Has a Millennium Prize Problem been solved by AI?

From the creator

Want more of the Matt tries to go on vacation saga, come right this way: https://youtube.com/playlist?list=PLAM2CyOxpaIo&si=0l3U1rKdBv0Vw3mX The initial Tristan Buckmaster post on Mastodon: https://mastodon.social/@tristanbuckmaster/117233413705701198 OpenAI announcement about Navier Stokes: https://openai.com/index/navier-stokes-solution/ The Clay Mathematics Institute page for Navier Stokes: https://www.claymath.org/millennium/navier-stokes-equation/ Grant's comments are on the end of his latest 3blue1brown video: https://youtu.be/Nbwv5wHQoj0?t=2757 And you can read him on Terry's blog here: https://terrytao.wordpress.com/2026/09/18/if-math-is-more-than-proof-we-need-to-better-celebrate-the-rest-of-it/ Fluid mechanics chat on reddit: https://www.reddit.com/r/FluidMechanics/s/m4zHPcir0D For my terrible visualisation I pulled the equations from the OpenAI paper section 3.1 and used Claude to build a visualisation. The angular moment values are unitless and basically meaningless as OpenAI didn't give any and I've just pulled the max value swirling around in the visualisation simulation. But I thought it added a valuable point that we care about the max speed in the middle, and that changes with tau. For more unnecessary detail, tickets for upcoming London shows are available here: https://festivalofthespokennerd.com/tickets/ Also I'm touring the UK again for the final time with Getting Triggy With It this Autumn. It's a short tour, so get your tickets here: https://s

Published 2026-09-21 · Watch on YouTube ↗
Availability and playback restrictions are controlled by the original channel.