AI & Maths watch GitHub

Retrospective research digest

Updates on artificial intelligence for mathematical research, proof discovery and formal verification, with links to papers, code and tools.

At a glance1 entry · 1 linked source Partial coverage

One dated research-culture essay qualifies for September 10, 2026. No additional mathematical result or tool release was established for this day; archival and social-media coverage remains incomplete.

Research updatesSelect an entry to read more

Matt Baker examines how AI changes mathematical research, exposition and graduate training Matt Baker Published: 1 source

Baker’s dated essay argues that mathematical research includes choosing problems, creating definitions, developing theories and communicating insights, beyond producing proofs. He anticipates greater roles for researchers as facilitators and curators of AI-assisted work, while raising concerns about opaque discovery processes, unequal access to compute and the training of graduate students. His practical emphasis is on preserving understanding, exposition and mentoring as AI capabilities advance.

Research relevance:

Relevant to research planning and graduate supervision: the essay asks how departments should value mathematical understanding and sustain opportunities for students to develop research judgment.

Claim status:
A directly read, dated opinion essay. It announces no theorem, proof, formalization or independently evaluated AI capability. Its descriptions of earlier fluid-equation announcements are the author’s framing, not certification by this digest.
Limitations:
Predictions about the profession are speculative. Baker explicitly declines to adjudicate the provenance dispute surrounding OpenAI’s earlier announcement. The essay provides neither a reproducible workflow nor a new technical artifact.

Sources & further reading

Link to this entry
Coverage and limitations
  • Public web searches reconstructed September 10, 2026, within the supplied Paris civil-day window. Original page dates were used; indexing and crawling dates were not treated as publication dates.
  • Read Matt Baker’s dated essay and Anima Anandkumar’s September 10 guest post on Terence Tao’s blog. The Euler guest post was excluded because its exact URL and underlying project already appear in the supplied previous items, and no distinct new result or correction was established.
  • Consulted Gowers’s September blog archive, Scott Armstrong’s official research website, François Charton’s official biography, Kevin Buzzard’s Imperial webpage, and official websites for Epoch AI, Harmonic and Axiom. Project Numina’s research listing was available through indexed search text; its September 2026 Kakeya entry lacked an established day and was excluded.
  • Official webpages supplied social-account links for Armstrong and Charton and identified Buzzard’s Xena project. Identity checks were incomplete for the other supplied handles; no item relies on their assumed identities.
  • Grouped date-specific searches covered the supplied X handles, but did not provide comprehensive dated timelines. X profiles and third-party mirrors appeared in indexed results; these snippets were not treated as fully read posts or evidence of target-day publication.
  • The Xena September archive, Lean homepage and Buzzard’s Bluesky profile returned retrieval errors. Tao’s Mathstodon profile exposed no usable post history. These gaps prevent an exhaustive historical edition.
  • Trail of Bits’ Lean vulnerability disclosure was found through indexed search text but dated September 9, so it was excluded. September 10 secondary recaps of earlier fluid-equation announcements were also excluded. No proof was verified and no Lean code was executed.

Requested coverage window: 2026-09-10T00:00:00+02:00 — 2026-09-11T00:00:00+02:00.

Public web research; coverage is not exhaustive. A source link is not a certification of a claim.

Archives 32
Older editions (2)