Introduction for and Reactions to Plan A
Zvi's AI Roundups · Zvi Mowshowitz · 2026-07-11
Zvi Mowshowitz introduces and responds to community reactions to 'AI 2040: Plan A,' a proposal from the AI 2027 team calling for a US-China deal to slow superintelligence development through international compute monitoring and research transparency.
Appears in
Extraction
Topics: ai-governancesuperintelligenceinternational-ai-coordinationai-safety-policyplan-a
Claims
- Plan A's core proposal is a US-China deal to jointly control chip supply, audit data centers, and share research information to slow development of superintelligent AI.
- The central crux dividing Plan A supporters and detractors is whether superintelligence will arrive soon enough to justify Plan A's costs — those who believe it won't should oppose Plan A, and those who do should engage seriously with it.
- Vitalik Buterin argues that Plan A detractors implicitly assume AI is 'normal technology' while failing to apply consistent concern about power concentration to the possibility of superintelligence itself.
- Every viable scenario for handling superintelligence involves some unlikely events, and the alternatives to human-steered outcomes are either AI-steered outcomes or the equilibrium where AIs end up in de facto control.
- The authors of Plan A and Zvi both agree there are no 'non-naive' options and that criticizing Plan A as naive is not a meaningful objection unless the critic proposes a better plan.
Key quotes
Daniel Kokotajlo: 'We think it's still good to recommend what would actually be good, even if you think that your audience is probably not going to listen.'
vitalik.eth: 'Detractors say things like AI 2040 is naive about human coordination ability and a threat to freedom, but don't seem to see any naivety in assuming that the ASI transition will just go well by default, don't seem to see ASI itself as a massive power concentrator risk, and don't seem to feel fear of humanity's hard power dropping to zero if ASIs can do literally every task better than we can.'
Ryan Greenblatt: 'Plan A seems like a good plan for handling powerful AI, or at least the best plan anyone's written up. Many choices initially seem crazy, but are actually pretty carefully considered. Plan A isn't likely to happen, but pushing for something like this seems worthwhile.'