The Upstarts Podcast Axiom’s Carina Hong: Solving Math’s Hardest Problems With AI, And AI's Problems With Math
Apr 2, 2026
Carina Hong, founder and CEO of Axiom Math and former Morgan Prize–winning mathematician, leads work on AI that performs and verifies advanced mathematics. She discusses building AI that solves and formalizes hard proofs, Axiom’s real-time Putnam success, and using formal verification to make AI-generated code and math trustworthy.
AI Snips
Chapters
Transcript
Episode notes
Verified Math AI That Produces Trustworthy Proofs
- Axiom built an AI that does math at a superintelligent level and produces verified outputs you can trust.
- Carina Hong frames this as a “verified AI” or perfect prover able to crack Olympiads and research problems while ensuring correctness.
Verification Multiplies AI's Scientific Power
- Verification scales AI usefulness beyond catching hallucinations to compounding superintelligence.
- Hong compares verification to mentoring Ramanujan into rigorous proof writing, turning intuition into reliably usable results.
Axiom Aced The Putnam Exam In Real Time
- Axiom achieved a perfect score on the Putnam exam in real time, a feat no prior AI matched.
- Hong notes other labs tried and failed, while Axiom got 120/120 evaluated by Math Arena versus competitors' lower scores.