Mathematics is becoming AI's marketing surface

Mathematics has become a proving ground for AI labs, and the credit disputes now surfacing are how that pressure registers inside a working discipline. The pattern is not that machines suddenly match human proof-craft. It is that headline-grade problems, long-standing yardsticks for what counts as hard thinking, now function as trillion-dollar advertisements for whoever plants the flag first.

Steven Strogatz, the Cornell mathematician and co-author of the forthcoming book "Big Math," put the mechanism plainly in his WIRED interview: OpenAI's recent Navier-Stokes-adjacent result is a "marketing device" worth perhaps a trillion dollars to the lab that can claim superiority over Anthropic. Tristan Buckmaster at NYU posted solutions to three closely related problems days before OpenAI's announcement, then alleged the lab "tried to influence who got credit for the work." Anthropic, for its part, said Claude formalized Fermat's Last Theorem and proved 29,500 supporting theorems last week, a project other mathematicians had pursued for years.

The repeatable mechanism: a prestigious open problem is declared solved, a small set of human contributors is credited or erased, and the announcement lands on financial pages before it lands in journals. Working mathematicians absorb the cost. Corporate labs harvest the narrative. Strogatz calls 2026 a candidate annus mirabilis or annus horribilis, but the diagnosis and the disease are now the same thing: mathematics is being absorbed into corporate competition, and the credit system is the part that breaks first.