This company seems to be trying to automate the "hard problem solving" aspect of math. I've never been interested in that. To me, the interesting part of math is creating beautiful structures that help us understand reality better - and even more importantly, *explaining* these structures well, so more people can use them. If AI manages to automate *this*, I'll enjoy talking to it.
This sentence in the Business Insider article needs some context:
"Axiom, which recently said it solved two Erdős math problems that eluded mathematicians for decades, announced a $64 million seed round in September."
They're talking about Erdős problems #124 and #481.
Axiom's program solved a wrongly stated version of problem #124 that's much easier than the important version. Terence Tao (npub1gve…sduy) wrote:
"Recently, yet another category of low-hanging fruit has been identified as within reach of automated tools: problems which, due to a technical flaw in their description, are unexpectedly easy to resolve. Specifically, problem #124 https://www.erdosproblems.com/124 was a problem that was stated in three separate papers of Erdos, but in two of them he omitted a key hypothesis which made the problem a consequence of a known result (Brown's criterion)."
The harder version is still open: https://www.erdosproblems.com/124
The other problem, Erdős Problem #481, had been solved by David Klarner in 1982:
https://www.erdosproblems.com/forum/thread/481#post-1934
However, the Axiom team didn't notice this.
A 24-year-old Stanford dropout has wooed top Meta AI researchers to her nascent startup, which is building an AI mathematician. https://www.businessinsider.com/axiom-math-stanford-dropout-meta-ai-researchers-startup-2025-12 #artificialintelligence #math
