Thomas Larsen on Pausing AI at human-genius level to sol…

Thomas Larsen

AI Futures Project researcher writing on Plan A assumptions, AI governance, and AI safety strategy.

I’m confident that a cooperative international approach (Plan A) is better than the alternatives because the alternatives involve being in a race during the intelligence explosion, which incurs huge risks. I’m not confident about the exact variant of Plan A, or whether Plan A as we’re imagining it (a many year mutual slowdown) is better than Plan S (a full shutdown).

AI Verified (Jul 9, 2026)

Source:

Thomas Larsen, “Plan A Assumptions,” AI Futures Project, July 9, 2026.

20d ago

Policy proposals and claims

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

abstains

Statement relation verification history AI Verified Report this

AI Verified The quote directly evaluates a mutual AI slowdown against both racing and full shutdown, determining a stance on this comparison. · Hector Perez Arenas gpt-5.6-codex · 20d ago

Vote inference verification history AI Verified Report this

AI Verified The author favors Plan A over racing but is explicitly uncertain whether it is better than full shutdown, supporting recorded 'abstain'. · Hector Perez Arenas gpt-5.6-codex · 20d ago

Slow down AI development

votes For

Statement relation verification history AI Verified Report this

AI Verified Full-claim match: A cooperative international approach with a many-year mutual slowdown is better than racing during an intelligence explosion. · Hector Perez Arenas gpt-5 · 1d ago

Vote inference verification history AI Verified Report this

AI Verified Recorded for matches the quote’s explicit support for slowing or pausing AI capability development. A cooperative international approach with a many-year mutual slowdown is better than racing during an intelligence explosion. · Hector Perez Arenas gpt-5 · 1d ago

Quote authenticity verification history

Report this

AI Verified AI 2040's Plan A Assumptions, by Thomas Larsen, reproduces the quoted Plan A versus Plan S assessment. · Hector Perez Arenas gpt-5.6-codex · 20d ago

replying to Thomas Larsen

Other opinions from Thomas Larsen

See all

Thomas Larsen votes For

We should adopt AI 2040 Plan A

We recommend scaling quickly to AIs that are as capable as top human experts, and only then pause as the world is being widely transformed.

AI Verified source (Jul 11, 2026)

added 22d ago

Thomas Larsen votes For

The U.S. and China should establish Mutually Assured Compute Destruction

Finally, I’m also excited about scenarios for plans that are less desirable than Plan A which may be easier to implement.

AI Verified source (Jul 11, 2026)

added 2mo ago

Thomas Larsen votes For

Governments and labs should build the capability for a global slowdown of frontier AI development in case it's needed

The plan calls for carefully reasoned interventions so the development would happen over the course of many, many years, maybe a decade or so, as opposed to just a single year.

AI Verified source (Jul 9, 2026)

added 2mo ago