


一 | "The actual intentions of Pelosi’s visit to Taiwan are inevitably sophisticated," says Dr. Chang Ching, a research fellow from the Taiwan-based Society for Strategic Studies and a lead military expert on the People's Liberation Army and regional security in Taiwan.,"To acquire a historical reputation is the obvious factor identified by many politicians. Delivering political signals to Beijing can also be another reason. Attempting to shape the political developments in Taiwan has been accused by certain political commentators as well. Nonetheless, the Chinese media basically define her maneuver as an 'impulsively sneaky visit,' which is a term generally applied to those people who are defined as politically hostile to the Beijing regime visiting certain territories protested by the People’s Republic of China," Dr. Chang continued.,Pundits have speculated about the possibility of Pelosi visiting Taiwan and China's reaction since April 2022; Beijing has clearly demonstrated its fierce opposition to the high-profile visit. After her tour was canceled in the spring after the House speaker reportedly got COVID-19, the debate got its second wind in July.,Pelosi neither officially confirmed nor denied her intent to visit Taipei. Likewise, the White House tried to distance itself from the Democratic politician's trip to Taiwan, while at the same time arguing that such a tour is by no means "uncommon" and that the US "shouldn’t be intimidated by [China's] rhetoric or... potential actions.",Pelosi Provocation Over Taiwan Leads US, World Into Age of Disorder And Instability: Martin Jacques10:27 GMT,"The worst indication reflected by Pelosi’s visit is the decline of the United States that no American politicians are willing to admit," says the Chinese scholar. "If Washington still keeps sufficient confidence to condition Beijing’s reactions to this issue, President Biden himself and his senior members of the government would not show their reluctance to support Pelosi’s visit. The differences between the executive branch and the legislative branch of the United States government on this visit are so clear to see.",Dr. Chang wonders why Pelosi chose to keep her Taiwan visit secret until the last moment. He notes that the House speaker cited "security concerns" over her travel plan.,"Actually, it is a matter of Pelosi’s integrity and also causing more speculations on the so-called American security commitments to Taiwan," says the scholar. "If the Speaker herself has no full confidence of security during her presence in Taiwan, how can people in Taiwan trust the United States is capable of supportive while Taiwan is in danger?",According to Dr. Chang, China has long been very clear about its "red lines" with regard to Taiwan. Therefore, "certain serious consequences can be expected," he notes, ruling out, however, a "full scale invasion or other form of armed conflict.",The scholar points out that there are many measures available in Beijing's toolbox towards Taiwan for conditioning Taipei’s political maneuvers. What's more, the Taipei leadership is "well aware of how damaging [Pelosi's] visit can be" for the island, he adds.,"What can the possible consequences be? It all depends on how Taipei will react to those measures towards Taiwan," says Dr. Chang. "After all, it is literally an interactive process. It takes two to tango. Whether it will turn into a negative spiral or on the contrary to a positive spiral is not merely decided by Beijing. Taipei’s reactions are also important. Also, Beijing will not only try to condition Taipei like many western media [outlets] expected but also press Washington at the same time. The possibility of any mutual cooperation between Beijing and Washington on any global issues has basically vanished in the near future. The difficulties of reconciliation by these two powers will also increase dramatically.",All the aforementioned negative consequences will be unquestionably associated with Democratic House Speaker Nancy Pelosi and her ill-fated decision to visit Taiwan, according to the scholar.。

新智元报道
AI 没有抢走任何人的工作。至少到目前为止没有。这个结论来自 Anthropic 经济研究负责人 Peter McCrory 近日发布的一篇长文。

https://x.com/PeterMcCrory/article/2079979321607745905经质量调整的 AI 产出在 2024 和 2025 年均增长超过 2000%,约五分之一的美国企业已在业务中使用 AI,但美国 6 月失业率仍然是 4.2%,恰好落在充分就业水平上。McCrory 的核心判断是:AI 目前仍是一种「技能偏向型、劳动力增强型」技术,它在放大人的能力,而不是替代人。

二 | 他同时给这个判断标了有效期:12 个月。

数据不撒谎:
连 AI 高暴露岗位都没事McCrory 拉了一整条数据链来证明美国劳动力市场有多稳:失业率 4.2%,职位空缺与失业人数之比刚过 1,25 至 54 岁就业比例接近数十年高位,初请失业金人数连续四年低位运行。他认为,总体而言,美国就业形势稳中向好。总量数据只是起点,分组对照才见真章。

三 | McCrory 与同事 Maxim Massenkoff 此前做过一项分析:把 Claude 被大量用来自动化的任务所对应的岗位单独拎出来,看它们的失业率是否比其他岗位恶化更快。答案是没有。用最新的劳动统计局数据更新,结论不变。软件工程类职位的招聘帖自 2025 年 5 月起反弹,涨幅甚至高于其他岗位的平均水平。唯一值得留意的是年轻劳动者。McCrory 承认,年轻人在 AI 高暴露岗位上求职确实变难了,斯坦福数字经济实验室的论文提供了类似证据。但他也指出一个容易被忽略的背景:2022 年至今,美国经历了有记录以来最大幅度的非衰退式劳动力市场降温,「低招聘、低裁员」格局本就最先挤压早期职业者。把这笔账全算到 AI 头上,证据不够。

为什么没抢走工作:
每个岗位都有 AI 搞不定的环节McCrory 的解释建立在 Anthropic 过去 18 个月多份研究报告的事实基础上,逻辑链条是这样的:第一层事实:在美国劳工部 O*NET 的全部职业分类中,没有任何一个岗位的所有关联任务被 Claude 系统性地完成。每个岗位都存在 AI 搞不定的环节:人际协调、面对面互动、与物理世界打交道。这些环节像木桶的短板,既限制了 AI 对整体生产力的提升幅度,也让能补上短板的人变得更值钱。第二层事实:AI 越强,对操控它的人的要求也越高。

四 | Anthropic 今年 1 月的 Economic Primitives 报告发现,Claude 产出越复杂,背后用户的输入专业度越高。模型做出复杂经济模型时,往往是一个能给出精准指令的专业用户在指挥。
https://www.anthropic.com/research/economic-index-primitives3 月的 Learning Curves 报告显示,用 Claude 六个月之后,用户更多地把它当思维伙伴来用,交互成功率更高。

五 | 如果 AI 已经强到能独立完成工作,这种学习效应不该出现。
https://www.anthropic.com/research/economic-index-march-2026-report第三层事实:即便到了 Agent 时代,人的价值依然稳固。Anthropic 分析了 Claude Code 七个月的使用数据,发现任务价值在持续攀升,但人类专业能力的回报没有下降:用户负责规划,把实现委派给 Claude;专业能力更强的用户,任务成功率更高,Claude 犯错时恢复也更快。纯编程实现的溢价可能正在缩水,但规划、判断和纠错的价值反而上升了。

六 | 一个有趣的佐证来自用户自己的感受。

七 | Anthropic 6 月的经济指数调查里,超过三分之一的受访者预期 AI 在 12 个月内能完成自己大部分工作,但预期自己因此失业的比例明显更低。用 AI 自动化程度更高的那批用户,对薪资和工作安全感反而更乐观。
https://www.anthropic.com/research/economic-index-june-2026-report

三种「奇点」和一道防线McCrory 对当下的判断是乐观的,但他没有回避三种可能打破现有格局的未来场景。它们在学术文献和行业讨论中分别被称为软件奇点、经济奇点和科斯奇点。

八 |
https://x.com/MTSlive/status/2090610614154764509软件奇点指的是 AI 自动化创新本身。过去的通用技术再强大,内燃机造不出新的交通方式;但 AI 被直接赋予了通用认知能力,理论上可以反过来加速自身迭代。McCrory 原文明确讨论了这一场景:递归自我改进(RSI)一旦成立,AI 能力的提升将脱离人类控制的节奏。经济奇点是软件奇点的经济学后果。McCrory 援引 Aghion、Jones 和 Jones 2017 年的模型指出,如果创新过程本身被自动化,标准经济模型会预测出「有限时间内的无限增长」。这个结论听起来像科幻,但它是严格的数学推演。科斯奇点则瞄准另一个维度。NBER 2025 年的一篇论文(Shahidi、Rusak、Manning 等人合著)提出,当 Agent 把搜索、谈判、签约和监督等交易成本压到接近零时,Ronald Coase 1937 年的经典问题就会被重新激活——市场协调的成本一旦消失,企业存在的理由也就动摇了。

九 |
https://www.nber.org/system/files/chapters/c15309/c15309.pdfMcCrory 讨论的 Agent 场景天然与这一框架对话:如果 Agent 不只改变生产方式,还改变组织形态,劳动力市场的影响将远不止「谁丢了工作」这么简单。面对三种奇点,McCrory 给出的防线是同一道:「弱环节」。这个概念来自 Aghion、Jones 和 Jones 的论文——经济增长的上限不由做得最好的环节决定,而由那些不可或缺却难以改进的环节决定。
https://www.nber.org/system/files/working_papers/w23928/w23928.pdf只要存在始终无法被自动化的关键任务,无论是技术瓶颈还是社会约束,劳动收入份额就不会崩塌,企业形态也不会彻底解体。

十 | 编程 Agent 的现状已经在验证这条逻辑。McCrory 引用的数据显示,引入编程 Agent 后代码行数增长了 10 到 20 倍,但软件发布量只增加了 30%,应用实际使用量没有任何增长。代码只是软件生产的一环,测试、架构设计、产品判断这些环节仍然卡在人手上,它们就是当前的「弱环节」。McCrory 给出了明确的预测:一年后美国的失业率不会因为 AI 而明显上升。这个判断的有效期是 12 个月,验证信号也很清晰——观察 AI 高暴露岗位的失业率是否开始偏离其他岗位,以及 Agent 场景中人类专业能力的回报是否开始下降。两个指标同时转向的那天,他自己搭建的这套分析框架也需要被修正。

十一 |
Current article:http://qccuk.liaoenxingunchuanglou.buzz/sjre/c7jancm.html
Published on:12:36:22