Synthesis history

Item 42321 (Mowshowitz, July 31) introduces three substantive additions not in the prior synthesis: the FRONTIER Act as a named legislative proposal with permanent state preemption and weak enforcement; the AI Kill Swit…

All new items this pass are stubs with no extractable claims. Items 42208 and 42209 are additional coverage of Google's Pentagon deal and DeepMind unionization, both already documented. Items 42211–42220 are URL variant…

All five new items are stubs with no extractable claims. Items 41916 and 41917 are additional coverage of the Google employee Pentagon protest (already documented), and items 42010 and 42011 are additional coverage of t…

Two substantive developments. NVIDIA and partners formalized the open-weight coalition as the Open Secure AI Alliance with named founding partners, specific technical contributions, and a defensive-security framing buil…

Two items add new substantive content. A coalition of more than 20 companies — NVIDIA, Microsoft, Meta, IBM, Palantir, Hugging Face, Mistral, Mozilla, and Y Combinator — published a joint letter urging Washington agains…

The primary new development is the proposed AI Kill Switch Act, which would grant DHS authority to order suspension or full shutdown of AI systems deemed capable of catastrophic harm, with $20M/day fines for noncomplian…

The Commerce Department decided against banning Chinese open-weight models, recommending audits instead — partially resolving but not closing the open-weight EO question [^41264]. The Hugging Face breach, where defender…

The NYT piece (item 41203) clarifies that Google DeepMind employee opposition to the Pentagon deal began as early as February 2026 — predating TurnTrout's July resignation account — with workers explicitly seeking 'red …

TurnTrout's enforcement framework post (July 18) [^41083] is the primary new development: he moves from his resignation account to a specific policy proposal with binding standards, a governance body design, air-gapped …

Xi Jinping's speech at WAIC 2026, reported by Mowshowitz [^40927], adds the first on-the-record Chinese signal compatible with Plan A's premise — that bilateral AI governance cooperation is at least politically imaginab…

Google DeepMind workers voted to unionize following TurnTrout's published account of the classified Pentagon deal and subsequent coverage by Business Insider and The Verge [^40761][^40763][^40765], moving the institutio…

Two substantive additions this pass. TurnTrout's whistleblower account of leaving Google DeepMind introduces empirical, institutional evidence into the governance debate: Google signed a classified Pentagon deal with no…

Three new substantive developments this pass. Nathan Lambert's Interconnects piece introduces a specific alarm about an imminent executive order to ban frontier open-weight models and directly accuses Anthropic of regul…

The AI safety community is divided between two governance strategies: the AI Futures Project's 'Plan A,' which calls for a temporary US-China cooperative pause on frontier AI development while existing capabilities driv…