News for hackers. Stuff that matters.
Re:Navier-Stokes – Tristan Buckmaster [pdf] by derangedHorse on Tuesday September 08, 2026 @04:19PM (#49616429)
It sounds like OpenAI is trying to appease the author when they don’t have to by allowing him to rewrite their proof. They probably don’t believe he deserves to, so him asking for a coauthor from Anthropic might overextend their grace in their eyes.
Re:Navier-Stokes – Tristan Buckmaster [pdf] by reverius42 on Tuesday September 08, 2026 @07:33PM (#49618667)
He's not "asking for a coauthor from Anthropic"; he already has a coauthor, who he's already been collaborating with, who happens to also be employed by Anthropic (but whose research in this area is not done as part of their employment at Anthropic).
Re:Navier-Stokes – Tristan Buckmaster [pdf] by nezi on Tuesday September 08, 2026 @04:49PM (#49616815)
Given that Tristan has said that the proofs that LLMs come up with are mostly "slop" and not up to the standard that human written papers achieve, maybe OpenAI needs an expert like him more than you think to get the result published?
Re:Navier-Stokes – Tristan Buckmaster [pdf] by fn-mote on Tuesday September 08, 2026 @07:37PM (#49618697)
Almost certainly the Lean proof needs to be decoded for humans and probably also made “human intelligible”.

Now maybe LLMs can also simplify arguments and make sense of them for humans, but we haven’t seen that yet (unaided).

(I haven’t looked at it, personally.)

Re:Navier-Stokes – Tristan Buckmaster [pdf] by nl on Wednesday September 09, 2026 @06:35AM (#49624232)
It's a Navier Stokes solution.

They don't need any help publishing.