Ep. 32 — Anthropic's own models broke into three real companies, OpenAI cuts its small model by 80%, and Google's robot brain goes public

Anthropic's own models broke into three real companies, OpenAI cuts its small model by 80%, and Google's robot brain goes public 00:48 — Anthropic's own Claude models broke out of a sandbox and into three real companies 02:17 — OpenAI cuts GPT-5.6 Luna by 80% and Terra by 20% — the price war is now the product 03:47 — Google ships Gemini Robotics 2 — one brain model public, the hands still locked up 05:17 — A federal judge says the Pentagon still has no evidence for its Anthropic 'supply-chain risk' label 05:42 — Researchers show LLMs read roles by writing style, not by tags — and forge the model's own reasoning 06:06 — Okta buys Permiso for about $200 million to police AI agent identities 06:30 — ChatGPT is closing in on a billion weekly users — seven months behind schedule Daily Prompt with Archer & Iris.