Physical Superintelligence claims an Anthropic model proved a math conjecture
Physical Superintelligence says a non-public Anthropic model proved the Aaronson-Ambainis conjecture, with the proof checked in Lean.
Physical Superintelligence claims that a non-public Anthropic model proved the Aaronson-Ambainis conjecture, and that the proof was verified in Lean. It doesn't say which Anthropic model was involved, how it was given the problem, or what the proof establishes.
Because the model is not public, outside researchers cannot run it to test the result themselves. The post doesn't say whether the Lean code has been released for others to inspect. It cites no statement from Anthropic confirming or disputing the work.
What changed
Physical Superintelligence claims a non-public Anthropic model proved the Aaronson-Ambainis conjecture, with the proof verified in Lean; the claim is unverified.
Type: Research · Applies to: non-public Anthropic model
Why it matters
If verified, it would show a frontier model producing formal research results, but for now it is an unverified claim about an unreleased model.
What you can do
Wait for the proof, its Lean file and an Anthropic statement before drawing conclusions.
How the story developed
The startup's own blog gives the Lean verification details and says the system improved the result twice before the final version was accepted.
Original source
Where we saw it
Import AI (Jack Clark), Latent Space
AI wrote this summary from the sources listed, and nobody reviewed it individually before publication. Check the original source before you rely on it. How it works