A Caltech undergraduate-led AI math event has become a national flash point after more than 700 academics and students signed an open letter urging it to be canceled. OpenAI, one of the event’s sponsors, then backed out and took back its $1 million in tokens.
The event, called a 'mathathon,' was organized by Caiman Moreno-Earle, a 19-year-old sophomore math major at Caltech, and another undergraduate organizer. They first approached AI companies because the event needed massive computing power and they did not have much capital themselves.
Moreno-Earle said he asked Anthropic for $1 million in a Zoom meeting, expecting far less. Anthropic said yes, and later OpenAI agreed to provide another $1 million. That put $2 million in computing power behind the event and, in Moreno-Earle’s words, turned it from a 'fun little experiment' into something with a responsibility to the math community.
The organizers say they consulted professors at Caltech and other figures in the field to build a responsible framework for using AI. Even so, once the flyer went live on September 4, critics argued the event would promote AI 'slop,' encourage rushed work, and serve as a marketing opportunity for AI companies.
The open letter said the event, as sponsored by Anthropic and OpenAI, 'is likely to have destructive impacts for the mathematical community.' It also challenged the organizers’ framing, especially the line asking, 'What is the role of a mathematician when AI can solve conjectures faster?'
OpenAI’s research lead, Dan Roberts, said on X: 'We recognize that the rapid progress of AI in mathematics is disruptive.' The event is still set to continue, with some slight alterations, after organizers decided not to cancel.
What’s really on the line here is not just one campus event. It is whether the next generation of mathematicians learns to work with AI or gets forced to watch the field change around them. Moreno-Earle said young mathematicians do not have the same luxuries as older ones and do not know whether graduate school openings will still be there when they reach Ph.D. age.
That is the hard truth for founders, operators, and students watching AI move faster than institutions can absorb it: delay is a strategy, but rarely a winning one. If the tools are already changing the work, the real advantage goes to the people who learn the new playbook first.
