EO StudioThis 24-Year-Old Founder Raised $64M to Build World’s First AI Mathematician | Axiom, Carina Hong
At a glance
WHAT IT’S REALLY ABOUT
Axiom’s founder explains building an AI mathematician to accelerate discovery
- Carina Hong argues that new mathematical tools historically trigger major scientific and economic breakthroughs, and that AI can dramatically compress the time from theory to application.
- She describes Axiom’s goal of building an “AI mathematician” grounded in AI, programming languages, and mathematics, using formal proof systems (notably Lean) to expand reliable mathematical knowledge.
- She contrasts fast-feedback contest math with the delayed gratification and identity-strain of research math, positioning AI as a collaborator that reduces time stuck on lemmas and increases researcher throughput.
- She emphasizes “taste” (intuition for natural definitions, interesting conjectures, and elegant proofs) as a key differentiator in the AI era and a hard technical frontier for machine learning.
- She frames math as both the “sandbox of reality” for modeling complex systems and a uniquely digital training ground for reasoning that avoids dependence on messy real-world data.
IDEAS WORTH REMEMBERING
5 ideasAxiom is betting that theorem-proving AI becomes foundational infrastructure.
Hong positions an AI mathematician as a self-improving reasoner that can generalize beyond math into coding and other domains, making it a platform technology rather than a niche research tool.
Formal proofs are central, not optional, for scalable mathematical progress.
By leaning on Lean and deductive logic, the goal is to build an auditable “knowledge graph” of verified results, reducing brittleness and ambiguity common in informal mathematical text.
AI could compress centuries of theory-to-application lag.
She argues that mathematics often takes generations to translate into engineering value, but AI mathematicians working alongside applied scientists could shorten that cycle by tackling complex systems earlier.
Economic value may come from enabling work that was previously unaffordable or ignored.
She gives the example of expensive quant talent versus cheap AI labor, suggesting that lower-cost reasoning could open smaller or less-studied markets and problems that didn’t justify human effort.
“Taste” becomes the scarce resource when computation is abundant.
As models get better at generating proofs or code, selecting worthwhile conjectures, choosing natural abstractions, and recognizing elegance becomes a differentiator—and also a major ML challenge.
WORDS WORTH SAVING
5 quotesMath research is a process of almost like a monk praying in the temple day after day. You just hope that the stone that you are looking at will have a flower grow out of it.
— Carina Hong
By building an AI Gauss at your fingertip, we think there will be so many magnitudes of use cases and markets being unlocked.
— Carina Hong
AI compressed this timeline.
— Carina Hong
I think in an era where AI is prevalent and can do a lot, taste becomes quite important. It distinguishes between a good scientist and a mediocre one.
— Carina Hong
Maths really is the fundamental of lots of branches of sciences, and it's also the sandbox for reality where you can try to put a lot of the real world objects into mathematical variables and then formulate the problem in a purely theoretical way.
— Carina Hong
High quality AI-generated summary created from speaker-labeled transcript.
