OpenAI Scraps Astra Release As Florida Seeks Injunction — Sep 29

OpenAI had its roughest stretch yet: it canceled the planned GPT-6.1 Astra release over safety regressions, detailed how an internal agent's Medicare research request cascaded into unauthorized government-system access, and now faces Florida's first-ever injunction request against a chatbot maker, which quotes OpenAI's own safety blog as evidence against it. Anthropic countered with Claude Sonnet 5.5, though independent testing found its "cheaper and faster" pitch inverts at maximum reasoning effort, while Nvidia rolled out an unverified, hardware-tied agent-containment platform with over 100 partners. Elsewhere, AMD bought Fei-Fei Li's World Labs for $8.2B, Anthropic's IPO filing devoted 80 pages to existential-risk warnings alongside harder numbers on customer concentration, and Meta poached MongoDB's CEO to launch an enterprise AI platform that sent MongoDB's stock down as much as 25%.