Gary Marcus breaks the OpenAI Navier-Stokes claim into three separate questions

GaryMarcus · x · 2026-09-22

Gary Marcus breaks the OpenAI Navier-Stokes controversy into three questions: (1) did OpenAI solve it per the Clay Institute's criteria — yes; (2) did training rely on unpublished work — unclear; (3) did they solve what physicists actually need — no. A thermofluids professor cited in the thread says the solution is not physically reproducible and teaches nothing about turbulence, a mathematical curiosity meeting the letter but not the spirit. Economist Noah Smith notes a human doing this would have won the $1M prize, while Marcus argues Sabine's accusations of disingenuousness unfairly conflate these points.

Related event: OpenAI's Navier-Stokes claim sparks debate over math validity and hype(9 posts)→

Original post →

More from Models

Models channel →