Behind the Curtain: AI godfathers converge on regulations

The chiefs of Google DeepMind, Anthropic, and OpenAI each published memos agreeing that frontier AI models need independent testing before release, a single US-led regulatory body, and restrictions on dangerous models. They differ on enforcement: Anthropic wants an FAA-style agency, DeepMind a FINRA-style body, and OpenAI an IAEA-style forum.