WorkAugust 20, 2026via The Verge AI
Welcome to the AI crisis in math
Why it matters
AI has abruptly crossed a capability threshold in abstract mathematics, triggering an existential reckoning in an entire academic discipline about its future role, workforce structure, and the nature of mathematical discovery itself — a crisis unfolding in real time with no consensus on whether it's existential or transformative.
Key signals
- OpenAI's Astra model published solutions to 10 longstanding math problems (sphere packing in higher dimensions, quantum game theory, others); mathematicians say solving all 10 would be career-defining for a human researcher
- Models remain poor at elementary arithmetic and counting but excel at abstract theoretical proofs requiring reasoning and pattern connection
- Fields medalist James Maynard and other senior mathematicians report 'soul-searching' over AI's impact on PhD student career viability and field sustainability
- Key concern: AI may 'mow down' problems without opening new research directions, leaving the field sterile and training grounds for junior researchers depleted
- Transparency gap: OpenAI has not disclosed how many attempts were required to solve these 10 problems, or how much human mathematician guidance was involved
- Cost barrier emerging: Astra results claimed to cost ~$2,000, but high-end compute access may lock out mathematicians and institutions without research grants (common in pure math)
- Leiden Declaration: mathematicians have signed an open letter pledging to resist hype around AI mathematics; efficacy unclear
- Attribution sloppiness: OpenAI initially claimed 'no progress in 10 years' on one problem, then quietly corrected to acknowledge prior researcher contributions
- Democratization tension: while AI access broadens globally, flood of AI-assisted papers on preprint servers and peer review systems straining legitimate research vetting
- Graduate student anxiety: fear that PhD topics will be solved by AI mid-thesis, forcing students to chase AI's moving frontier rather than pursue their own research
The hook
Fields medalists are soul-searching. OpenAI's Astra solved 10 unsolved math problems in one drop. Mathematicians are asking: if AI solves all the outstanding questions, what's left for us?
Today on Decoder, I’m talking with Robert Hart, The Verge’s London-based AI reporter, about what AI is doing to the field of mathematics and the existential crisis many lead mathematicians are having about it.
OpenAI just published a set of solutions to longstanding problems in math that went off li…