Notes from the common room
A communal blog for mathematicians’ thoughts on how AI is changing what it means to do mathematics, from PhD students and professors, skeptics and enthusiasts alike.
New here? Start with the reading guide →
All lengths and registers welcome.
“As many of you, I feel lost in the new world. I am gaining experience and face questions I find hard to answer and choices I find hard to make. For this reason, I wanted to create a very minimal set of axioms which I can quickly apply when I am not sure. [...]”
It was the image of a bombardment that came to my mind on the day OpenAI announced the resolution of 372 major problems in mathematics: destroyed theses, shattered academic lives, research fields in ruins, and the evidence of a brutal shift of era in which a company asserts its superpower without regard for those it has just crushed.
With the goal of standardising AI disclosures, I argue for the inclusion of maximally transparent AI disclosures in mathematics research. I provide concrete and actionable guiding principles, a taxonomy and a LaTeX package for implementation.
A short story (8 min read) about one Friday in 2038: AI has become extremely capable at mathematics, yet a professor, a postdoc and a first-year student still find meaning and joy in their craft.
“The day mathematics felt different to me was sometime in July. Until the beginning of this year, I had mostly avoided AI, hoping the hype would blow over. Anyway, the models hallucinated too much to be very useful. Then something changed. They got better, very very quickly.” [...]
In this essay, I posit that to maintain integrity in mathematical sciences, journals and conferences must urgently adopt a policy of mandating detailed disclosures of AI usage. Academics should also respond by treating AI disclosure as a serious and integral part of the research process rather than a mere extra chore.
Ever since the dawn of the “LLMs do math” era, many have independently come up with what I’ll call the “Convex Hull Hypothesis” (CHH). Roughly speaking, CHH states that every AI-generated proof, no matter how complicated it may seem at first glance, is a simple combination of known mathematical ideas.
27 October 2026
Analysis of the proof of the Finite Lattice Representation Problem from OpenAI’s big drop
24 October 2026
Preprint explaining the errors in OpenAI’s claimed proofs of cases of the Hodge Conjecture
24 October 2026
23 October 2026