Researcher Accuses Prominent AI Figure of Rebranding Already-Solved Math Problems as AI Breakthroughs

BlancheMinerva · x · 2026-09-09

AI safety researcher BlancheMinerva alleges that a prominent AI researcher twice announced AI solutions to math problems that had actually been solved before — in one case the AI proof was literally the human's proof. The defense offered was that "found solutions" doesn't necessarily imply novelty; the poster calls that technically true but not something an honest person would use as a shield, adding that the figure has a reputation for dishonesty.

Related event: Ex-DeepMind researcher casts doubt on disputed AI math claims(2 posts)→

Original post →

More from Fun

Fun channel →