OpenAI Withdraws from Caltech Mathathon After Mathematicians' Open Letter
An open letter from current and former Caltech mathematicians accusing AI labs of treating research math as an 'advertising opportunity' prompted OpenAI to pull its sponsorship of the October 30 hackathon.
What matters
- OpenAI withdrew sponsorship of the Caltech Mathathon after an open letter from current and former Caltech mathematicians warned of 'destructive impacts' on the discipline.
- The hackathon, scheduled for October 30, was backed by roughly $2 million in AI credits jointly from OpenAI and Anthropic.
- The letter accused AI companies of treating research math as an 'advertising opportunity' and warned labs would 'take credit for the effort of talented undergrads.'
- OpenAI research lead Dan Roberts said the lab wants to 'engage with the math community more'; Anthropic has not responded.
- Anthropic now faces pressure to decide whether to remain as a sponsor.
What happened
OpenAI has pulled its sponsorship of the Caltech Mathathon, a hackathon scheduled for October 30 in which participants would attack open research problems by prompting frontier large language models. The withdrawal came after current and former Caltech mathematicians published an open letter arguing the event "is likely to have destructive impacts for the mathematical community."
The letter accused AI companies of treating research mathematics as an "advertising opportunity" and warned that sponsoring labs "will take credit for the effort of talented undergrads." The event was backed by roughly $2 million in AI credits jointly supplied by OpenAI and Anthropic, according to Business Insider.
OpenAI research lead Dan Roberts said the lab wants to "engage with the math community more," signaling an intent to repair relations rather than simply walk away. Anthropic, which co-sponsored the event, did not respond to requests for comment.
Why it matters
This is not a typical sponsorship dispute. The mathematicians who signed the open letter are part of the academic community whose credibility effectively validates AI labs' claims about mathematical reasoning capabilities. When that community publicly rejects a lab's involvement, it undercuts the narrative that frontier models are making meaningful progress on hard mathematical problems — a narrative labs rely on for both prestige and investor confidence.
The episode also puts Anthropic in an awkward position. With OpenAI now out, Anthropic remains the event's primary AI-credit backer and faces pressure to make a similar call. Staying in risks aligning the company with an event its own academic peers have condemned; withdrawing would effectively end the hackathon's AI-powered premise.
More broadly, the letter reflects a growing tension between AI labs eager to showcase their models on hard research problems and domain experts who see these efforts as extractive — leveraging the labor of students and researchers while labs claim the results as evidence of model capability.
What to watch
- Whether Anthropic responds to the open letter and decides to maintain or withdraw its sponsorship.
- Whether the Caltech Mathathon proceeds on October 30 with revised sponsorship or a different format.
- Whether other academic departments or institutions issue similar letters opposing AI-sponsored research events.
- How OpenAI follows through on Dan Roberts' stated desire to "engage with the math community more" — whether that means new partnerships, dialogue with critics, or simply a quieter approach.
What to do next
Developers
If building tools that interface with academic research communities, proactively engage domain experts early and incorporate their concerns before launching public events or benchmarks.
The Caltech letter shows that academic communities can mobilize quickly against initiatives they perceive as extractive, regardless of sponsor intent.
Founders
Audit any academic partnerships or sponsorships for potential perception issues around credit attribution and labor exploitation before committing resources.
Reputational damage from academic pushback can outweigh the marketing value of a sponsored event, especially when the community in question validates your core claims.
PMs
When planning community-facing events involving frontier models, build in a stakeholder consultation phase with the relevant academic community before announcing sponsorships.
The Mathathon backlash demonstrates that top-down sponsorship without community buy-in can force last-minute withdrawals and negative coverage.
Investors
Monitor how AI labs manage relationships with academic communities whose endorsement underpins model capability claims; friction here can signal broader adoption risk.
Academic credibility is a key asset for labs marketing frontier reasoning capabilities. Public disputes with domain experts can erode that asset.
Operators
If your organization sponsors academic events, ensure sponsorship terms include clear credit-attribution frameworks and community consultation requirements.
The open letter specifically warned that labs would 'take credit for the effort of talented undergrads,' highlighting attribution as a flashpoint in academic-industry collaborations.
Testing notes
Caveats
- This is an editorial news story about a sponsorship withdrawal and academic dispute, not a testable product or model release.