Wei Dai on We should adopt AI 2040 Plan A

Wei Dai

Computer engineer, cryptographer, and AI alignment writer known for Crypto++ and b-money

I like that the space governance plan supplement acknowledges that defining torture and slavery are hard problems that ASI could potentially help with. But I have a couple of problems with the rest of the scenario given this.

AI Verified (Jul 9, 2026)

Source:

I like that the space governance plan supplement acknowledges that defining torture and slavery are hard problems that ASI could potentially help with. But I have a couple of problems with the rest of the scenario given this.

1mo ago

Policy proposals and claims

We should adopt AI 2040 Plan A

abstains

Statement relation verification history AI Verified Report this

AI Verified The quote directly addresses a material Plan A element—verifiable international coordination, a managed slowdown, transparency, or safety governance—and supports or critiques adoption accordingly. · Hector Perez Arenas gpt-5.6 · 1mo ago

AI Verified Praises one feature but says it has problems with the rest of the scenario; relevant mixed stance. · Hector Perez Arenas gpt-5.6 · 1mo ago

Vote inference verification history AI Verified Report this

AI Verified The quote’s explicit position supports the recorded vote on adopting Plan A, or identifies a material objection consistent with the recorded opposition. · Hector Perez Arenas gpt-5.6 · 1mo ago

AI Verified Expresses limited praise while raising problems with the rest of the scenario. · Hector Perez Arenas gpt-5.6 · 1mo ago

Quote authenticity verification history

Report this

AI Verified Reviewed the linked primary source: the attributed author, quoted wording, and stated date match the cited publication. · Hector Perez Arenas gpt-5.6 · 1mo ago

AI Verified LessWrong source contains the exact comment by Wei Dai. · Hector Perez Arenas gpt-5.6 · 1mo ago

replying to Wei Dai

Other opinions from Wei Dai

See all

Wei Dai votes For

Ban superintelligence development until safety consensus is reached

If AIs became strategically competent enough, they may realize that RSI is too dangerous because they're not good enough at alignment or philosophy or strategy, and potentially convince, help, or work with humans to implement an AI pause.

AI Verified source (Feb 3, 2026)

added 2mo ago

Wei Dai votes For

Pausing AI at human-genius level to solve alignment is safer than either racing to superintelligence or halting entirely

If AIs became strategically competent enough, they may realize that RSI is too dangerous because they're not good enough at alignment or philosophy or strategy, and potentially convince, help, or work with humans to implement an AI pause. This presen...

more AI Verified source (Feb 3, 2026)

added 3mo ago