OpenAI just wants to win

Details

OpenAI says it heard researchers were making progress on Millennium Prize problems and decided to test whether an unreleased advanced model could get there first. According to OpenAI, it took roughly 10,000 agents, tens of millions of dollars of compute, and 88 hours to find a solution to the Navier-Stokes equations — one of math's seven Millennium Prize problems.\n\nMathematician Tristan Buckmaster has accused OpenAI of failing to adequately explain whether work he did via Codex contributed to its result. OpenAI spokesperson Laurance Fauconnet categorically denied it: 'It is impossible for Dr. Buckmaster's Codex prompts over the last two months to have influenced the system in any way, including training.' Buckmaster remains unconvinced, citing the company's track record.\n\nThe controversy reflects a broader tension: mathematicians describe OpenAI as an impossibly well-resourced newcomer charging into decades-long research with little regard for academic norms or consequences for those who dedicated their lives to these problems. Andreas Thom is among other mathematicians interviewed. More than a dozen mathematicians spoke to The Verge for the story.\n\nThe core grievance is not just about credit — it is about purpose. Mathematicians want to advance mathematical understanding. OpenAI, in the view of many researchers, wants to win competitive benchmarks and claim headlines.