在 2026 年 2 月，科技创业家与投资人 Matt Shumer 在 X 上发布一则爆红贴文，题为「Something Big Is Happening」，主张像 Anthropic 的 Claude Code 与 Claude Cowork 这类 AI agents 很快就会重创专业工作，并呼吁人们每天花 1 小时练习使用 AI 以保持领先。该贴文广泛传播、浏览量突破 80 million，且发布时点正逢金融与软体等被视为容易被 AI 取代的股票出现剧烈抛售。文章将此描绘为另一个「ChatGPT moment」：戏剧化叙事与个人示范（例如快速建网站或自动化传讯）加剧大众焦虑与投资人情绪波动。
该专栏认为，恐慌更多是由轶事而非数据驱动：Shumer 的文章长达 4,783 字，却没有提供任何可量化的证据证明 AI 很快就会取代数以百万计的白领，反而依赖诸如离开一下回来程式码看似就「完成」了、或初阶律师被取代等故事。文中指出市场反复出现由叙事驱动的剧烈反转，例如 2025 年 11 月中旬因「AI bubble」恐惧，道琼指数几乎下跌 500 点；以及 Anthropic 在 2026 年 2 月初发布 11 个 Claude Cowork plugins（其中包含一个用于法律任务）后又出现另一波抛售。在这样的背景下，作者引用相对稳定的指标：全国生产力略有上升但仍在历史区间内；Yale Budget Lab 表示自 ChatGPT 上线以来未见可辨识的广泛劳动市场冲击；一项 METR 随机对照试验发现，资深软体开发者使用 AI 工具完成任务反而多花了 19% 的时间。
文章警告，大胆预测可能出于自利并误导决策，并举出多个高知名度说法，例如 Anthropic CEO Dario Amodei 预测 AI 将在 1 to 5 years 内消灭一半入门级白领工作，以及微软 AI 负责人 Mustafa Suleyman 表示「most if not all」专业任务可能在 18 months 内被自动化。文章将这些言论与组织行为对照：Harvard Business Review 对超过 1,000 位高阶主管的调查发现，许多裁员是为了迎接 AI 而提前进行，但只有 2% 表示裁员是因为实际导入 AI；并引用 Klarna 尝试用 AI 取代 700 名客服人员，之后因品质下降而需要重新雇用人类。结论呼吁优先依靠可衡量的证据、追踪生产力与招募率等具体指标，并预期 AI 对劳动的影响将不均且渐进，呼应 2000 年代初 dot-com 热潮虽具变革性但实现时间比预测更久的经验。
In February 2026, a viral X post by tech entrepreneur and investor Matt Shumer titled “Something Big Is Happening” argued that AI agents like Anthropic’s Claude Code and Claude Cowork will soon decimate professional jobs, urging people to practice using AI for 1 hour a day to stay ahead. The post spread widely, surpassing 80 million views, and landed amid sharp market selloffs in finance and software stocks seen as vulnerable to AI replacement. The piece frames this as another “ChatGPT moment,” where dramatic narratives and personal demos of fast website-building or automated messaging intensify public anxiety and investor mood swings.
The column argues the panic is being driven more by anecdotes than data: Shumer’s essay runs 4,783 words yet offers no quantifiable evidence that AI will displace millions of white-collar workers soon, relying instead on stories like code appearing “finished” after stepping away and junior lawyers being replaced. It points to repeated narrative-driven whiplash in markets, including the Dow falling nearly 500 points in mid-November 2025 on “AI bubble” fears and another selloff after Anthropic released 11 Claude Cowork plugins in early February 2026, including one for legal tasks. Against that backdrop, the author cites relatively stable indicators: national productivity is up slightly but within historical ranges, Yale Budget Lab reports no discernible broad labor-market disruption since ChatGPT’s launch, and a METR randomized controlled trial found experienced software developers took 19% longer to complete tasks when using AI tools.
The piece cautions that bold predictions can be self-serving and mislead decision-making, highlighting high-profile claims such as Anthropic CEO Dario Amodei forecasting AI will wipe out half of entry-level white-collar jobs within 1 to 5 years and Microsoft AI head Mustafa Suleyman saying “most if not all” professional tasks could be automated within 18 months. It contrasts rhetoric with organizational behavior: a Harvard Business Review survey of more than 1,000 executives found many layoffs were made in anticipation of AI, while only 2% said they cut jobs because of actual AI implementation, and it cites Klarna’s attempt to replace 700 customer-service staff that later required rehiring humans after quality declined. The takeaway is a call to prioritize measurable evidence, track concrete metrics like productivity and hiring rates, and expect AI’s labor impact to be uneven and gradual, echoing how the early-2000s dot-com hype proved transformative but took longer than predicted.