OpenAI says ~10k concurrent agents solved the Navier-Stokes Millennium Prize Problem
DKokotajlo · x · 2026-09-15
OpenAI announced a solution to the Navier-Stokes Millennium Prize Problem — unresolved for 90 years — produced by a group of agents running on a next-generation model claimed to be more capable than GPT-6 Astra, using 10k concurrent agents. Former OpenAI employee Oleg Murk (5 years there) shared the news while publishing an "AI Pragmatist Manifesto": AI could compress the first half of the 20th century into the next 5 years, and he argues for sane regulation without doom or utopia, noting such models may soon run locally on high-end consumer hardware.
More from AGI Musings
- Economist: Superhuman AI communism could deliver prosperity, but jail for wrongthink remains bad — paulnovosad · 2026-09-15
- AI safety insider: labs' accelerate-to-control camp is the weirder, often anti-human one — erikphoel · 2026-09-15
- AI safety expert: loss of control was never researchers' top concern, but that may be shifting — DavidSKrueger · 2026-09-15
- 20-year AI veteran Margaret Mitchell: don't slow the pace, change the game — mmitchell_ai · 2026-09-15
- When AI makes one step 10x faster, the queue behind it becomes the bottleneck — AnneliesGamble · 2026-09-15
- Eric Topol spotlights the case for a science of the AI mind — EricTopol · 2026-09-15