The clash began with a bleak calculation by students entering a profession being rapidly reshaped by machines. Caiman Moreno-Earle, a 19-year-old Caltech math major, helped devise a multiday “mathathon” in which participants would deploy AI on hard theoretical problems. For young researchers, he argued, waiting for senior academics to settle the rules is not an option: “We don't have the luxury of waiting around to find out.”¹

The organizers secured an extraordinary $2 million in computing support — $1 million each from Anthropic and OpenAI — and sought advice from Caltech faculty and other mathematicians as they built a framework for the event. But when the flyer appeared on September 4, its question — what mathematicians should do when AI solves conjectures faster — triggered a revolt.

More than 700 academics and students signed an open letter calling for cancellation. Their concern was not simply that AI belongs outside mathematics, but that the event’s framing exaggerated what systems can reliably do, rewarded rushed results and offered AI companies a conspicuous marketing stage. The alarm drew high-profile backing: Yann LeCun amplified an appeal to read an open letter signed by 24 Fields Medalists, titled “A Severe Misalignment of AI in Mathematics.”Yann LeCun (ylecun)²

OpenAI then withdrew its $1 million in tokens, saying through research lead Dan Roberts that it recognized rapid progress in AI mathematics was “disruptive.” The pullout followed the backlash to the Caltech event.³

Yet the organizers declined to cancel. Alvan Arulandu said the dispute reaches beyond one competition: mathematics must preserve its human dimension while adapting to increasingly capable systems. To them, collaboration with AI companies is a chance to shape that transition rather than surrender it. The event remains scheduled to start October 30, with organizers seeking replacement computing support.