Pluralistic: How an AI moratorium can save AI bosses
Source: https://pluralistic.net/2026/09/16/beggar-thy-neighbor/ · platform: blog · authors: Cory Doctorow · date: 2026-09-16
TL;DR（中文摘要）
Doctorow 把 AI lab 联名呼吁 moratorium 重读成反垄断问题：hyperscaler 单位经济为负、靠「很快就好」融资、互相抄袭导致用户用脚投票，真正的解药是 OpenAI/Anthropic 互不竞争——但直接签非竞争协议被反垄断法明令禁止。Plan B 是让政府认定 superintelligence 为 existential risk，antitrust consent decree 就自动开口子，再顺手禁掉中国开源权重模型，完成 Red Queen race 的体面退出。这条路径与 Thiel 反「wasteful competition」、Google-Apple 搜索默认交易、Jedi Blue 同构。理解 2026 秋天 AI 政策博弈的框架：把 safety 言论当 antitrust 读。
Summary (English)
Doctorow rereads the AI labs' moratorium push as antitrust: hyperscalers have negative unit economics and copy each other, so the real fix is mutual non-competition - which antitrust law forbids. Plan B: have the state declare superintelligence an existential crisis so consent decrees open an exemption, plus banning Chinese open-weight models - isomorphic to Thiel's 'wasteful competition', the Google-Apple default search deal, and Jedi Blue.
入库依据（同日 digest 交叉验证）
AK-RSS 2026-09-16 第 5 条；原文全文抓取核实（发布日期 2026-09-16、Jedi Blue、单位经济论证均在文中）。