文章以 2026 年「六个可怕预测」为轴，从 2025 年 12 月 19 日的观察出发，对比「三年前」到 2023 年 1 月 Google 首次大规模裁员的节点，推测 OpenAI 也可能在 2026 年初出现类似的人力调整。文中强调扩张速度与规模：OpenAI 两年内人数「五倍成长」到约 4,500 人，且作为成立约 10 年的组织，在多线作战与投资重配下，裁员与重组可能被外界视为泡沫降温讯号。
在基础设施层面，作者担心反资料中心建设的社群动员可能被中俄以假讯息「加乘」，而生成式 AI 让影像与影片的量产成本更低、扩散更快。产品层面则预期 2026 各大展会充斥机器人示范：以大型语言模型降低训练量、提升家务类任务的成功率，但短期多为 demo，真正商用仍需额外测试以降低损害风险。工作场域方面，可能出现以监控员工点击、卷动、打字来搜集资料、训练工作代理人的新一波「bossware」，并与「永远开启」的会议听写/摘要工具一起，把同意、隐私、诉讼与资料外泄风险推向 2026 年的焦点。
在自驾车趋势上，文章给出最清楚的量化路径：Waymo 预期在 2026 年底每周提供超过 100 万趟乘车服务，从目前的「数十万」级别上升，服务版图可能从 5 座城市扩到约 25 座，并传出为约 150 亿美元募资。虽然官方估计显示自驾相关事故「每月数十起」，但资料也指出 robotaxi 罕为肇因、致死事件占比很小；因此作者的反直觉预测是：2026 年扩张会很大，但更可能仍「无重大致命事故由电脑负责」。
The piece frames 2026 around six “scary” AI predictions, written on December 19, 2025 and anchored to a prior inflection point: after Google sounded alarms in late 2022, it carried out its first sweeping layoffs in January 2023. It asks whether OpenAI could mirror that pattern in early 2026, especially after rapid scaling. OpenAI is described as roughly a 10‑year‑old organization that has grown about fivefold in two years to ~4,500 employees, making a “reset” via restructuring or layoffs plausible and easily read as a bubble-deflation signal.
On infrastructure, it warns that anti–data-center activism could become a target for Chinese and Russian amplification, with generative AI lowering the cost and speed of producing persuasive images and video. On products, it expects 2026 conferences to overflow with robot demos that pair LLM-style models with physical machines to reduce training and improve task accuracy—impressive in staged settings, but still short of safely shipping into homes at scale. In workplaces, it anticipates more agent-training via surveillance software that records clicks, scrolls, and keystrokes, alongside always-on meeting tools—raising consent, privacy, breach, and lawsuit risk.
On autonomous mobility, it offers the clearest quantitative trajectory: Waymo expects to exceed 1 million rides per week by the end of 2026, up from the “hundreds of thousands,” while potentially expanding from 5 cities to about 25 and reportedly seeking ~$15 billion in funding. A common fear is a first at-fault fatal robotaxi crash, yet federal estimates cite dozens of self-driving-related crashes per month while suggesting robotaxis are rarely the cause and deaths are a small share. The article’s wager is rapid scale-up without a headline-making, computer-at-fault tragedy in 2026.