OpenAI Drops 722 AI-Generated Math Papers in One Day, Sparking Backlash

In early October, OpenAI released 722 mathematical manuscripts generated by an undisclosed internal model in a single drop, hosted in the openai/math repository on GitHub (Apache-2.0 license, over 13,000 stars), covering number theory, algebraic geometry, analysis, combinatorics, and more. The math community reacted strongly, with debate centering on the field's capacity to digest the output and on career prospects for early-career researchers.

Confirmed

• The 722 manuscripts came from testing the internal model on roughly 4,000 research problems and are grouped into 372 families of related results; OpenAI says each result took on average about 3 hours of ChatGPT Pro thinking compute under standard processes, and only about 42% of the results have been formalized in Lean.
• Major results include: a 175-page set-theoretic proof of the four-dimensional Kakeya conjecture (previously only the sticky case was known in 4D, with the lower bound stuck at 3.059), a 97-page three-dimensional maximal-function version of Kakeya (stronger than the set version proven by Wang Hong and Zahl), a complete proof of the UGC conjecture, the L=BPL derandomization, and a new O(n^(9/4)) matrix multiplication bound, among others; two manuscripts numbered 074 total 272 pages and carry a single word as authorship.
• NYU math professor Tristan Buckmaster told CNBC and argued in multiple venues that these papers "destroyed an entire research program" and ended the careers of early-career mathematicians; UT professor Scott Aaronson also joined the discussion, and Terence Tao was reported to be leading a boycott (this last point appears only in Chinese media accounts).
• The Verge interviewed more than 30 mathematicians, most of whom agreed that AI is now producing new mathematics faster than the community can understand, verify, and absorb it.

Unconfirmed

• The motive behind the release is disputed: one researcher (deredleritt3r, relayed by RexDouglass) clarified that the results were not a deliberate attempt to "destroy mathematical research" but a byproduct of stress-testing the internal model with the hardest open problems — apart from top unsolved problems like the Millennium Prize problems, other questions no longer challenge the model.
• Claims such as a "mass mathematician boycott" or "Terence Tao leading a boycott" mainly come from Chinese media aggregations; the original scope and wording require further verification.

Why it matters

• This is the first time AI has produced near-frontier new mathematics at such scale, directly disrupting the traditional academic order in which "original proofs belong to humans," especially for young mathematicians still building their careers — some have compared their situation to industrial workers hit by NAFTA.
• Nassim Taleb shared related coverage with commentary, while Perry Metzger publicly rebutted Buckmaster, showing clear divisions within academia; keunwoochoi pointed out from a technical angle that mathematical formulas are "the most autoregressive, least noisy corpus," suggesting these results could be used to reverse-engineer structural properties of mathematics itself.
• A formalization rate of only 42% means a large share of proofs still need manual verification, so mechanisms for auditing the credibility of AI-generated mathematics will become a central issue going forward.

2026-10-09 ~ 2026-10-10 · 17 related posts

Full story(15 episodes)→

• Episode 1: Debate flares over how OpenAI should release AI-generated math proofs(2026-10-06, 5 posts)
• Episode 2: Rumor: Unreleased OpenAI Model Proves Quasi-Riemann Hypothesis(2026-10-07, 3 posts)
• Episode 3: Mathematician Clarifies Overhyped Hodge Conjecture Progress by OpenAI and Anthropic(2026-10-07, 3 posts)
• Episode 4: OpenAI Paper Claims Quasi-Riemann Hypothesis Proof, Hailed as Fields Medal-Level(2026-10-07, 5 posts)
• Episode 5: Overlooked compute details of OpenAI's math research spark ASI debate(2026-10-07, 11 posts)
• Episode 6: Rumor: OpenAI Has Two More Batches of Math Results Unreleased(2026-10-07, 3 posts)
• Episode 7: OpenAI Drops 722 Math Manuscripts Overnight, Stunning Mathematicians(2026-10-08, 19 posts)
• Episode 8: OpenAI Retracts 3 Math Papers Just Two Days After Publishing Repo(2026-10-08, 8 posts)
• Episode 9: Researchers Improve OpenAI's Math Proof, Verified in Lean(2026-10-08, 2 posts)
• Episode 10: Developer extends OpenAI's math breakthrough with AI, verifies optimal result in 13 dimensions via Lean(2026-10-08, 2 posts)
• Episode 11: OpenAI's Math Results on GitHub Spark Debate Over Scientific Publishing(2026-10-09, 2 posts)
• Episode 12: OpenAI Drops 722 AI-Generated Math Papers in One Day, Sparking Backlash(2026-10-09, 17 posts)
• Episode 13: Mathematicians' Boycott of OpenAI Results Draws Backlash(2026-10-09, 3 posts)
• Episode 14: Hundreds of Mathematicians Warn AI Is Destroying Math Research(2026-10-10, 2 posts)
• Episode 15: Mathematicians Slam AI Labs for Using Them as Free Proof Checkers(2026-10-10, 3 posts)

• OpenAI's math results spark boycott calls; Buterin and Drake warn crypto at risk — Wes Roth · 2026-10-09
• NYU math prof: OpenAI's 722 AI-generated papers "destroyed" early-career mathematicians — tekbog · 2026-10-10
• NYU math professor says OpenAI's 722 AI-generated math papers 'destroyed entire research programs' — eigenrobot · 2026-10-10
• OpenAI's ~400 math results were a side effect of benchmarking its internal model — RexDouglass · 2026-10-10
• OpenAI Open-Sources Internal Model's Math Results: Proofs, Lean Artifacts and Reasoning Traces — keunwoochoi · 2026-10-10
• Math Formulas Are the Most Autoregressive, Least Noisy Corpus Ever, Researcher Argues — keunwoochoi · 2026-10-10
• OpenAI's 722 Math Papers Spark Global Backlash and Mathematician Boycott — 新智元 · 2026-10-10
• OpenAI's 272 Pages Take Aim at 4D Kakeya, Weeks After Wang Hong's 3D Proof — 新智元 · 2026-10-10
• [source] OpenAI quietly releases 722 math manuscripts produced by an unreleased internal model — FinanceYF5 · 2026-10-10
• OpenAI releases 722 math manuscripts spanning 372 results from ~4,000 tested problems — FinanceYF5 · 2026-10-10
• OpenAI drops 722 math manuscripts claiming 4D Kakeya proof; Tao-led mathematicians revolt — 创业邦 · 2026-10-10
• NYU mathematician won't feed original ideas to AI as models race to publish novel math first — rohanpaul_ai · 2026-10-10
• [source] OpenAI floods GitHub with 722 AI-generated math papers, and mathematicians say they can't keep up — 机器之心 · 2026-10-10

4 near-duplicate retellings: tedmitew · inductionheads · FinanceYF5 · zealcaiden