x.com/m4h007/status/2108934812849537356
2 corrections found
He was willing to explain his approach and present his proof, things that OpenAI hasn't done.
By October 10, 2026, OpenAI had already publicly released its mathematical manuscripts, supporting proof artifacts, and summaries of the model’s reasoning, so it is incorrect to say OpenAI had not presented its proofs or explained its approach.
Full reasoning
This claim is contradicted by OpenAI’s own public release from October 6, 2026, four days before this post was published.
OpenAI stated that it was publishing the results in a GitHub repository and sharing formalizations of many of the proofs in Lean. It also said it was publishing “10 summaries of the model’s reasoning” and additional details about how the results were obtained. Those are exactly the kinds of materials that count as presenting proofs and explaining an approach.
The linked public GitHub repository likewise says it contains “mathematical manuscripts and supporting proof artifacts” produced by an internal OpenAI model. It further says the collection includes hundreds of manuscripts, a preprints/ directory containing PDFs and source files, and released abridged summaries of the model’s reasoning.
So while one can argue that OpenAI’s materials still need validation, it is factually wrong to say OpenAI had not presented proofs or not explained its approach at all.
2 sources
- Sharing AI progress in mathematics | OpenAI
For this release, we’re publishing the results in a GitHub repository... As part of our GitHub repository, we are sharing formalizations of many of the proofs in Lean... We are also publishing additional details about how we obtained the results... These include 10 summaries of the model’s reasoning.
- GitHub - openai/math
This repository contains mathematical manuscripts and supporting proof artifacts produced by an internal OpenAI model... The current catalogue contains 719 manuscripts organized into 372 families... We are also releasing abridged summaries of the model's reasoning.
American mathematicians were not interested in reading it.
That overstates the reaction. Contemporary accounts show the proof quickly became a major topic of discussion, and mathematicians did in fact study the preprint, give talks on it, and write expositions.
Full reasoning
This sentence is too categorical and is contradicted by historical accounts of the reaction to de Branges’s Bieberbach proof.
A 1984 report by Paul Zorn says that news of the proof spread quickly through the mathematical world and that at the August 1984 Joint Meetings of the MAA and AMS in Eugene, Oregon, “talk of de Branges’ achievement was heard everywhere.” That is not consistent with the claim that American mathematicians were simply “not interested in reading it.”
Richard Askey likewise wrote that once a preprint was available, “many mathematicians went through the details of his argument, gave talks on it, and some wrote their own accounts.” Askey was himself an American mathematician, and his account directly contradicts the blanket statement that American mathematicians were uninterested.
It is fair to say that de Branges’s proof was initially met with skepticism and that Soviet mathematicians played an important role in validating and simplifying it. But saying American mathematicians were not interested in reading it erases documented American engagement with the proof.
2 sources
- Bieberbach Conjecture Proved
News that the conjecture had been verified spread quickly through the mathematical world last spring and summer as de Branges lectured in Europe and as preprints and informal communications circulated. At the August 1984 Joint Meetings of the MAA and the AMS in Eugene, Oregon, talk of de Branges’ achievement was heard everywhere.
- History and Philosophy of Modern Mathematics, Vol. 11 (snippet quoting Richard Askey)
After a version of de Branges’s proof was available in preprint form, many mathematicians went through the details of his argument, gave talks on it, and some wrote their own accounts.