—and Mr Acemoglu and his co-authors raise it in their conclusion. But it can get lost and can seem unsophisticated when set against Mr Acemoglu’s reams of maths.
A determined critic can find holes in almost anyone’s work—and Mr Acemoglu is the biggest target in economics. Yet fame-induced scrutiny cannot explain a second problem: that his ideas can seem underwhelming. His public commentary, especially on American politics, is unadventurous. His “unified theory of Trump”, for instance, asserts that the president operates “by elevating executive power and destroying constraining institutions and norms”. Well, indeed.
More worryingly, his grand thesis may not reveal much. Nations prosper when institutions are good, and stagnate when they are bad. True. But what, exactly, is an institution? Rules, norms, enforcement, culture—everything, really. Where do institutions come from? “Critical junctures” and “institutional drift”.
Reviewing one of Mr Acemoglu’s books in 2024, Tyler Cowen of George Mason University noted that the institutional changes it describes seem only to come from other institutional changes, making the core argument regress for ever—turtles all the way down. Duncan Green of the London School of Economics has said that the framework works mainly in hindsight. Francis Fukuyama of Stanford has argued that the book waves away the example of China, which has had blistering economic growth alongside institutions that can plausibly be described as extractive.
Mr Acemoglu retorts that the book devotes two chapters partly or mostly to China. As for the turtles: “To understand institutional change from an institutional
perspective you must consider the specific institutions in the past and the historical events...that influence them,” he says. “I don’t think this is tautological at all.”
To the extent that Mr Acemoglu breaks from received wisdom, it is mainly to espouse remarkably glum views about technology. “Power and Progress”, written with Mr Johnson and published in 2023, treats a thousand years of innovation largely as a story of elite capture and unintended consequences. The cotton gin was the handmaid of slavery; the Haber-Bosch process created weapons. Ordinary people apparently gained only when they could steer capitalists to consider their interests. The book plays down a more compelling claim: that technological progress in itself has made the average person many times richer than in pre-industrial times.
Mr Acemoglu’s arguments about AI are also pessimistic to the point of losing credibility. In a paper published in 2024 he calculates that AI will boost American productivity by only a touch over ten years. The paper reaches this conclusion by ignoring the potentially transformative effect of new AI products. “He leaves out entirely in that analysis the possibility that we will have more rapid scientific progress, more rapid social-scientific progress, or better decision-making because of artificial intelligence,” according to Lawrence Summers, president emeritus of Harvard (and a former treasury secretary).
“This is a valid criticism,” says Mr Acemoglu. “I did not foresee agentic AI and it was not incorporated in my predictions.” He also notes that AI models may be more useful in scientific research than he imagined. But he stands by his methods of estimating the productivity effects of AI, as well as injecting “some realism into the claims about huge coming productivity improvements”.
These academic disputes could end up mattering a great deal. Mr Acemoglu may not think AI is about to turbocharge productivity, but he does worry about its effects on democracy and jobs. His new book focuses on its potential social harm rather than its potential benefits. His proposal to limit that harm, “pro-worker AI” that would augment the value of human labour rather than replacing it, sounds terrific—but also obvious.
Mr Acemoglu’s influence in the AI debate is clear in a recent statement, signed by dozens of prominent economists, which argues that “we must act now” to “steer AI in a direction that complements humans and benefits society”. But who are “we”? And who is to decide what sort of AI does or does not complement humans? Even economists who signed the petition say they are not entirely sure. So great is Mr Acemoglu’s stardom that it can sometimes blind the critical faculties. ■
Government bonds
Yielding
NEW YORK
Bond investors are global markets’ party-poopers
BOND MARKETS are unsettled, and so, in turn, is America’s government. On August 19th the Treasury said that from next month it would increase its own purchases of longer-dated debt. This reflects “increasing administration unease” about yields, believe analysts at Deutsche Bank.
Officials have good reason to worry about rising borrowing costs. The yield on ten-year Treasuries reached its highest since January 2025 on August 18th. Scarier still, that on 30-year bonds briefly passed 5.3%, its highest since 2007. And America is not alone. Bondholders are demanding more from governments across much of the rich world (see chart). Yields on British, French and German long-dated bonds have all reached levels not seen in more than a decade. Japanese 30-year yields, long the lowest in big economies, are close to an all-time high. The rout eased somewhat on August 19th. But yields are unlikely to fall much soon. What is amiss?
For one thing, unlike equity investors, bond traders appear to be reading the news. The Strait of Hormuz remains largely shut and looks likely to stay that way for a while. Fuel prices in America—particularly for diesel, on which much of commercial haulage relies—have soared in turn. The fund managers surveyed monthly by Bank of America, most of whom are heavily invested in stocks, are sanguine. On average, they expect Brent crude, the global oil benchmark, to trade at $76 a barrel by the end of the year, only slightly up from before the start of the war. Bond markets seem less sure.
Accordingly, they are pricing in continued inflation. On August 19th Britain re- ▶
Looking steep
30-year government-bond yields, percentage-point increase since Mar 9th 2020*
Source: LSEG Workplace “Black Monday” stock-market trough
60 Finance & economics
► ported consumer-price inflation of 2.9% for the year to July, up from 2.6% in June, as higher energy prices bit. In America, core inflation (which excludes food and energy) eased last month. But bond traders seem to have pared back bets on future rate rises after recent remarks by the Federal Reserve's new chairman, Kevin Warsh.
Another recent concern is that government debt-issuers have new competition. In recent months large tech firms have sold some $75bn-worth of bonds to fund investments in data centres for artificial intelligence. The spree has already pushed their combined debt issuance to nearly twice last year's total, estimates Goldman Sachs, a bank. All sorts of businesses are getting in on the bond bonanza. Some 40% of large-scale debt issuance (ie, more than $10bn) by investment-grade issuers has come from outside tech, Goldman calculates. So bond-buyers have a surfeit of investment-grade options from which to choose. Many believe the AI boom will push up interest rates by increasing competition for capital.
Still, the biggest reason for creeping yields is of long standing: concerns over government debts and deficits. On August 19th America's Treasury said federal debt had passed $40mn (130% of last year's GDP) for the first time. The government's deficit is around 6% of GDP. Such worries also show up in differences between countries. The spread between French and German ten-year yields has reached its widest since 2012. Bond markets now expect a higher yield for Japanese debt than for Chinese debt, reversing the conventional order. In both France and Japan investors assess that politicians lack the will to trim spending or raise taxes meaningfully.
Governments have few other good options. At the Fed's meeting in June members of its rate-setting committee were briefed on how ownership of Treasuries has shifted from "relatively price-insensitive official-sector holders to more price-sensitive private investors". That is likely to increase the premium bondholders expect for long-term debt. In response to such pressures America, Britain and Japan have increased their sales of shorter-term bonds with lower yields. But as a result their debt stocks will roll over more often, raising the risk that such moments coincide with high interest rates.
A second unenviable option is for central banks to buy back more debt. But many of them had hoped to shrink their balance-sheets, not to expand them. They may therefore be reluctant to rely too heavily on bond purchases as a tool to reduce yields. All this means that bond markets are likely to remain wary. Interventions such as the Treasury's may help for the time being. They are unlikely to placate buyers for long. ■
Unemployment in China
Goodhart's law with Chinese characteristics
However you measure it, the job market is weak
STEERING AN ECONOMY has never been easy. In a candid presentation in 1975 Charles Goodhart, then of the Bank of England, shared some of Britain's "unhappy" experiences after sterling lost its link to the dollar four years before. Instead of defending an exchange rate, the bank had hoped to use interest rates to control the supply of money, variously measured. But the monetary authorities had ignored what the presenter playfully called Goodhart's law: "Any observed statistical regularity will tend to collapse once pressure is placed upon it for control purposes."
Others have put this more snappily: when a measure becomes a target, it ceases to be a good measure. The law poses particular difficulties for China, which has many economic targets and thus few good measures. In 2018, for example, it began publishing a new monthly gauge of urban unemployment, based, like similar statistics elsewhere, on surveys of the population. But unlike other countries' figures, China's measure has remained remarkably stable. July's rate of 5.2%, published on August 17th, was typical. It was less than 6% but at least 5% for the 99th time in the past 115 months. Even during the covid-19 pandemic, when American unemployment climbed well into double digits, China's measure never exceeded 6.2%.
One reason may be that the survey covers only urban workers. The flow of mi-
grants into and out of China's cities gives the urban labour market a homoeostatic quality. When job creation is strong and the unemployment rate is set to fall, more of them turn up, expanding the denominator. Conversely, when the labour market is weak, the jobless seek work in the countryside, dropping out of the survey.
There may be a second reason for the measure's stasis: the pressure placed on it for control purposes. Since 2018 the government has set an annual target for surveyed urban unemployment, usually around 5.5%, rising to 6% in 2020. That gives officials an incentive to keep the rate around that figure, by hook or by crook. A "possibility, which cannot be proved but is difficult to dismiss, is that the surveyed unemployment rate is being managed in some fashion to reduce volatility", argues Andrew Batson of Gavekal Dragonomics, a consultancy, in a recent note.
The stability of the unemployment rate is thus a false comfort. It makes it needlessly difficult to interpret the biggest job market in the world. Some economists suspect that China's export boom is not generating equally strong manufacturing employment. Others think artificial intelligence may be killing some entry-level white-collar jobs. Another worry is that wage growth is not strong enough to stop the economy falling back into deflation. All these suspicions would be easier to ►
Stably employed by any measure
Finance & economics 61
An unpalatable measure
China, urban unemployment, m
Sources: National Bureau of Statistics; Haver Analytics
► verify or dismiss if China's unemployment figures were as fluid as its economy.
Instead, economists must rely on a variety of more or less unsatisfactory proxies. Mr Batson has spotted a new one—or rather, a revived one. He advises looking again at an older gauge, which the modern survey-based figure was supposed to supersede. This measure, which has been collected since 1978, counts only those who have registered as unemployed with local authorities, a first step towards applying for jobless benefits. For decades, this was also a government target, and also uncannily stable. With the arrival of surveyed unemployment, it fell into obscurity. The government stopped targeting it and even stopped releasing the percentage rate. But once a year the raw number of people registered as unemployed still appears in the official database, albeit with a substantial lag. The number jumped at the end of 2025 to 12.7m, an increase of almost 16% from the year before (see chart).
That total must be interpreted with care. Rule changes in 2020, just before the pandemic struck, made it easier to register, especially for people living far from their place of birth. But no tweaks in the rules explain the more recent increase in registrations, argues Mr Batson. The rise would instead seem to represent "real stress in the labour market".
If so, then this uncharacteristic jump in a once inert statistic may also illustrate a deeper principle. "When a measure ceases to be a target," Mr Batson suggests, "it becomes a good measure again." Call it Batson's corollary to Goodhart's law. If the registered unemployment number attracts too much fresh attention, of course, the government may stop updating it. It has form. When youth-unemployment figures became a lightning rod for criticism in 2023, the government solved the problem by suspending them. Tracking China's economy has never been easy. When a measure becomes a target, it ceases to be a good measure. And when a measure becomes too embarrassing, it ceases to be. ■
Trade politics
Trash talk
SINGAPORE
The White House's absurd claim of a Chinese transshipment "scam"
IN G.K. CHESTERTON'S parable of the fence, those eager to tear down seemingly useless structures are urged to understand why they were built in the first place. Chesterton's fence has long been received wisdom among American conservatives of a certain age, but Peter Navarro, Donald Trump's chief trade ideologue, seems to have missed the memo. In a new paper published by Mr Navarro's office in the White House, titled "The Great Transshipment Scam", dozens of countries from Japan to Jordan stand accused of abetting a Chinese plot to dodge American tariffs. Absurd as this claim is, the paper is nevertheless a vivid exposition of the Trump administration's muddled thinking on trade.
Start with the "great reallocation", Mr Navarro's term for the shifts that followed Mr Trump's first wave of tariffs on Chinese goods in 2018. At the time, free-traders warned that a gap in tariff rates between China and the rest of the world would backfire. It would create an irresistible incentive to evade duties, legally or otherwise, while crimping growth. America understood this well when, in the 1940s, it was the primary author of a global trading system based on non-discrimination.
Mr Trump charged ahead anyway. The average tariff-rate differential between China and the rest of the world widened from 0.9 percentage points at the start of 2018 to 29 points last year (see chart). The predictable happened: the share of American imports coming from China tumbled from 21% to 9%, while that from other countries soared. In "connector" economies such as Vietnam, Malaysia and Mexico, imports from China and exports to
America rose in tandem. Messrs Trump and Navarro now face a problem of their own making.
Mr Navarro calls the pursuit of lower tariffs the "financial engine behind the Great Transshipment Scam", echoing Mr Trump's threat last year to impose a 40% tariff on "transshipment" (a loosely defined term that usually refers to Chinese goods re-routed through third countries). Yet rejigging supply chains to shift a shipment's country of origin and reduce its tariff liability is no scam—unlike concealing its origin by simply slapping on a new label.
The traditional legal test is "substantial transformation", requiring a "fundamental change in form, appearance, nature or character", according to American trade authorities. This is not a clean distinction. Most free-trade deals include painstakingly negotiated "rules of origin" (ROO) for each product line. Haggling over ROOs for the Trans-Pacific Partnership, a 12-country trade deal scuttled by Mr Trump in 2017, took the better part of a decade, notes Deborah Elms of the Hinrich Foundation, a think-tank in Singapore.
Even Mr Navarro offers a cursory nod to the difference between "legitimate manufacturing and substantial transformation" on the one hand and "pass-through trade and origin shifting" on the other. Indeed, he collates five private- and public-sector estimates of "illegal transshipment flows", amounting to as much as $303bn—an improbable 56% of America's imports from China in 2018—each year. Yet even the most convincing of these, an analysis by Goldman Sachs finding that $40bn-worth of Chinese goods were "superficially re-exported" to America in 2023, does not claim that the flows are illegal. (Granted, some will be customs fraud, which America is within its rights to clamp down on. Mr Trump's Trade Fraud Task Force claims to have recovered $1bn from tariff-dodgers since launching last year.)
Were Mr Navarro exercised about illegality alone, third countries building factories making goods from Chinese components should pose no problem. Yet he rails against "goods that are not necessarily declared as Chinese" but which involve "China-origin inputs or components, Chinese ownership or financing, relationships with Chinese suppliers or manufacturers, China-based production steps [or] China-origin routing histories".
This is a staggeringly broad objection. Modern manufacturing is structured as a global network of value chains, distributing capacity across countries in each stage of production. But if proximity to Chinese inputs or firms is evidence of complicity in an anti-American scam, then America is at war with trade itself. It is a "redefinition of trade such that everything is transshipment", says Ms Elms. ■
Trade transgression
US, trade-weighted average tariffs, %
Source: Peterson Institute for International Economics
62 Finance & economics
Mr Navarro's paper describes the "functional architecture" of the "shadow transshipment network" that stretches across 43 countries, making up over 70% of America's non-Chinese imports. From assembly hubs in South-East Asia to logistics centres in the United Arab Emirates and Canada, the supposed scam snakes through "production-side transformation claims, logistics-side routing channels, processing zones, maritime gateways, overland corridors, bonded warehouses and re-invoicing
systems". This is, more or less, the entire infrastructure of global trade; policing it would require a panopticon. Still, Mr Navarro is willing to try. He vaguely threatens setting up an AI-enabled "detective border" system.
After all this, you may wonder whether Mr Navarro thinks imports to America should contain any Chinese content at all. If so, how much? An extreme (though coherent) answer would be to set a maximum threshold for Chinese content, and ban or
tax all imports that breach it. Mr Trump's 40% transshipment tariffs could then be used to strong-arm countries whose exports contain too much from China.
That has not come to pass, presumably because doing so would drive up prices for an American public already enraged about the cost of living. Yet Mr Navarro's extraordinary paper has performed a service. It is a further illustration of the mess his boss's attack on the global trading order has brought about. ■
BUTTONWOOD
Lies, damned lies and stock markets
Benchmark indices are becoming divorced from equity reality
STOCK MARKETS, Wall Street-wary politicians and like-minded commentators repeat ad nauseam, are not the real economy. The S&P 500 index of America's biggest firms is up by 12% this year, setting one all-time high after another. Meanwhile, they note, though American GDP growth has been steady and unemployment is fairly low, consumers feel more downbeat than at just about any time since the University of Michigan started keeping track in 1952.
All true—and nauseatingly familiar. Less remarked is the fact that the S&P 500 and the world's other benchmark indices are no longer the real stock market, either. Rather than mirroring the universe of domestic equities, they increasingly reflect the fortunes of a few corporate giants, which in turn mostly rise and fall with the unstable outlook for the artificial-intelligence revolution. This is making many benchmarks—and, by extension, the stock portfolios and pension pots of investors everywhere—considerably more volatile.
Take the S&P 500. The biggest constituent, Nvidia, accounts for 8% of its value. Imagine the $5.3trn chip-designer's share price moves up or down by 3% (as has happened 30 times this year) and the other 499 stocks are flat. The index then shifts by 0.25% in the same direction. No single name in the S&P 500's history has packed a bigger punch.
Yet in comparison with the index it is a featherweight next to TSMC, which manufactures Nvidia's AI processors. The semiconductor hulk, whose market capitalisation has doubled to $2trn over the past year, accounts for over 40% of the Taiwan Stock Exchange's TAIEX index. To let investors partake in TSMC's success, last year the country's markets regulator dropped the 20-year-old rule
that index-tracking exchange-traded funds (ETFs) must hold no more than 30% of their assets in a single stock. As a result the TAIEX, which comprises just over 1,000 stocks, is behaving like an equal-weighted index of just six firms (or fewer if you consider that other big Taiwanese companies like MediaTek and Foxconn are on the same AI steroids as TSMC and their movements are therefore partly correlated).
Things are only slightly less top-heavy in South Korea. The same bit of financial maths (which, for the curious, involves calculating the inverse of the Herfindahl-Hirschman index for component stocks) implies that the 830-strong KOSPI benchmark acts as if it contained just 11 firms besides the country's two muscular memory-chip champions, Samsung Electronics and SK Hynix. The pair's combined weight in the KOSPI briefly exceeded 40% in June, up from a quarter, give or take, in the preceding few years. (South Koreans' love of leveraged ETFs, which use borrowed money to amplify gains but also losses, increases the KOSPI's jumpiness.)
Between May and July, as traders con-
constantly repriced the size of global AI spending, including on memory chips, the KOSPI's volatility was four times last year's average. The index moved by 5% or more once every three sessions, on average. On July 28th it sank by 11%. Three days later it soared by 18%. In the first six months of 2026 the Korea Exchange had to suspend trading five times and curb it on nearly 30 other occasions, more than in the whole of 2008, at the height of the global financial crisis.
Concentration and volatility need not go hand in hand. Nearly half the value of the broad Swiss index, SPI, sits in five firms: Roche, Novartis, Nestlé, ABB and UBS. But their industries are varied, and representative of Switzerland's listed firms. Almost one in two Swiss public companies makes drugs, food or drink, machinery or loans. In Taiwan and South Korea, by contrast, only one in six and one in 200, respectively, churn out chips or other technology hardware.
The top three Swiss stocks are also stodgy ones with stable earnings and regular dividends—and so inherently less skittish than tech shares. AI will transform them, as it will all sectors. But their fortunes are not hitched to the vagaries of the AI capex cycle.
Like a championship belt, dominance can be forfeited. Novo Nordisk took over the Danish bourse thanks to Ozempic, then shed its index weight as Eli Lilly's better obesity drug ate its lunch. That serves as a reminder that the stock market aims to capture investors' best collective guess at tomorrow's equity reality. It mirrors today's only when the future is likely to look much like the present. In the age of AI, and of abundant private capital that feeds a long roster of unlisted pretenders which benchmarks omit, that is increasingly not the case.
Finance & economics 63
The phantom tollbooth
Freedom of navigation is worth preserving
PFLZGRFENSTEIN CASTLE, near the village of Kaub in Germany, had a lucrative site. Situated on an island in the Rhine, the waterway that connects the most prosperous bits of western Europe, the redoubt lies on one of the river's shallowest stretches. The tightness of the route past it meant that chains could be lowered from the battlement to prevent passage and demand payment. Pflzgrfenstein was a toll castle, one of many along the course of the river.
Rhine tolls were a classic example of an economic rent, a payment that some entity is able to extract by making a cheap resource artificially scarce. The river cost the Electors Palatine who usually controlled the castle absolutely nothing, so the toll they earned from stopping others from making their way upriver was pure rent. In the aftermath of the latest Gulf war, there is a risk that tolls of this sort will come back. The Iranian government is considering levying a toll on ships traversing the Strait of Hormuz, through which around 20% of the world's oil supply passed before the war began in February. Oman, on the other side of the Strait, may join in. America, too, may want to get in on the act: President Donald Trump has said that his country should consider levying fees as "THE GUARDIAN OF THE HORMUZ STRAIT".
You might call this the political economy of the rent-seeking sea, to paraphrase the title of a paper by Anne Krueger, an American economist, published in 1974. Restrictions on trade create rents and that creates a new type of competition, in which different groups compete to capture those rents rather than doing anything productive. The trade restrictions that Ms Krueger observed in the "rent-seeking society" typically meant lobbying, bribery or legal chicanery. For charging tolls on passing maritime traffic the rent-seeking activity could be far more deadly: piracy, drones and so on. During the age of sail territorial waters ended three nautical miles (5.6km) from the coast, roughly the range of a cannon shot. It was a recognition that sovereignty could stretch only so far.
For the past couple of centuries, however, freedom of navigation has been the norm. That has typically been a public good provided by naval powers capable of asserting their sovereignty, or extending its reach beyond their shores. Tolls were abolished at
Pflzgrfenstein castle in 1867: the Prussian government controlled both sides of the river and was at last able to make good on a diplomatic promise made years earlier. The rent extraction ended. In 1868 an international treaty abolished all tolls on the Rhine. Elsewhere change was also under way. For years the Danish crown had earned much of its revenue by levying tolls on access to the Baltic Sea, until in 1857 the great powers forced the end of that money-making scheme.
On the high seas freedom of navigation was assured by hegemonic powers, able to project their authority everywhere. For much of the 19th and 20th centuries the hegemon was Britain, notably in the Persian Gulf. The British Empire used a crackdown on piracy to strong-arm emirs into signing treaties, starting in 1820; their fiefs became the Trucial States. The empire secured access to the shipping lanes without formally incorporating them. The torch passed to America after 1971, as Britain withdrew from east of Suez. The Gulf states got their independence but also fell under America's shield. America occupied the former British naval base in Bahrain.
Keeping waterways toll-free, then, is an example of the hegemonic stability theory formulated by Charles Kindleberger, another American economist: to remain in good order, the global economy requires certain public goods. The hegemon is the only one with both the ability and the desire to provide them; it captures some of the large gains from a functioning world economy.
Kindleberger, who was writing about the Depression, was primarily concerned with financial stability, but freedom of navigation also fits the model. In his theory, and later versions, only the hegemon was able to both bear the costs of providing such goods and capture a big chunk of the benefits: "Our economic interest is in maximum freedom of movement for merchant vessels at minimum possible cost," wrote officials in America's State Department in 1969. Now America is retreating from that role, partly for lack of will—Mr Trump is reluctant to commit troops to the Gulf—partly for lack of ability. Cheap drones can project power much farther and more cheaply than the three-mile cannon shot of yore.
Is there a third option, neither anarchy nor hegemonic domination? Perhaps collections of other states—the "middle powers" referred to by Mark Carney, Canada's prime minister—could club together to provide public goods. The treaty guaranteeing a toll-free Rhine was signed by a group of German states, France and the Netherlands. But a paper last year by Anwesha Banerjee, Ottmar Edenhofer and Ulrike Kornek, three economists based in Germany, sounds a note of scepticism; a more equal balance of power between countries encourages free-riding and jockeying for position. It may be unlikely, then, that middle powers would do any better at keeping Hormuz toll-free.
History doesn't repeat itself but it does Rhine
Then there are the truly global public goods. Whether or not there are tolls on the Rhine is, for the time being, a moot point. Europe's extended heatwave and drought, now into its third month, has reduced the water level at Kaub to just six centimetres, below the previous record of 25cm in 2018. That is barely enough for children to paddle in, let alone for heavily laden tankers to pass. This will knock 0.2 percentage points off German GDP this year, estimates Oxford Economics, a consultancy. The danger is not only that states manage to extract rents from what was once free, but that the world is less capable of providing public goods at all. ■
64
Science & technology
Global warming and aviation
Bluer cooler skies
A trial over the Atlantic aims to show how much Earth could be cooled by contrail-free flying
THE 40M scheduled passenger flights which took place in 2019 added up to about 61bn kilometres of travel—exceeding the distance from Earth to the Sun every day of the year. More than 40% of those flights left contrails behind them for at least some of the journey as water vapour in the exhaust from their engines froze into tiny particles of ice.
Many of these trails were ephemeral. They streamed out behind their progenitor planes before fading quickly away. But over about 5% of the total distance flow—comfortably enough for ten round-trips a day to the Moon—they persisted and widened as water vapour already in the atmosphere joined in the freezing. Thin white lines became broad translucent stripes of cirrus cloud. Wind those stripes around Earth like yarn around a ball and you will find that, at any one time, persistent contrails cover maybe a thousandth of the sky.
In doing so they warm the planet.
Contrails are white because their constituent ice crystals reflect sunlight. That can have a cooling effect during daytime. But the crystals also absorb infrared radiation and re-emit it, rather as greenhouse gases do. Some is thus radiated into space, but the rest travels back towards Earth. This produces 24-hour warming that, overall, handily outweighs the cooling (something which is true for non-contrail cirrus, too). The exact amount of net warming is hard to calculate, but current evidence
→ ALSO IN THIS SECTION
66 Probing Antarctica's ice
66 Dinosaurs' stomach stones
67 A new subatomic particle
67 Well Informed: When working, move it!
suggests aircrafts' contrails and the cirrus clouds they induce may do more to raise the planet's temperature than carbon dioxide emitted by the aircrafts' engines.
This warming is, though, to some extent, optional. For contrails to persist the air they are in must be well below freezing point and yet contain a significant amount of not-yet-frozen water vapour. The thing which stops this water vapour freezing is a lack of particles that might "seed" the process by acting as nuclei around which crystals can form. A contrail provides such seeds by the trillion, and so creates cloud where there was none.
The nucleus of the problem
Such ice-supersaturated regions (ISSRs) tend to be found at altitudes of 8-13km. They are a few hundred kilometres across, but only a few hundred metres deep. Keep aircraft out of them and you should get rid of a lot of persistent contrails—which means getting rid of some warming, too.
On August 18th a consortium based in Britain announced Operation Blue Skies, the most important trial of this approach to cooling the climate so far undertaken. It will take place in Shanwick, a piece of airspace that covers 1.8m square kilometres of the North Atlantic west of the British Isles (see map on next page).
Science & technology 65
There have been smaller trials of the idea. One was run by Germany's aerospace agency in 2021 and another, in 2025, by American Airlines and Contrails.org, a not-for-profit research organisation largely funded by Breakthrough Energy, a network of investors founded by Bill Gates. The results were promising. Blue Skies, which is considerably bigger, should show whether ways to predict and route around ISSRs are both effective and suitable for operational use.
Shanwick is a bit of airspace well suited to such a project. It is prone to the right sort of damp chilliness at airliners' cruising altitudes, and it is used by a lot of planes. In combination, those factors mean that, although it covers just 0.4% of Earth's surface, models suggest Shanwick is responsible for about 5% of contrail-associated warming. What is more, European and American weather satellites monitor the area day and night in both visible-light frequencies and infrared, making it easy to see what is going on.
The Winter's tale
The plan is that on between 20 and 40 days over the coming winter and again the one after that air-traffic controllers at NATS, the company which looks after British airspace, will alter the flight paths of aircraft passing through Shanwick to keep them out of suspected ISSRs. These changes will route the planes a few hundred metres below their flight plans when conditions suggest it and when there is no reason not to (such as turbulence).
Comparing contrails made during periods when these changes are happening and periods when they are not should provide reliable data on how much the contrails are being reduced. "The perfect outcome", says Sebastian Eastham, part of the team working on Blue Skies at Imperial College, London, "is that a satellite image robustly and reliably has fewer contrails in it, time after time after time."
As well as NATS and Imperial, the consortium behind Blue Skies consists of Contrails.org, Google (which, as well as its software expertise, is chipping £1.4m, or $1.0m, of hardware and computing time into the £5m budget), the University of Cambridge and Britain's Met Office.
If contrail avoidance is to go mainstream, the Blue Skies trial needs to show not just that days with avoidance measures in place have fewer contrails, but also that integrating the process into normal operations can be done safely and straightforwardly. That would open the way to more widespread, and eventually everyday use.
Adoption does not have to be universal to be useful. The warming attributed to contrails is concentrated. By one estimate 80% of the temperature-raising effect on the climate comes from just 2% of flights.
Awkwardly, you cannot know in advance exactly which 2% they are. But Marc Shapiro, who runs Contrails.org, says that a system eliminating 70% of the world's contrail warming might require changing the flight plans of fewer than 5% of flights.
According to a recent paper by Jessie Smith of Cambridge and colleagues elsewhere, if aviation does not change its ways it will be responsible for about 0.01°C of total warming above preindustrial levels in 2050. If the world is seeing about 2°C of warming by then, that contribution comes in at 5%, with more than half of it induced by contrails. A 75%-effective contrail-avoidance programme that began in 2035 would reduce the total by about a third, though there is a lot of uncertainty.
Fractions of fractions sound small, but Dr Smith and her colleagues argue that the effect is similar in size to that of any other single measure proposed to reduce warming as a result of air transport—and could be realised faster.
The benefits are not free. Flying at lower altitudes increases the amount of fuel used, which is a cost to airlines, and the amount of carbon dioxide emitted, which is a cost to the planet. But neither cost is huge. Estimates from models suggest that avoiding ISSRs makes flights only slightly
longer and increases the amount of fuel used by less than 2%.
Using 2% more fuel on 5% of flights would increase the industry's total fuel bill by a thousandth. In the American Airlines trial the difference in fuel use between the flights that were diverted and those which were not was undetectable.
Similar calculations hold for warming. According to Olivier Boucher, a climate scientist who has worked on the matter for 30 years and now runs Klima, a small French company dedicated to it, "If you really focus on the most impactful flights, maybe you can save like 100 tonnes of carbon-dioxide equivalent for an extra tonne of carbon dioxide."
There are also opportunity costs. Some worry that focusing on contrails will distract attention from airlines' carbon-dioxide emissions and efforts either to reduce them (by using sustainable aviation fuels, or SAFs) or offset them (by removing CO₂ from the atmosphere and putting it into permanent storage). Eliminating contrails without attacking these points—or reducing the amount of flying—provides a one-off gain but does not change the fundamental fact that, in the long term, more flying means more warming.
The ideal is the enemy of the good
Doubters argue as well that SAFs can be designed to produce fewer particles around which ice can form, thus reducing contrails as well as carbon dioxide. This is true. But steering clear of ISSRs looks like a far more cost-effective approach to contrail reduction. The paper by Dr Smith and her colleagues found that adding SAFs to a system which already practised avoidance resulted in at most 0.001°C of cooling.
There is, then, no free lunch. But the idea that giving passengers a minute or two longer to savour the in-flight cuisine that they have paid for on a few of the world's many many flights might result in some good is about as appealing as climate interventions get. ■
Ice or no ice?
Forecasted impact of contrail suppression by effectiveness of intervention Increase in global surface temperature, °C
Source: "The climate opportunities and risks of contrail avoidance", by J. Smith et al., Nature, 2026
66 Science & technology
Glaciology
A signal from the noise
The sound of giant tractors may replace dynamite for probing Antarctic ice
THE PISTENBULLY 300 Polar Antarctic vehicle is a 8.5-tonne heavy-duty specialised monster built to withstand extreme polar conditions. It features wide tracks, massive pulling strength, powerful cranes and multi-purpose tools designed for the toughest ice and snow operations. No Antarctic base should be without one. Research led by Liu Guofeng at China University of Geosciences in Beijing, however, suggests the PistenBully may be capable of more than just fetching and carrying. As Dr Liu writes in the Journal of Glaciology, it is also able to do actual science.
Antarctica's ice sheet is the largest reservoir of freshwater on Earth, but it is shrinking. Just above the ice proper, but below the surface snow, is a transition zone 50-100 metres thick called the firn layer. This is where snow is compressed into ice. It is also where meltwater from above is captured in pores and refrozen. Knowing the firn layer's exact thickness and understanding how it behaves is crucial to predicting what will happen, over the long term, to the ice sheet and thus to the world's sea levels. Visualising this subsurface melange is difficult, though.
Current practice, borrowed from oil- and gas-prospecting, is to use an "active source". This is geologist-speak for dropping a stick of dynamite down a drill hole and listening to the explosion's sound, propagated through the rock—or, in this case, ice. The listening is done by sensors called geophones. Since sound travels at different speeds through water, ice and snow, a computer analysis of when and how powerfully the echoes arrive at different geophones permits construction of a high-resolution image of the firn layer.
This works, but is not ideal. First, someone has to drill the hole down which to drop the dynamite. That is not always possible in polar conditions. Then, the explosive must be purchased, brought to Antarctica and the actual dropping carried out. Moreover, detonating dynamite in sensitive environments is not to be done lightly. An alternative source of sound would thus be good. And, after hearing the din created by a convoy of PistenBullys, Dr Liu thought they might be the very thing.
To this end, he and his colleagues experimented with geophones already in place in the Larsemann Hills of East Antarctica. They commandeered some PistenBullys and arranged for them to be driven
Mover and shaker
around the hills. They found that if the vehicles were sent along both sides of a line of geophones, parallel with the sensors, they got the effect they were after.
Since the Larsemann Hills' firn had already been mapped the old-fashioned way, Dr Liu was able to compare his machine-created images with those obtained using dynamite. They were pretty much identical. He thus seems to have devised a way of mapping the firn's structure that is simpler than blowing it up, but equally good. ■
Dinosaur gastroliths
The guts for flight
Stones in dinosaur stomachs may help understanding of the origins of birds
HOW DO YOU know what a dinosaur ate? Guts rot away after death, so the details of dinosaur digestion remain obscure. The best indicators are teeth. Sharp, pointed teeth probably ripped chunks of flesh off prey for swallowing whole. Robust grinding surfaces suggest lots of chewing was involved. Yet many dinosaurs belonging to a group called the theropods, which gave rise to birds, lost their teeth altogether and evolved beaks. This has made it hard to work out what they ate and how they digested it.
There is, however, a second clue. Lots of dinosaurs had stones, known as gastroliths, in their stomachs. These are assumed to have assisted with digestion, and thus have their own tales to tell. And modern birds and crocodilians (which are both, like dinosaurs, members of a larger evolution-
ary group called the archosaurs) often have gastroliths, too.
Takasaki Ryuji at Okayama University of Science, in Japan, decided to investigate. In a study published in Paleobiology, he and his colleagues examined the gastroliths of 46 modern archosaur species (42 birds and four crocodilians) and compared them with those from 16 species of dinosaur.
The birds included herbivores, omnivores and carnivores. The crocodilians were all carnivorous. The shapes of the stones varied. As might be expected, if an animal had a highly muscular stomach (ducks and geese), its gastroliths tended to have been worn into the sorts of rounded shapes displayed by beach pebbles. Weaker stomachs (seabirds and crocodiles) contained more angular stones.
Looking at the dinosaurs, Dr Takasaki and his colleagues found that ornithischians (think Triceratops), a group which their teeth suggest were herbivores, retained angular gastroliths. These creatures, this implies, relied mostly on their jaws to grind up their food, leaving their stomachs to apply merely the finishing touches.
Long-necked sauropods (think Diplodocus) also had angular gastroliths. That is more of a puzzle, for these animals had pencil-like teeth using which, it is assumed, they stripped the leaves off tree branches that less cervically elongated animals were unable to reach. The explanation, Dr Takasaki suspects, is that like those of modern ungulates, sauropods' stomachs depended mainly on fermentation to break food down.
Stones in the guts of theropods tell yet another story. Tarbosaurus, a hypercarnivore similar to Tyrannosaurus, and similarly endowed with sharp, pointed teeth, had angular gastroliths. However, theropods with beaks (of which science recognises at least four groups) had rounded ones. Indeed, Dr Takasaki saw a clear relation between tooth loss and stone roundness, as the work of grinding up food was transferred from jaws to stomach.
That shift may have helped make possible the evolutionary success of birds. Teeth are heavy, as are the jaw muscles those teeth require to do their job. This puts a lot of weight near an animal's front. Shifting the job of grinding, and the muscles and hard surfaces involved, to the stomach centralises an animal's centre of gravity in a way that would make it easier for it to rise from the ground.
Curiously, it is clear from the fossil record that the very earliest birds of all did not have beaks. But the fact that so many of their theropod relatives did, and that birds themselves evolved them at least twice, suggests the transition was easy. It may thus be that part of the secret to taking successfully to the skies was, quite literally, having the stomach for it. ■
Science & technology 67
Particle physics
Make it stick
Physicists say they have found an elusive particle: the glueball
A NEW SUBATOMIC particle has made the leap from theory to reality. So say the scientists of the Beijing Spectrometer III (BESIII), a Chinese physics experiment. Speaking on their behalf at the International Conference on High Energy Physics (ICHEP) on August 5th, Jin Shan, of Nanjing University, announced that the team have discovered a glueball—a hitherto hypothetical particle that is unlike any which physicists have seen before.
The nuclei of atoms are held together by a phenomenon called the strong nuclear force. What better name, then, for the particle responsible for carrying this force than the “gluon”? Gluons interact with quarks, another type of fundamental particle, wrangling them into pairs and triplets to create protons, neutrons and other such “composite” particles. But gluons also interact with other gluons, raising the possibility of gluons gluing themselves together into a different kind of composite particle: the glueball.
Physicists’ theories predict glueballs. Finding one is another matter, since those theories also suggest glueballs will be difficult to distinguish from the torrents of more mundane quark-based particles produced in machines like the Beijing Electron-Positron Collider II, to which BESIII is attached. Almost half a century of searching has left researchers empty-handed.
No more, say the operators of BESIII. In 2011 a new particle, called X(2370), showed up in their detectors. It looked a lot like a glueball, but other interpretations were possible. So the team got to work on ruling them out. They have spent the intervening years analysing over 10bn processes in which X(2370) could play a role. The result—published online in July and presented at ICHEP by Dr Jin—is a complete picture of the new particle’s properties. The team is now convinced it is a glueball. “No other interpretation can explain all these properties simultaneously,” says Dr Jin.
Detecting the glueball—if it has, indeed, been found—is not just a matter of physicists adding a new specimen to their by now extensive collection. Its discovery will provide valuable data to test the theory behind the strong force. That theory, good as it is, contains many mysteries. One is how the strong force generates mass—for gluons have no mass, but glueballs do. On the answering of such questions, the glueball’s discovery could tip the scales. ■
Well Informed
Should you stand at your desk?
Sitting too much is unhealthy, but standing instead is no fix
PROLONGED SITTING rightly has a bad rap. A review in 2012 concluded that the most sedentary were more than twice as likely as the least to have or develop diabetes, and also faced a much higher risk of premature death. Another, in 2015, found that sitting for long stretches was associated with “deleterious health outcomes regardless of physical activity”. Three years after that an analysis of news articles in the British Journal of Sports Medicine found hundreds of pieces which referred to claims likening sitting to smoking.
That last claim is overblown. In 2018 the American Journal of Public Health reckoned sitting’s true risk to be a tenth of that of smoking. But this is still bad. Hence the appeal of standing desks and, more recently, “active seats”: chairs and stools that tilt, wobble or rock to force continual postural adjustments. Is either actually healthier?
An analysis in 2018 found standing at a desk burned just nine more calories per hour than sitting. Nor are the cardiovascular gains impressive. A review in 2020 pooled nine trials involving 877 people. Participants stood about 1.3 hours more a day and were followed for roughly four months. Body fat and blood-sugar control improved slightly, but blood pressure, insulin, cholesterol and triglycerides did not. The study did
not, however, look at blood flow and the activation of postural muscles, both factors which might contribute to the benefits of standing over sitting.
That said, standing can have downsides. Work published in the International Journal of Epidemiology in 2024 followed 83,013 adults in Britain for an average of nearly seven years. Those who stood for more than two hours a day had a higher risk of varicose veins and venous ulcers, perhaps because of pooling of blood in the legs.
What matters far more is movement. A study of nearly 482,000 Taiwanese adults in 2024 found that just 15–30 minutes of extra daily exercise could erase most of the excess mortality risk associated with sitting at work. Desk-bound workers may therefore wonder if active seating can make a difference. Some findings are encouraging.
Consider a trial at Georgia Southern University. In this 20 adults performed reading and typing tasks on an ordinary chair, an exercise ball and a seat mounted on top of an inflatable bladder. Energy expenditure on the ball was similar to normal sitting. But on the bladder, participants’ heart rates and calorie burn rose by 6–13% and 19–40% respectively. Writing in Ergonomics in 2019, the researchers attributed this to muscle contractions required to maintain balance.
A longer trial found another benefit. For six months 133 office workers in Bangkok sat on either a foam seat pad or a slightly unstable inflatable one designed to encourage frequent postural changes. Those using the latter were 81% and 84% less likely to develop neck and lower-back pain, respectively. Writing in the Scandinavian Journal of Work, Environment & Health in 2024, the researchers described the pad as a “light-exercise device”. They reckoned it reduced fatiguing muscle contractions and improved blood flow to compressed tissues.
There are caveats. Many studies of dynamic sitting have been small or brief. Results have varied with seat design, and some participants have reported discomfort in the derrière. That said, there does seem to be a lesson here: the real distinction is not between sitting and standing, but between remaining still and moving.
68
Culture
Man's best frenemy
What dogs can teach us about AI
Two books ponder whether the technology will be tame or savage
The Rise and Fall of the Artificial State.
By Jill Lepore. Liveright; 336 pages; $29.99.
Allen Lane; £25
What Happened to Liberal Democracy?
By Daron Acemoglu. Dutton; 416 pages; $32.
Profile Books; £25
WHEN MULLING how AI will transform society, consider an earlier technology that gave humans a helpful companion with superior capabilities: the dog. By domesticating wolves, humans were able, indirectly, to smell game far away and hear enemies approaching at night. They became better hunters than Neanderthals and better warriors: hard to ambush, hard to elude. Dogs are one reason why Homo sapiens outcompeted their prehistoric cousins and conquered the world, some anthropologists believe.
Back when dog technology was first being developed, it was far from obvious that
it was safe for people—who were literally inviting wolves into their caves. They had to train them not to eat babies. Thousands of years later, dogs still sometimes escape and go rogue. But overall, people are happy with how things have worked out. (It did not end so well for the Neanderthals.)
The optimistic case for AI is that it will be like dogs: complementing rather than replacing human labour, and remaining obedient to its masters. But what if it isn't? AI evolves much faster. Soon it could be
creating new, improved versions of itself at a pace dog-breeders can only dream of. So whereas people can be confident that dogs will not enslave them—by 2030, pooches will still not know how to use a tin-opener—they cannot be so sure about AI.
Two new books, one by a Nobel-prize-winning economist and one by a Pulitzer-garlanded historian, grapple with the question of how AI will shape the future. Neither uses the dog analogy, but both are easier to understand with reference to it.
Jill Lepore, a Harvard historian, focuses on the men who train the dogs. In her telling, AI tycoons are horrible, irresponsible people whose creations may break their leashes and savage humans. In "The Rise and Fall of the Artificial State", she argues that tech bosses are trying to build "a future in which humans are ruled, and many replaced, by machines owned by corporations". This, she says, implies "both the desertion of liberal democracy and the destruction of the natural world".
She bases her case partly on the moguls' own words. Sam Altman, the boss of OpenAI, has said it would be "awesome" to replace human presidents with an AI that could "go around and talk to every person on Earth, understand their exact preferences" and "optimise" for all of them. Elon Musk has said humans may end up like "pet labradors" for superintelligent robots. ▶
→ ALSO IN THIS SECTION
69 Punctuation's (huge) impact!
70 A Renoir portrait's riveting story
70 Cheers to Swedish winemakers
71 Japan's listening bars go global
72 Forgotten Founding Fathers
Culture 69
Mr Altman has suggested that most of Earth's surface may soon be covered with data centres. Mr Musk wants to solve this problem by putting data centres in space—and blasting millions of people to live on Mars. The tech titans all seem to have escape plans, notes Ms Lepore: "Elon Musk to Mars, Mark Zuckerberg to the metaverse...Jeff Bezos to space, Sam Altman's brain uploaded to the cloud." This makes them reckless about the risk of creating dystopia here on Earth, she fears.
She is right that today's tech bosses are not people you would trust to write the rules for the future of civilisation. But she overestimates the importance of individual billionaires. How AI shapes society will not depend on Mr Musk's eccentric political views, any more than the effects of cars depended on Henry Ford's. Rather, it will hinge on the power of the technology itself and the uses to which people put it.
That is a question Daron Acemoglu addresses in "What Happened to Liberal Democracy?". Mr Acemoglu is among the world's most-cited economists. He is best known for "Why Nations Fail", a sweeping economic history arguing that the main factor keeping poor countries poor is predatory institutions. His new work focuses on the present day and the interplay between technology and democracy.
Liberal democracy has underpinned progress, he argues. It has fostered innovation and economic growth, spread the benefits (through taxes and public services) and protected voters with regulations.
For a long time, the system more or less met voters' expectations. But "then computers happened." Digital technologies, in Mr Acemoglu's account, aggravated inequality by raising the demand for brains over brawn. They also boosted the "social power of college graduates", allowing a new liberal elite to impose its values on a resentful working class, provoking a backlash that led to the election of Donald Trump. And they changed how people interact. In America time spent engaging with friends in person has plunged by two-thirds in the past 20 years. Political debate has grown more extreme and toxic.
AI could make everyone much richer by supercharging innovation. But it could also supercharge the harms of digital technology, Mr Acemoglu (rather familiarly) fears. It could lead to widespread joblessness, creating "huge social-status gaps" between "makers" and "takers" and a full-blown crisis of democracy. (It could also accelerate social atomisation by making loverbots and political fakery more seductive.)
His proposal is "pro-worker AI". By this, Mr Acemoglu means finding ways for workers to use AI to make their labour more valuable, like prehistoric hunters combining their spear-making and throwing skills with a dog's nose and speed.
Maybe AI could enable electricians to troubleshoot increasingly sophisticated grids, spotting problems via sensors and helping humans analyse them with a database of past glitches and solutions. Perhaps AI could improve education: an AI assistant could assess pupils' answers to quizzes and suggest ways to split the class into smaller groups, who could be offered tailored instruction, thus increasing the demand for human teachers.
It is hard to quarrel with this. AI will certainly cause less social turmoil if it complements human labour rather than making it redundant. But Mr Acemoglu leaves unresolved the question of how this is to be accomplished. Ultimately, both books are disappointing. Inadvertently, they illuminate how hard it is to write persuasively about a technology that is changing so fast, and whose future capabilities are so utterly unknown.
Politically, AI has become unpopular. Twice as many Americans are pessimistic about its long-term effects on society as are optimistic, and only 2% think it is developing "too slowly". But one way or another, this new wolf will be invited into the cave. The countries pioneering AI technology, America and China, are rivals. Neither wants to be the second to develop a super-intelligent ally, for fear of ending up like those hapless, dogless Neanderthals. ■
Punctuation
To the point
On the Mark. By Florence Hazrat. Basic Books; 352 pages; $32. Profile Books; £22
"IN THE BEGINNING was the word, and the word was with God, and the word was God." Yet in the beginning there was no punctuation. In early Latin manuscripts, the Bible was written in scriptura continua, without spaces or marks between words. This was a problem: pause at the wrong point, and you could sever God from "the word", throwing Christ's divinity into doubt. So in the Middle Ages, Christian scholars set about punctuating the holy book, using symbols and sub-clauses as "lanterns to words" to guide preachers.
"The history of punctuation is the history of humanity," argues Florence Hazrat, a researcher, in a rollicking new book. Often dismissed as "the invisible ushers of the sentence", dots and dashes have shaped religious texts, legal documents, and literature. "On the Mark", like punctuation itself, makes readers pause for
thought, showing how full stops have displeased and commas caused controversy.
This pointed history starts in the ancient world. Cicero, like other Roman scholars, thought punctuation was amateurish. For skilled orators, half the fun was in interpreting obscure writing. Ms Hazrat races across time—offering digressions about how, for instance, Irish monks in the eighth century first put spaces between words—before reaching the "punctuation boom" of the Renaissance. Scholars in Italy invented curved commas, semicolons and (do not forget) parentheses. The printing press helped spread and standardise these signs, and by around 1600 all the major marks were in place.
"On the Mark" is crammed with fascinating facts. Take the exclamation mark, about which Ms Hazrat wrote a previous book. Invented by an Italian poet in the 1360s to signal "admiration and wonder", by the 19th century it had become a sign of emotion and even hysteria. Editors added them to old texts, including Shakespeare's "Antony and Cleopatra". (In the First Folio, the play had only 16 exclamation marks, but modern versions contain about 200, most of them clustered in Cleopatra's shouty-sounding speeches.) In the 20th century the exclamation mark fell into "soundless cymbal-crashing", as Theodor Adorno, a German philosopher, put it. Nazi propagandists aggressively attached clusters of them to the end of assertions!!!
Writers have long squabbled over scribbles. Emily Dickinson felt "defeated" when an editor cut a dash; Charles Baudelaire told his proofreader "I absolutely want to keep this comma." Mark Twain even threatened to have a proofreader shot who dared change his punctuation. This is a delightful tale of marks' evolution, even if Ms Hazrat's own attempts to use them inventively fall a dash flat. ■
70 Culture
Money, sex and bigotry in France
Get the picture?
The Renoir Girls. By Catherine Ostler.
Atria Books; 432 pages; $35.
Simon and Schuster; £30
IT WAS CALLED “the capital of the 19th century” and “modern Babylon”. To those living during the Belle Époque, which stretched from the 1870s to the onset of the first world war, Paris felt like the centre of the world. And at the very centre of Parisian life were the Cahen d’Anvers, a Jewish banking family. They held salons, threw lavish parties and counted celebrated writers, including Marcel Proust and Guy de Maupassant, as friends.
The childhoods of the three Cahen d’Anvers daughters were immortalised by Pierre-Auguste Renoir in commissioned portraits. One, of Alice and Elisabeth holding hands (see picture), hangs in the Museu de Arte de São Paulo. It looks as though everything about the girls’ lives, save their ruffled dresses, was smooth. In fact, the painting “reminds us that nothing and no one is quite safe from the hands of time or the evil within humanity, apart, sometimes, from art itself”, writes Catherine Ostler, a historian, in a new book.
When you gaze at Renoir’s pretty paintings, several things are obscured from view. One is paternity. Rumours abounded that Louise, the matriarch of the Cahen d’Anvers family, was the paramour of Charles Ephrussi, an art-loving critic and collector. It was a time of “rebellious, emancipated, intellectual feminism”, when wealthy women were free to hold salons and live more independent lives. There were whispers that Alice might be Ephrussi’s daughter. Ms Ostler uses a DNA test of a descendant to show that she probably was (the same may have been true of her brother, who was named Charles). This is not, then, your average portrait.
The painting also masks the antisemitism that would change the girls’ lives. When it was painted in 1881, France was ostensibly in a tolerant period (though after receiving a late payment, Renoir wrote to a friend complaining that the payment was “stingy” and declaring that “I really give up on the Jews”). Then the Dreyfus affair—in which a Jewish army officer was falsely accused and convicted of passing secrets to Germany—laid bare people’s prejudices and jealousy. Between 1807 and 1914 some 880 Jews in Paris converted to Catholicism; about a quarter of them did so in just three years, 1895-97,
when the Dreyfus affair was most heated.
One of them was Elisabeth (in blue), who chose to be baptised at the age of 20. Spending most of her life as a Christian, however, did not protect her. The collaborationist Vichy regime began to rule France, and a mayor whose family had mingled with the Cahen d’Anvers for generations certified that Elisabeth, then 69 years old, was Jewish.
Suffering from severe arthritis and unable to walk without help, she was arrested by the Nazis, who sent her to Auschwitz, where she was probably murdered on arrival, if she had survived the 55-hour train journey. Ms Ostler focuses on the many betrayals that facilitated this fate. The Cahen d’Anvers had fought in the first world war and been generous patrons of France, including donating their chateau to the state. And yet the country turned on them.
Ms Ostler came upon the Cahen d’Anvers while reading “The Hare with Amber Eyes”, which masterfully traces art through time (the author, Edmund de Waal, inherited sculptures bought by Ephrussi). She has doggedly dug up clips and correspondence that evoke the Belle Époque. Sometimes, though, “The Renoir Girls” gets too bogged down in social minutiae that must have preoccupied the Cahen d’Anvers. Few really want to know which Rothschilds were at weddings, what colour the bridesmaids wore and whether the press preferred Louise’s previous hairstyle. But the broader themes—about the hidden history of objects, and what art captures and obscures—are heart-rending and memorable. You will not look at another 19th-century portrait the same way again. ■
All dressed up with everywhere to go
Wine
Vin-dinavia
HOGANAS
World-class wine is being made in a corner of Sweden. Yes, Sweden
AT THE NOBEL PRIZE banquet in Stockholm in December, not all the winners were seated at the table. A victor of a different kind could be found in the diners’ wine glasses. In 2024 a blind tasting had pitted a dozen of Europe’s finest wines against a dozen wines from Sweden. An international panel of 18 independent experts declared Immelen, a white wine from Kullabergs, a winery in the Skane region of southern Sweden, the overall winner, above famous whites from France, Italy and Spain. That made Immelen the obvious choice for the Nobel banquet. It was a coming-out party for Swedish wine.
Sweden is “a young wine country”, says Victor Dahl, chief executive of Kullabergs (pictured on next page). EU approval for winemaking in Sweden and Denmark, required for commercial production, was granted only in 1999. There are now about 50 commercial wineries in Sweden. Sales are still tiny—just a drop in the barrel by global standards—but growing quickly.
At the Nobel banquet, Mr Dahl moonlighted as a waiter so he could hear what the diners thought. Such was their level of interest that he was soon unmasked and invited to introduce the wine. “They had so many questions,” he recalls.
And no wonder. That wine can be made in Sweden at all may come as a surprise. Global warming has helped. But “climate change is also bad for us,” says Mr Dahl. Although the climate is warmer overall, it is also unpredictable: since 2018 every growing season has been different. Sweden’s long summer days provide plenty of sunshine to ripen grapes. Indeed, slower ripening may enhance flavour.
Immelen is a blend. It is two-thirds Solaris, a descendant of Riesling and Pinot Gris created in 1975, and now Sweden’s most widely planted varietal. The rest is Souvignier Gris and a dash of Muscaris (a cross between Solaris and Muscat). The resulting wine is surprisingly full-bodied, with none of the under-ripeness often found in cool-climate wines. It reminded this correspondent of a high-end Sancerre.
Just as significant as climate change has been the rise of New Nordic cuisine, a gastronomic culture centred on Copenhagen and Malmo. Skane is sometimes referred to as the “Nordic Tuscany” and is being promoted as a tourist destination for foodies. In this part of the world seasonal tasting menus abound, and foragers gather ▶
Culture 71
Grape expectations
► local ingredients for restaurants. Magnus Nilsson—whose previous restaurant, Faviken, earned two Michelin stars—has recently opened a new one in Skane, as has Titti Qvarnstrom, the first Swedish woman to win a star. Being able to serve local wine with local food has obvious appeal.
History has also played a part. Sweden's government ran a campaign from 1957-85 to encourage Swedes to drink wine, rather than spirits, for health reasons. This created a strong wine culture just as European travel was getting cheaper, prompting Swedish wine buffs to engage in wine tourism. But now they can do it at home. The covid-19 pandemic helped, too: Swedish wine critics who were unable to go abroad started to pay more attention to the viticulture scene on their doorstep.
Oddly the Systembolaget, the government monopoly on the sale of alcohol through a national network of shops, has also helped. It bans Swedish wineries from selling bottles over the counter, though they can sell wine for consumption on the premises, for example at an on-site bar. To get into the Systembolaget shops, Swedish winemakers must achieve consistent quality and volume. Systembolaget acts as "an extraordinary quality-control mechanism", says Joe Fattorini, a British wine guru who now lives in Sweden. Once a winery wins approval, as Kullabergs has, distribution is in effect handled by the government.
The contrast with neighbouring Denmark, which lacks such restrictions, is telling. Denmark has fewer hectares of vines than Sweden, but roughly three times as many producers and a wider range of styles. That means more experimentation, but greater variation in quality.
Things are changing in Sweden, however. After much lobbying, the law was changed to allow "cellar-door sales" by
wineries from June 2025. There are still restrictions—sales are permitted only as part of an educational experience lasting at least 30 minutes—and the rule-change is subject to a six-year trial. Sweden's wine industry is still in its infancy and lacks the long history of winemaking in other parts of Europe. But you do not need a Nobel prize to see its potential. ■
Listening bars
For the record
Japan's music cafés are striking a chord abroad
SPACE TALK, located in the London district of Clerkenwell, takes music seriously. Before the bar opened in 2024, its owner enlisted an architect, designer and sound engineer to create the perfect acoustic environment. Resembling the deck of a wooden spaceship, the bar is divided into four zones. On a recent evening, your correspondent sipped cocktails while trippy electronic music floated out of the bespoke sound system. In its ethos, Space Talk is one of many spots around the world loosely inspired by Japan's ongaku kissa (music cafés). These venues are commonly known as listening bars or hi-fi bars.
The first ongaku kissa opened in Japan during the 1920s. Visitors drank tea or coffee as music—often classical, tango or swing—played through gramophones. In the decades after the second world war, cafés devoted to jazz grew popular. Imported records and audio equipment were expensive, and jazz kissa offered a place for
enthusiasts to hear the latest album by Miles Davis or John Coltrane. They were places for quiet, reverential listening.
Owners would (and still do) display a fussy attention to detail. High-quality speakers are a must; the nerdiness often extends to amps, turntables and styluses. (Gramophones have gone, but vinyl is still very much in vogue.) This music is not merely for ambience: the act of listening is central to the experience. Many kissa ask visitors to remain silent. Kawasaki Toshiya, who runs his own listening bar in Tokyo, says the atmosphere this creates is akin to "sitting in meditation at a temple".
The listening bars that have sprung up in the West, South America and Africa—most of which have opened in the past decade—rarely have stringent rules on chitchat. Space Talk, for instance, has not banned nattering, but it has implemented a "no photos or videos" policy. All Blues, a venue in New York, politely asks visitors to "let the music do the talking". Some bars host dedicated events where guests can hear an album in full or listen to music tied to a particular artist or theme.
The "real magic" of ongaku kissa culture is to create an environment "where you can be genuinely moved to the core by music", says Nick Dwyer, the director of a documentary on Tokyo's music cafés. What makes this new wave of venues interesting, he says, is that proprietors have "their own take" on the format and are adapting it into "something that works in London, New York, Barcelona or wherever".
There are three reasons why the ongaku kissa model has succeeded as a cultural export. One is Japanomania. A record 43m foreign holidaymakers visited the country in 2025. Over several decades, Japan has established itself as a trendsetter and taste-maker in everything from video games to fashion and design, meaning that where the Japanese lead, others will follow.
The second reason is that listening bars and ongaku kissa offer an appealing alternative to nightclubs, which are in decline. These bars provide a relaxed, stylish environment for retired ravers and youngsters alike. Here, enjoying music need not mean enduring punishing opening hours (and punishing hangovers).
The third reason is that these bars offer a distinctive, curated experience to suit a wide variety of tastes. Jim Hanmer, the co-founder of Jazu, another listening bar in London, made a pilgrimage to Japan to visit several jazz kissa. ("Jazu" is a nod to the Japanese pronunciation of "jazz"). He was drawn to the venues that reflected the owner's lifelong passion. In these places, "you could listen to a record that you've heard 30 times before, but because you're in this unique environment, it sounds totally different." At a listening bar, even an old track can find a new groove. ■
72 Culture
American history
Finding Founding Fathers
Two new biographies resurrect seminal but forgotten revolutionary figures
The Lost Founder. By Jesse Wegman.
Celadon Books; 384 pages; $30
Radical Duke. By Danielle Allen.
Liveright; 464 pages; $37.50. W.W. Norton; £28
A STROLL THROUGH the gallery of America's Founding Fathers yields familiar faces. Here is brave, disciplined George Washington alongside sociable, inventive Benjamin Franklin. There is John Adams—principled and prickly—squabbling with brilliant, curious Thomas Jefferson. Two new books seek to draw your gaze to two forgotten figures, each with strikingly expansive views of democracy.
The founders, as a whole, were inconsistent. They allowed slavery to endure. They wanted out from under Britain's thumb, but not universal suffrage. Roger Sherman, the only person to sign all four of the founding documents, felt "the people...should have as little to do as may be about the government." Benjamin Rush, another founder, called democracy "the devil's own government".
Charles Lennox (pictured left) and James Wilson (pictured right) disagreed. They shared a faith in democracy and popular wisdom uncommon at the time. Their lives were markedly different: Lennox was an aristocrat and parliamentarian, while Wilson was a poor Scot who emigrated to America and, on the strength of his intel-
intellect, became consequential.
The "tall, rich and beautiful" Lennox was an unlikely radical. Yet he was the first parliamentarian to propose letting all adult men vote in Britain and the first member of the House of Lords to suggest recognising American independence. He was also a patron of Thomas Paine, a tax collector in Sussex who became a vociferous proponent of independence after emigrating to America. Paine's pamphlet, "Common Sense", argued that Americans "have it in our power to begin the world over again". Widely read, it galvanised support for revolution.
In "Radical Duke", Danielle Allen, a political philosopher at Harvard, makes the novel argument that Lennox and Paine were at the centre of "a network of writers" that formed Junius, a pseudonymous author. Junius published essays critical of King George III's government that roiled British politics in the early 1770s; they argued that royal "authority rested upon the consent of the governed". Both men "cut their teeth as radicals by playing Junius" several years before the publication of "Common Sense".
Lennox's radicalism exemplifies Ms Allen's shrewd broader argument that the roots of the American Revolution lay not in Boston but in Britain. Protests against unpopular and unjust taxation, popular demonstrations against royal authority, the
publication of essays promoting revolution: all of these happened in Britain in the 1760s-70s. "Every act that provided the drumbeat of the march to revolution in the colonies had been foreshadowed in Britain," she writes. That argument may provoke some grumbles: Americans consider their revolution home-grown. But few geopolitical events happen in isolation.
Ms Allen began researching Lennox's life after she discovered a copy of the Declaration of Independence in a record office in Sussex. It had belonged to Lennox and was probably commissioned by James Wilson—the subject of Jesse Wegman's biography. Wilson had emigrated to America in 1765, when he was 23, and was the only signer of both the constitution and the Declaration of Independence to serve as a Supreme Court justice.
Mr Wegman, a journalist, situates Wilson within the Scottish Enlightenment, which linked commerce with goodness—a comfortable pairing for mercantile America. ("The road to virtue and that to fortune," wrote Adam Smith, "are, happily in most cases, one and the same.") Scottish nobles had long defended self-government; in 1320 barons declared their king's authority stemmed from "the due consent and assent of us all".
In the run-up to the revolution, Wilson's rigorous essays provided a legal argument for independence. Parliament had no authority over the American colonies, he stated, because colonists did not elect them. "All men are, by nature, equal and free," so "no one has a right to any authority over another without his consent."
Mr Wegman suggests that Jefferson paraphrased this sentiment in the Declaration of Independence. He notes that Wilson was the only founder to use the phrase "created equal" in later writing and that he coined the constitution's opening phrase ("We the people"). In his battles for direct elections for Congress and the presidency, Wilson was "by a long shot the most democratic of all the founders".
Why, then, is he not better known? First, because he was a creature of the law, not a politician. And second, alas, his roads to fortune and virtue diverged. Addicted to land speculation, he ended up in a debtors' prison in New Jersey while a justice. Then, to escape his creditors, he fled to North Carolina, where he died of malaria in 1798.
Were Wilson and Lennox failures because their ideals did not carry the day? That is a crabbed view of reform. Battles for social progress are rarely won in fighters' lifetimes. Martin Luther King did not live to see a racially equal society; he was not wrong to fight for one. In their efforts to expand the franchise and their faith in democracy, Lennox and Wilson were both ahead of their time—and their stories are a welcome rediscovery today. ■
Remember them?
73
Economic & financial indicators
| Gross domestic product | Consumer prices | Unemployment rate | Current-account balance % of GDP, 2026† | Budget balance % of GDP, 2026† | Interest rates 10-yr giv't bonds latest, % | change on year ago, bp | Currency units | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| % change on year ago latest | quarter* | 2026† | % change on year ago latest | 2026† | % | % | per $ | % change on year ago | |||||
| United States | 2.1 | Q2 | 1.5 | 3.4 | Jul | 3.2 | 4.1 | Jul | -3.5 | -6.5 | 4.7 | 35.0 | - |
| China | 4.3 | Q2 | 3.6 | 0.5 | Jul | 1.0 | 5.2 | Jul** | 3.2 | -5.4 | 1.5 | -16.0 | 6.73 |
| Japan | 0.7 | Q2 | 1.1 | 1.7 | Jun | 1.7 | 2.5 | Jun | 4.0 | -1.8 | 2.9 | 129 | 158 |
| Britain | 1.2 | Q2 | 1.7 | 2.9 | Jul | 3.2 | 4.9 | May** | -3.6 | -5.1 | 5.0 | 37.0 | 0.73 |
| Canada | -0.1 | Q1 | -0.1 | 3.0 | Jul | 2.5 | 6.4 | Jul | -0.3 | -2.0 | 3.7 | 23.0 | 1.38 |
| Euro area | 1.0 | Q2 | 1.8 | 2.9 | Jul | 2.8 | 6.3 | Jun | 2.3 | -3.4 | 3.3 | 51.0 | 0.86 |
| Austria | 0.9 | Q1 | 0.8* | 2.7 | Jul | 3.0 | 5.7 | Jun | 1.1 | -4.1 | 3.5 | 41.0 | 0.86 |
| Belgium | 0.5 | Q2 | nil | 3.7 | Jul | 3.6 | 6.3 | Jun | -2.8 | -5.4 | 3.9 | 55.0 | 0.86 |
| France | 0.7 | Q2 | 0.7 | 2.4 | Jul | 2.3 | 8.2 | Jun | -0.4 | -5.3 | 4.1 | 65.0 | 0.86 |
| Germany | 0.9 | Q2 | 0.9 | 2.8 | Jul | 2.7 | 3.9 | Jun | 4.5 | -4.2 | 3.3 | 51.0 | 0.86 |
| Greece | 2.0 | Q1 | 1.0 | 2.7 | Jul | 3.7 | 8.0 | Jun | -5.5 | 0.8 | 3.9 | 47.0 | 0.86 |
| Italy | 1.0 | Q2 | 0.8 | 2.9 | Jul | 2.6 | 5.7 | Jun | 0.8 | -3.0 | 4.1 | 48.0 | 0.86 |
| Netherlands | 1.3 | Q2 | 1.6 | 3.0 | Jul | 3.1 | 4.0 | Jul | 9.3 | -2.6 | 3.3 | 41.0 | 0.86 |
| Spain | 2.7 | Q2 | 2.8 | 3.9 | Jul | 3.3 | 10.1 | Jun | 2.4 | -2.5 | 3.6 | 37.0 | 0.86 |
| Czech Republic | 2.2 | Q1 | 0.8 | 1.7 | Jul | 2.0 | 3.0 | Q2* | 0.3 | -2.6 | 4.9 | 60.0 | 20.7 |
| Denmark | 4.7 | Q2 | 1.1 | 1.7 | Jul | 2.0 | 3.1 | Jun | 12.1 | 1.3 | 3.0 | 43.0 | 6.40 |
| Norway | 1.7 | Q1 | 1.4 | 3.0 | Jul | 3.2 | 4.5 | May** | 14.8 | 9.0 | 4.4 | 51.0 | 9.33 |
| Poland | 3.8 | Q2 | 3.6 | 3.0 | Jul | 2.8 | 5.8 | Jun* | -1.4 | -6.9 | 5.9 | 49.0 | 3.70 |
| Russia | 1.3 | Q2 | na | 6.0 | Jul | 6.4 | 2.2 | Jun* | 2.2 | -3.2 | 16.0 | 269 | 85.1 |
| Sweden | 2.5 | Q2 | 5.7 | 0.2 | Jul | 1.7 | 9.9 | Jun* | 5.5 | -2.3 | 3.1 | 67.0 | 9.44 |
| Switzerland | 0.5 | Q1 | 2.6 | 0.4 | Jul | 0.5 | 3.1 | Jul | 6.9 | 0.2 | 0.5 | 13.0 | 0.80 |
| Turkey | 2.5 | Q1 | 0.5 | 31.8 | Jul | 32.1 | 7.4 | Jun* | -3.0 | -3.6 | 32.4 | 325 | 47.9 |
| Australia | 2.5 | Q1 | 1.1 | 3.8 | Jun | 3.7 | 4.5 | Jul | -3.7 | -1.5 | 5.0 | 77.0 | 1.40 |
| Hong Kong | 4.3 | Q2 | -2.5 | 1.9 | Jun | 2.3 | 3.7 | Jun** | 4.6 | -2.5 | 3.6 | 47.0 | 7.84 |
| India | 7.8 | Q1 | 6.4 | 4.4 | Jul | 4.8 | 6.2 | Jul | -1.6 | -4.5 | 6.8 | 31.0 | 95.5 |
| Indonesia | 5.3 | Q2 | 4.4 | 2.9 | Jul | 3.5 | 4.7 | Feb* | -1.6 | -3.5 | 7.0 | 66.0 | 17.800 |
| Malaysia | 6.0 | Q2 | 9.3 | 1.8 | Jul | 1.9 | 3.0 | Jun* | 3.1 | -3.5 | 3.8 | 38.0 | 4.06 |
| Pakistan | 4.8 | 2026** | na | 9.2 | Jul | 9.0 | 6.9 | 2025 | -1.1 | -4.7 | 12.1 | 11.0 | 278 |
| Philippines | 2.3 | Q2 | 2.4 | 6.2 | Jul | 6.8 | 4.7 | Q2* | -4.0 | -6.3 | 7.3 | 131 | 61.8 |
| Singapore | 5.9 | Q2 | 5.7 | 1.9 | Jun | 2.0 | 2.0 | Q2 | 15.6 | 1.3 | 2.4 | 41.0 | 1.27 |
| South Korea | 3.7 | Q2 | 2.5 | 2.8 | Jul | 2.8 | 2.6 | Jul* | 13.6 | -1.9 | 4.4 | 157 | 1.389 |
| Taiwan | 12.9 | Q2 | 5.7 | 2.5 | Jul | 2.0 | 3.3 | Jun | 26.6 | 1.8 | 1.9 | 50.0 | 31.9 |
| Thailand | 1.9 | Q2 | -0.7 | 1.9 | Jul | 2.3 | 0.9 | Jun* | nil | -5.5 | 2.1 | 79.0 | 32.9 |
| Argentina | 2.3 | Q1 | 3.0 | 33.8 | Jul | 32.4 | 7.8 | Q1* | -0.4 | 0.1 | na | na | 1.497 |
| Brazil | 1.8 | Q1 | 4.5 | 4.4 | Jul | 4.8 | 5.4 | Jun*** | -2.6 | -7.3 | 14.8 | 74.0 | 5.18 |
| Chile | -0.2 | Q2 | -0.1 | 3.5 | Jul | 3.6 | 9.4 | Jun*** | -1.2 | -2.5 | 5.7 | 9.0 | 922 |
| Colombia | 3.4 | Q2 | 5.2 | 6.0 | Jul | 5.7 | 8.0 | Jun* | -2.4 | -6.6 | 12.0 | 23.0 | 3.053 |
| Mexico | 2.2 | Q2 | 6.1 | 3.1 | Jul | 3.8 | 2.8 | Jun | -0.4 | -4.0 | 9.2 | 1.0 | 16.9 |
| Peru | 3.5 | Q1 | 3.2 | 4.1 | Jul | 4.0 | 4.3 | Jul* | 2.2 | -2.4 | 6.2 | 18.0 | 3.36 |
| Egypt | 5.0 | Q1 | -24.0 | 14.9 | Jul | 13.8 | 5.8 | Q2* | -4.7 | -6.0 | 25.6 | 50.0 | 50.7 |
| Israel | 6.9 | Q2 | 15.4 | 1.5 | Jul | 1.9 | 3.1 | Jul | 1.0 | -3.5 | 3.8 | -34.0 | 2.98 |
| Saudi Arabia | 4.6 | 2025 | na | 1.8 | Jul | 2.1 | 3.1 | Q1 | -3.6 | -6.9 | na | na | 3.76 |
| South Africa | 1.9 | Q1 | 2.2 | 4.1 | Jul | 4.5 | 33.6 | Q2* | -0.7 | -4.4 | 8.6 | -101 | 16.1 |
Source: Haver Analytics % change on previous quarter; annual rate †The Economist Intelligence Unit estimate/forecast Not seasonally adjusted †New series Year ending June ††Latest 3 months 3-month moving average *5-year yield ††Dollar-denominated bonds. Note: Euro-area consumer prices are harmonised
Markets
| In local currency | Index Aug 19th | % change on: one week | Dec 31st 2025 |
|---|---|---|---|
| United States S&P 500 | 7,708.0 | -0.5 | 12.6 |
| United States NAS Comp | 26,331.1 | -1.0 | 13.3 |
| China Shanghai Comp | 3,894.4 | -1.3 | -1.9 |
| China Shenzhen Comp | 2,507.3 | -3.8 | -0.9 |
| Japan Nikkei 225 | 65,326.4 | -3.3 | 29.8 |
| Japan Topix | 4,012.3 | -3.1 | 17.7 |
| Britain FTSE 100 | 10,743.4 | -0.8 | 8.2 |
| Canada S&P TSX | 36,401.8 | -0.7 | 14.8 |
| Euro area EURO STOXX 50 | 6,444.5 | -1.4 | 11.3 |
| France CAC 40 | 8,501.9 | -2.0 | 4.3 |
| Germany DAX* | 26,091.3 | -0.9 | 6.5 |
| Italy FTSE/MIB | 52,618.2 | -2.0 | 17.1 |
| Netherlands AEX | 1,102.4 | -0.9 | 15.9 |
| Spain IBEX 35 | 19,847.5 | -1.8 | 14.7 |
| Poland WIG | 152,484.6 | -1.0 | 30.1 |
| Russia RTS, $ terms | 803.9 | -8.0 | -27.5 |
| Switzerland S&P | 14,386.6 | -0.4 | 8.4 |
| Turkey BIST | 14,459.0 | 2.5 | 28.4 |
| Australia All Ord. | 9,255.2 | -1.6 | 2.6 |
| Hong Kong Hang Seng | 25,495.1 | 0.2 | -0.5 |
| India BSE | 76,909.7 | -1.4 | -9.8 |
| Indonesia IDX | 6,394.1 | 0.3 | -26.1 |
| Malaysia KLSE | 1,731.3 | -0.6 | 3.0 |
| Index Aug 19th | % change on: one week | Dec 31st 2025 | |
|---|---|---|---|
| Pakistan KSE | 176,846.4 | -1.9 | 1.6 |
| Singapore STI | 5,694.2 | -0.5 | 22.6 |
| South Korea KOSPI | 6,471.2 | -1.6 | 53.6 |
| Taiwan TWI | 44,719.4 | -1.8 | 54.4 |
| Thailand SET | 1,609.8 | -0.2 | 27.8 |
| Argentina MERV | 2,874,493.0 | -4.2 | -5.8 |
| Brazil BVSP* | 167,830.3 | 0.2 | 4.2 |
| Mexico IPC | 63,999.3 | -2.7 | -0.5 |
| Egypt EGX 30 | 54,512.7 | -1.0 | 30.3 |
| Israel TA-125 | 4,040.0 | -0.6 | 10.3 |
| Saudi Arabia Tadawul | 10,925.7 | 0.8 | 4.1 |
| South Africa JSE AS | 116,051.1 | 0.9 | 0.2 |
| World, dev'd MSCI | 4,975.5 | -0.7 | 12.3 |
| Emerging markets MSCI | 1,668.4 | -0.8 | 18.8 |
US corporate bonds, spread over Treasuries
| Dec 31st 2025 | Last | Dec 31st 2025 | |
|---|---|---|---|
| Basis points | 95 | 93 | |
| Investment grade | 323 | 354 |
Sources: LSEG Workspace; Moscow Exchange; Standard & Poor's Global Fixed Income Research *Total return index
Commodities
| The Economist commodity-price index | % change on year | |||
|---|---|---|---|---|
| 2020=100 | Aug 11th | Aug 18th* | month | |
| Dollar Index | ||||
| All items | 152.1 | 153.0 | nil | 14.0 |
| Food | 149.1 | 152.9 | -1.6 | 3.6 |
| Industrials | ||||
| All | 154.5 | 153.1 | 1.5 | 24.3 |
| Non-food agriculturals | 151.3 | 151.4 | -0.4 | 21.4 |
| Metals | 155.4 | 153.5 | 1.9 | 25.1 |
| Sterling Index | ||||
| All items | 144.7 | 145.2 | -1.2 | 13.7 |
| Euro Index | ||||
| All items | 150.5 | 151.0 | -1.5 | 14.9 |
| Gold | ||||
| $ per oz | 4,392.1 | 4,360.6 | 7.0 | 31.2 |
| Brent | ||||
| $ per barrel | 89.0 | 91.0 | -0.1 | 38.3 |
Sources: CME Group; LME; LSEG Workspace; NOREXECO; NZ Wool Services; S&P Global Commodity Insights; Thompson Lloyd & Ewart; USDA *Provisional
For historical indicators data, visit economist.com/economic-and-financial-indicators
74
Mary de Rachewiltz
The poet, translator and tireless defender of her father, Ezra Pound, died on July 24th, aged 101
She was still awake when the tall bearded man came into her bedroom to say goodnight. She had hoped he would. Though she was nine, it was a rare treat to get a visit from her father. By the light of the kerosene lamp, he looked at her hands. Her fingernails were filthy. Then he started to rummage in his luggage, for sweets, she hoped. Instead he brought out scissors and, carefully, cleaned and cut her nails. He asked her then, where was her toothbrush? It was somewhere. The next day, he took her to buy one.
This happened in Gais, in South Tyrol, in the house of two kind farm-folk, Mamme and Tatte, who cared for abandoned children. But she was not abandoned, quite. Her parents, the poet Ezra Pound and his mistress Olga Rudge, a violinist (Tattile and Mamile to her then), lived in Venice and paid 200 lire a month for her keep. She did not mind the arrangement at all. From birth she had grown into a Tyrolean peasant child, a shepherdess with blonde plaits who had her own flock and, in secret partnership with her father, ran a miniature bee business for a while. But then at fairly regular intervals she would go to her parents in Venice, to a villa full of books in many languages, to the Lido via vaporetto and to feasts of ice cream. Her cool, willowy mother made her speak no German there, and wear white gloves. It was a small price to pay.
From the start, her two lives ran in counterpoint. She became skilled at defending one against the other. In a poem she wrote later she relished the Pustertal dialect she spoke, words as weighty as sausages and dumplings. Against city people she defended dung-heaps, insisting that a good one smelled mostly of leaves and moss. Tyroleans hated Italians; but her father, arranger of all her Italian treats, was an American Italophile and she adored him. “Babbo” taught her both Italian and English by making her translate “The Cantos”, the huge and obscure poem he spent his life working on, into Italian. For her, that was like being presented with a basket of exotic food she had no idea what to do with, but she tried. Each typed page she produced for him would come back entirely scribbled over, not only with corrections but with newly perplexing thoughts.
From the mid-1920s Babbo had also formed some controversial political ideas. He began to worship Italy’s fascist dictator, Benito
Mussolini, extolling him in articles and broadcasts and even presenting him with a copy of the early Cantos. (“Ma questo” said the Boss, “è divertente.”) For Pound, Mussolini embodied the strength and order Italy needed. But then he went further. The true threat to mankind was “international Jewry”. Racial purity should be preserved to strengthen the nation. On Italian radio he begged Franklin Roosevelt not to side with the Allies and said he should be hanged; those broadcasts earned him a charge of treason in absentia. When American troops eventually liberated Italy, in 1945, they captured Pound and caged him. For 13 years he was kept in a locked cell in a psychiatric hospital.
Mary now had two battles to fight. The first was to defend her father against the charge of treason and the imputation of insanity. The latter was clearly ridiculous, as he wrote her long, lucid letters from prison and continued working on “The Cantos”. She managed only one visit; otherwise she imagined him, still tall and dignified in the dining room among the slobbering inmates. His broadcasts and articles, though, were much harder to dismiss. What could she say of his vile remarks about “Jew slime” and “the dog-damn wop”? Only that he had been influenced by the opinions of the time. T.S. Eliot had been antisemitic, too. (She went to see shy, stooping Tom to ask for help; he offered her thin Sawia biscuits, Loneliness personified.) And obviously, later on, the “master of utterance” had lost his grip on words.
That was her second fight. Babbo had been carried away by words all his life. Before the debacle in Italy he had been ranked among the foremost poets of the age; now he was untouchable. But that could not be. She worked on with “The Cantos” and with his other poetry, turning everything into Italian and insisting on the beauty in it. “Beauty is difficult,” Babbo had written in Canto LXXIV; but, like light, it kept breaking through. In her memoir “Discretions”, in 1971, she cited his poetry whenever it could illuminate a phase of her life, or a memory. Her own poetry was a homage to the man who had trained her in it, not least on those evenings when the air had tingled, threatening explosion, as he recited his own. In 1985, 13 years after his death, she at last published the first dual-language edition of “The Cantos”.
She still had more to do. She preserved his papers and first drafts, most of which went to Yale, and lectured about him all over North America. Occasionally she unveiled long-resisted plaques to him. Most important, she had to keep up the re-evaluation. After her marriage in 1946 to a Russian-Italian minor nobleman, Boris de Rachewiltz, they bought Brunnenberg, a crag-bound castle in her beloved South Tyrol. It was a ruin when they found it, but over the years she turned it into a centre of Pound studies. Scholars from the world over came to discuss him amid the splendour of the Alps. She firmly told them, when they arrived, that they could say anything they liked about Pound. And then she would tell them what he really thought, and why.
Babbo himself had stayed there after his release from the asylum, hoping for peace in which to write “Paradise”, intended as the final Canto (with a nod to Dante). But he found the place too cold, and retreated again to Venice with Olga. Mary stayed, because South Tyrol was home. After her idyllic childhood in Gais, she had returned again and again to see Mamme and Tatte. Mamme, with her emotional, fearful Catholic heart and overflowing kitchen, offered true mother-love. But there could be only one father. “I have tried to write Paradise”, he declared, in his “Notes for Canto CXX”: “...let those I love try to forgive/ what I have made.” ■
CLASSIFIED
75
Your Hassle-Free Mayfair Pied-à-Terre
47 Park Street℠ offers a selection of luxuriously furnished, spacious apartments in Mayfair for the amount of time you personally require in London each year — at a fraction of the cost of whole ownership.
Residential Benefits Include:
- Fully integrated property management
- 24-hour concierge, maid, and room service
- Valet storage facility with wide usage options
- No stamp duty
- Individual and corporate fractional interests from £82,800
47 PARK STREET℠
Mayfair – London
GRAND RESIDENCES BY MARRIOTT®
Open daily, 9 a.m.–5 p.m.
Tel: +44 (0) 20 7950 5528
Email: enquiries@47parkstreet.com
www.47parkstreet.com
Furnishings may vary. Residence and property amenities are subject to change. Data protection: Your personal information and details will be stored by Marriott Vacation Club International, part of a global group of affiliated companies ("Affiliates"), and your personal information may be shared among the Affiliates and transferred outside of your country of residence. Personal information that is transferred outside of the European Economic Area is done under data transfer agreements that contain standard data protection clauses adopted by the European Commission that provide safeguards for such transfers. For more information about the use of your personal information, or to exercise your rights of opposition, access, rectification and deletion please review our Global Privacy Statement found at marriottvacationclub.com/privacy/ or contact our Privacy Office at privacy@nvwc.com. IQL 20-002. THIS ADVERTISING MATERIAL IS BEING USED FOR THE PURPOSE OF SOLICITING THE SALE OF FRACTIONAL RESIDENCE CLUB MEMBERSHIP. Subject to applicable terms and conditions. This is neither an offer to sell nor a solicitation to buy to residents in jurisdictions in which registration requirements have not been fulfilled or where marketing or sale of fractional residence club membership is prohibited and your eligibility and the membership clubs available for purchase will depend upon the jurisdiction of your residency. Prices are subject to change. Key information is available upon request by contacting 47 Park Street Grand Residences by Marriott, Mayfair, London, W1X 7EB, United Kingdom or email moreinfo@47parkstreet.com. Marriott Vacation Club International and the programs and products provided under the Grand Residences by Marriott brand are not owned, developed, or sold by Marriott International, Inc. Marriott Vacation Club International uses the Marriott marks under license from Marriott International, Inc. and its affiliates. ©2026 Marriott Vacation Club International. All Rights Reserved.
STRONGKEY
NOW DELIVERING RANSOMWARE DETERRENCE
GETSECURE@STRONGKEY.COM
The Economist
To advertise within the classified section, contact:
UK/Europe
Agne Zurauskaite
agnezurauskaite@economist.com
North America
Richard Dexter
richarddexter@economist.com
Asia
Chris Phang
chrisphang@economist.com
Readers are recommended
to make appropriate enquiries and take appropriate advice before sending money, incurring any expense or entering into a binding commitment in relation to an advertisement.
The Economist Newspaper Limited shall not be liable to any person for loss or damage incurred or suffered as a result of his/her accepting or offering to accept an invitation contained in any advertisement published in The Economist.®
Experience life in five European dream homes, effortlessly.
Discover a new way to enjoy second homes, without the hassle of traditional ownership.
augustcollections.com
@august.collections
Explora
JOURNEYS
*Maybe the best hotel
isn't a hotel.*
EXPLORE THE OCEAN STATE OF MIND
ALL OCEANFRONT SUITES | IMMERSIVE ITINERARIES | AWARD-WINNING DINING HOLISTIC WELLNESS | REFINED EUROPEAN DESIGN
EXPLORAJOURNEYS.COM
Contact your preferred travel advisor or call: +41 43 5086478
制度与权力批判深度研判：The Economist (2026-08-24)
▌ 本日全报思想与文化深思雷达 (Intellectual & Cultural Significance Radar)
在系统通盘梳理本期报纸所涉及的所有国际政治、经济、社会与文化报道后，以下 3 个具体事件在制度结构、权力技术与法理合法性层面最值得学者进一步深思：
-
【帕兰提尔（Palantir）的技术渗透与主权边界】
[F12_19 🔍]（版面：技术与主权）- 事件简述：技术公司 Palantir 通过与美国边境执法人员和以色列士兵合作，将大规模数据管理系统嵌入国家安全机构，并发布关于西方公民文化空心化的宣言。
- 制度批判切入点：分析私营技术公司如何通过定义“效率”与“文化危机”来重构公共治理的边界，探讨数据管理工具如何转化为一种实质性的权力行使机制。
-
【民主党初选日程的结构性重组】
[F5_9, F5_10]（版面：专业栏目/政治）- 事件简述：民主党全国委员会调整 2028 年初选日程，将南卡罗来纳州提前，而将传统的艾奥瓦州排除在首批名单之外。
- 制度批判切入点：探讨政党官僚机构如何通过对“程序时间表”的微观操纵，实现对候选人筛选机制的结构性控制，揭示制度设计如何通过地理空间的排斥来预设权力结果。
-
【剑桥大学对杰森·阿尔代的“符号化”处理】
[F15_21, F16_23, F17_25]（版面：专业栏目/学术）- 事件简述：剑桥教授阿尔代自杀，报道指出他在崛起时被支持者“非人化”为多样性符号，在陨落时又被反对者利用为文化战争的图腾。
- 制度批判切入点：剖析学术机构在 DEI（多样性、公平性与包容性）制度压力下，如何将个体转化为制度指标的载体，揭示官僚主义如何通过符号化管理掩盖对个体独立价值的审视。
▌ 精选核心专题深度思想论证 (In-Depth Dialectical Monograph)
专题一：技术利维坦的隐形殖民：Palantir 与治理逻辑的“算法化”
在当代治理逻辑中，权力不再仅仅通过法律条文下达，而是通过算法架构与数据库管理实现。Palantir 与美国边境执法及以色列军方的合作 [F12_19 🔍]，标志着一种新型的“技术-主权”共生关系的出现。从制度现实看，现代国家在处理海量安全数据时，倾向于依赖私营公司的技术能力。然而，这种依赖导致了严重的权力异化：当公共治理的底层逻辑由私营公司的软件定义时，治理的透明度可能被“商业秘密”所取代。
从话语技术剖析，Palantir 巧妙地利用了关于“西方公民文化空心化”的叙事 [F12_19 🔍]，将自身包装成文明的守护者，从而使其对数据的全视之眼（Panopticon）获得了某种道德合法性。这种策略将“监控”转化为“保护”，将“数据抓取”转化为“文明维护”。援引施密特的“例外状态” (State of Exception) 理论，这种运作模式实际上是在常规法律框架内创造了一个“技术例外区”，使得私营公司在实质上参与了主权权力的行使。
这种制度演进揭示了平台资本主义对主权国家的深层渗透：国家在某些领域不再是技术的使用者，而成为了技术基础设施的租户。当权力的行使标准从“法理正义”转向“算法优化”时，未来的制度危机将不再表现为政权的更迭，而表现为底层代码的失效或访问权限的被剥夺。
专题二：符号化治理的暴力：DEI 制度下的个体消解与学术官僚主义
剑桥大学教授杰森·阿尔代的悲剧 [F15_21, F17_25]，提供了一个分析现代学术机构如何通过“符号化管理”实施隐蔽制度暴力的样本。在制度现实中，顶尖学术机构在面对 DEI（多样性、公平性与包容性）压力时，容易陷入“指标崇拜”与“实质审查”的脱节。
在这种机制下，阿尔代在崛起期间被支持者“非人化” [F16_23 🔍]，他不再是一个独立的学术个体，而被制度性地转化为一个“符号”——他是机构多样化努力的可见证明。这种权力运作采用了典型的福柯权力谱系学 (Foucauldian Genealogy of Power) 逻辑：权力通过定义“谁是合格的多样性代表”来实施分类与管理。当阿尔代陨落时，制度的反应同样是符号性的：反对者迅速将其转化为“DEI 失败”的反面教材 [F16_23 🔍]。个体在整个过程中被彻底消解，其真实的学术能力与人格缺陷被掩盖在两种对立的政治叙事之下。
这揭示了韦伯意义上的“官僚制铁笼” (Iron Cage of Bureaucracy) 的当代变体：当多样性被量化为 KPI，制度本身就产生了一种排斥真实个体的惯性。大学在任命时的失效，并非简单的行政疏忽，而是制度在追求符号合法性时，主动选择性地忽略了对个体独立性的审视。最终，这种制度暴力不仅摧毁了被符号化的个体，也通过将少数群体学者置于“符号化”的高压之下，反向损害了其声称要保护的对象 [F16_23 🔍]。
▌ 面向学者的制度研究开放性追问 (Open Research Horizons)
- 【关于技术主权的解构】：当国家核心安全功能深度依赖于私营算法公司时，传统的“主权”概念是否已在事实上被“基础设施权力”所部分取代？
- 【关于符号化治理的反思】：在现代组织管理中，当“多样性”从伦理追求转化为制度指标（KPI）时，这种转化如何导致个体被再次非人化？
- 【关于程序正义的微观操纵】：民主党初选日程的调整
[F5_10 🔍]揭示了程序设计如何成为一种权力过滤机制，这种“程序性权力”在现代政党政治中如何影响民主的实质？
理性与情感辩证深度研判：The Economist (2026-08-24)
▌ 本日全报生活世界痛感与情感政治深思雷达 (Lifeworld & Affective Significance Radar)
在系统通盘梳理本期报纸所涉及的所有报道后，以下 3 个具体事件在个体生命体验、情感撕裂与生活世界崩解层面最值得学者进一步深思：
-
【AI意识的“人格幻象”与道德焦虑】
[F1_2, F7_13, F7_14]（生活领域：技术与存在）- 事件简述：关于AI能否产生意识的讨论持续升温。教皇利奥（Pope Leo）在通谕中明确划清界限，认为AI系统不会经历体验，也不会感受到喜悦或痛苦
[F7_14 🔍]。 - 情感政治切入点：探讨人类在面对日益强大的AI时，如何通过赋予其“人格”来寻求连接，以及这种投射与宗教/哲学界坚持的“非人格化”理性定义之间的冲突。
- 事件简述：关于AI能否产生意识的讨论持续升温。教皇利奥（Pope Leo）在通谕中明确划清界限，认为AI系统不会经历体验，也不会感受到喜悦或痛苦
-
【学术多样性理想与个体悲剧的撕裂】
[F15_21, F16_23, F17_24, F17_25]（生活领域：学术与认同）- 事件简述：剑桥教授杰森·阿尔代（Jason Arday）自杀身亡
[F15_22 🔍]。其生前被支持者非人化为多样性努力的象征，死后又被反对者利用来攻击DEI（多样性、公平性和包容性）政策[F16_23 🔍]。 - 情感政治切入点：分析个体如何被宏观的政治正确或意识形态标签“符号化”，从而导致其作为“完整的人”的生命体验被抹除，揭示了身份政治对个体真实性的异化。
- 事件简述：剑桥教授杰森·阿尔代（Jason Arday）自杀身亡
-
【地缘政治博弈中的肉身代价】
[F3_6, F4_8]（生活领域：战争与生存）- 事件简述：美国与伊朗未能达成和平协议，霍尔木兹海峡袭击导致水手丧生
[F3_6 🔍]；特朗普缩减与韩国的联合军演，理由是成本高昂且向朝鲜发出敌对信号[F4_8 🔍]。 - 情感政治切入点：分析大国之间基于“成本-收益”的理性计算（Logos）如何直接转化为前线个体（如水手）的生命消亡（Pathos），揭示了战略理性对个体生存权的绝对支配。
- 事件简述：美国与伊朗未能达成和平协议，霍尔木兹海峡袭击导致水手丧生
▌ 精选核心专题理性与情感辩证论证 (The Dialectic of Logos and Pathos Monograph)
专题一：符号化生存与主体性的消亡：杰森·阿尔代悲剧中的身份政治辩证法
在杰森·阿尔代（Jason Arday）的案例中，我们观察到一种极端的 Logos（意识形态理性）与 Pathos（个体生命痛感） 的错位。从系统理性来看，阿尔代教授被置于一个关于“多样性、公平性和包容性（DEI）”的宏观叙事中 [F17_24 🔍]。在这种逻辑下，他不再是一个具体的学者，而是一个代表某种社会进步或制度改革的“符号”。
这种符号化过程在阿尔代崛起时表现为支持者的“非人化”推崇，而在其陨落时则转化为反对者的政治工具 [F16_23 🔍]。这里的辩证摩擦在于：旨在通过DEI政策实现“包容”的理性设计，在实际操作中却可能导致个体被剥夺其作为独立个体的复杂性。当一个人被定义为“多样性的象征”时，他真实的优点与缺陷被掩盖，其主体性被意识形态的标签所取代。
从现象学视角看，阿尔代的自杀 [F15_22 🔍] 是这种异化的终极体现。当个体发现自己无论在成功还是失败时，都被他人视为某种“事物的象征”而非“完整的人”时 [F17_25 🔍]，其存在感在社会镜像中彻底崩塌。这揭示了当代学术与社会治理中的一个悖论：我们试图通过理性的分类和指标来保障公平，但这种分类本身可能成为一种新的、更隐蔽的非人化暴力。
专题二：技术主权与道德边界的模糊：AI意识争议中的理性定义与情感投射
在关于AI意识的讨论中，存在着一种深刻的 系统理性（Systemic Logos）与 存在焦虑（Existential Pathos） 的对立。技术理性试图通过功能定义来界定意识，例如教皇利奥认为AI不具备经历体验的能力，因此不应被赋予“人格” [F7_14 🔍]。这种观点将意识视为一种生物性的、不可模拟的特质，旨在通过划清界限来维护人类的独特性。
然而，这种理性定义在面对实际的人机交互时遭遇了挑战。尽管AI在技术上是概率分布的产物，但其表现出的能力（如宇树科技的人形机器人能够表演武术、跳舞 [F6_12 🔍]，以及LLM的交互能力）在用户端制造了一种强烈的“人格幻象”。这种幻象引发了深层的道德焦虑：如果AI能够模拟痛苦，我们是否在无意中创造了某种需要被伦理对待的实体？
这里的辩证冲突在于：理性告诉我们AI是工具，但情感体验告诉我们AI像人。这种温度差揭示了人类主体间性的脆弱性——我们极易将自身的情感投射到非人实体上。当帕兰提尔（Palantir）等公司将数据库管理与某种带有神秘色彩的“全视之眼”叙事结合时 [F12_19 🔍]，技术理性实际上在利用人类的情感恐惧与崇拜来构建权力。AI意识的争议本质上不是关于硅片的计算能力，而是关于人类在技术镜像中如何重新定义“灵魂”与“尊严”的边界。
▌ 面向学者的伦理与社会心理开放性追问 (Open Research Horizons)
- 【符号化暴力与个体救赎】：在DEI等社会工程实践中，如何建立一种机制，既能推动结构性的多样化，又能防止个体被简化为意识形态的符号，从而保障个体的“完整人格”？
- 【模拟共情与伦理脱节】：当AI能够完美模拟情感体验而无需具备真实意识时，人类社会是否会产生一种“伪共情”依赖，从而导致真实人际关系中责任感的进一步消解？
- 【战略理性与肉身痛感的不可通约性】：在国际政治的“成本-收益”计算中，如何将微观个体的生命痛感（Pathos）有效地引入宏观决策（Logos）模型，以修正纯粹的技术官僚主义倾向？