This repository was created as a public square for collecting and sharing trustworthy commentary on OpenAI's mathematics solution dump, to help mathematicians avoid retreating to closed-door discussions or being misled by online rumor mills. The goal is to amass the collective impressions of mathematicians in their own areas of expertise.
To start, open the list of pull requests above. Each manuscript group has its own pull request for discussion. For instance, commentary about the quasi-Riemann hypothesis can be posted under #3. You can search by group number or keywords.
Moreover, each manuscript addition or update has its own commit, and you can click on a commit and leave a comment somewhere specific in a tex file, e.g. you can click on the second commit, 3ac0106.
You must log in to GitHub to leave a comment. You can leave a general comment directly on the pull request or comment on a particular line or section of the manuscript source. For pull requests with thousands of Lean files, the easiest way to comment on a single manuscript is to click on the commit that added or last updated the file.
DON'T BE A MEAT PROXY: LLM-generated comments are not permitted and will be removed. Write your own commentary, and please be respectful and act as professional representatives of the mathematical community.
The goal of this repository is not to provide referee reports or in-depth mathematical discussions of the OpenAI papers. We are more interested in the meta-mathematical impact of these papers and their impact on experts who recently worked or were actively working on these problems. For example, comments on any of the following would be very welcome:
- General belief in the correctness. It is ok to say, without any commitment, whether the paper seems valid, and your level of confidence.
- Nature and context of the proof. Is it a synthesis of existing technology ("tying up loose ends"), and were humans already close to this result? Is this a breakthrough which will change the techniques used in the field?
- Nature of the exposition. Can experts who have worked on the problem understand the solution?
- Did you/your student get scooped? How far along were you? Were you using AI on the problem?
- Source of the ideas. Are there uncredited sources?
The questions above concern facts and expert judgments, and similar questions are welcome. It would help to structure your comments in two sections. In the first, please state the facts as you see them and distinguish them from your judgments or impressions. In the second, you might record your feelings on the matter.
Keep in mind that there may be only one or two people in the world who can accurately answer these questions. If you are one of them, feel free to correct or challenge a previous statement. Please keep the discussion cordial and professional.
Suggestions, bug reports, and contributions should be sent by email.