The two nuanced takes I’d add to the discussion: 1. The reason most people see math as different from curing cancer today, is the time lag between fundamental mathematical breakthrough and application is way too long, sometimes centuries. This will change in the next couple years, and then the intrinsic value of mathematics, in addition to the human value and community value, will be more visible. It is for the intrinsic value of mathematics that I consider the Caltech Mathathon a positive event. Discoveries should be encouraged even by participants who are less “mathematically mature” as long as we can Lean verify them. Ramanujan did not make his career by being mathematically mature but instead by his discoveries, using an incomprehensible workflow to get those magical formulas. 2. The problem is today human mathematicians are capable of producing paradigm-shifting ideas, and some view AI as there to finish the last mile. Examples of paradigm-shifting ideas include the Maynard Sieve Method. The win for humanity is not to shrink 70 million down to 600, but this method. This is a problem because if human mathematicians are overshadowed by AI advances, fundings get cut, no one stays in academia, young mathematicians use AI for exercises that are supposed to train them, as AI capabilities currently stand we will lose these revolutionary ideas (that are central to the headline breakthroughs). However, as AI capabilities improve, very soon we will see AI able to generate these ideas - I am hopeful. Optimism is the only rational take. Strategically, AI for math companies should invest in conjecturing, library learning, and other theory-building research directions. And we must involve human mathematicians to work on not only understanding and scientifically communicating the results - to “digest” them after they are proved, but also selecting them. *We need to select goalposts in a way that form theory.* We also need to advance technologies to map theory into applied science and engineering domains, with formalization of scientific computing one very obvious technical bet Axiom is taking right now. What, then, will the role of mathematicians be? One thing I am very certain of, is that in addition to being educators, they will play a meatier role than today in constructive economy, rather than taking a back seat. I think they will do what they love with day-to-day workflow more like developers. In fact, this is already happening at Axiom. I think this is a perspective many of us at Axiom share, which is somewhat between the current stances academia and big AI labs take. We are just really math-pilled.

@CarinaLHong believes math is training for any domain that involves structured reasoning, from working with constraints to coming up with abstractions. Watch this conversation with @mattmcilwain about how a deep love of math becomes the strategic instinct behind Axiom’s wins 💫

Was a pretty fun podcast where I got to nerd out with @mattmcilwain about math a bit! We also covered the history of ATP, ITP, and how they converge with modern AI. Thanks to friendos at Madrona for supporting Axiom since day 1.

I'll be speaking at #STOC2026# on AI for theory, alongside industry pioneers @WittedNote (Chief Technologist, Google), David Woodruff (CMU), @SebastienBubeck (OAI), and my friend @MarkSellke (OAI) at the very kind invitation of the organizers. See you in Salt Lake City on 6/27!

This morning at SAIR Science x AI Summit, @CarinaLHong spoke alongside Nobel Laureate Barry Barish, Fields Medalists @wtgowers, Terrence Tao, and @leanprover Founder @leodemoura on Frontiers of AI for Math. Tune in at 2h34m30s for AxiomProver's results in its first 100 days! 💯

"Ono is an American Mathematical Society (AMS) Fellow, a previous AMS vice president, and founding mathematician at Axiom Math. "Kudos to the AMS and AIM for stepping up and performing this vital service. Your tireless efforts to bridge the gap between the mathematics community and our nation's leaders are deeply appreciated by us all.""

@KenOno691 on behalf of Axiom just briefed Congress on the progress of AI for mathematics. A rising tide lifts all boats. In the AI era, we share our understanding of the frontier here at Axiom with Washington to advocate for investment in math and fundamental science research.

The Ultimate List of Artificial Intelligence "Neolabs": May 2026. A Neolab is a pre-revenue scale startup working on long-term AI breakthroughs, usually with a $1B+ valuation. There are now 63 of them!

We raised a Series B led by @sequoia at a $2 billion valuation and I’m excited to welcome @andrew__reed to our board. All our existing investors doubled down: @kleinerperkins, @IndexVentures, @khoslaventures, @firstround, @sparkcapital, @AbstractVC, and @terraincap.

Continuing to model the world through long-form podcasts! @CarinaLHong This episode is full of the "beauty of mathematics". She is also our youngest guest to date — Hong Letong, born in 2001, known online as the "Math Girl". This is her first-ever interview in Chinese. We talked about mathematics as invented versus discovered, proofs from the Book of Proofs, the entrepreneurial journey of the most unlikely founder, and of course, AI for Math. ✍🏻✍🏻 A 4-hour Interview with Carina Hong: AI for Math, Lean, Proofs from The ... 来自 @YouTube