You've dropped the important part of the title (OpenAI, mathematicians), and kept the clickbait.. how about: OpenAI's blow to mathematicians: 'They have annihilated my field of research'
* Don't release a slop paper that humans can't understand, in a obvious attempt to claim credit, even though they verbally deny it.
* Announce that the model has solutions to these problems, and work with mathematicians on how to release them.
* Do whatever Anthropic is doing. If you have to train them, keep it completely private.
The idea that OpenAI couldn't do anything differently is ridiculous. A few days ago they released a slop solution to ETH -- we now know it was an attempt to outrun Anthropic, which independently found the solution, but was busy tidying it up with mathematicians, so humans could consume it. How silly of them! That's why Sam Altman is such a winner.
Have people who are genuinely interested in the research they are doing. And let them polish the manuscripts to the level of clarity generally expected from publishable results.
The vast majority of the results they have released so far have no practical value beyond marketing. And that marketing aspect didn't work too well, because they didn't understand the target audience.
I don't see how OpenAI owe the level of interest or polish similar to human mathematicians to anyone.
1. OpenAI do model training on verifiable tasks that push the boundaries.
2. The work from the previous step produces groundbreaking results.
3. OpenAI can keep the unpolished results private, publish them or branch off a mathematics research arm which produces human-level quality of output. I think they should be free to not do that last one.
And it's been less than a month, it would be shocking for any of it to have practical value, even when we take into account the fact that most new mathematical proofs have no practical value whatsoever for a very long time. Take the previous few hundred proofs produced by humans recently and show me their practical value.
The novel thing is the production of proofs in a way that isn't coupled with the simultaneous expansion of human understanding. Mathematicians are massively affected in how they work and what their careers will look like but they're not victims that OpenAI has somehow maliciously targeted like some of the discourse seem to imply.
The question was what OpenAI could have done differently, so that their results would have been received better. Whether they owe anything to anyone is an unrelated discussion.
I've heard claims that OpenAI has spent hundreds of millions in pure mathematics research this year. It's difficult to understand what they are trying to achieve. They can't possibly expect results with sufficient practival value in the timescale the company cares about. And it doesn't work as a marketing campaign, as they failed the marketing part, either due to ignorance or arrogance.
7 comments
[ 2.0 ms ] story [ 16.3 ms ] threadRecent related:
OpenAI Withdraws 3 Math Papers (382 points, 24 days ago, 612 comments) https://news.ycombinator.com/item?id=50002650
A look at the math that OpenAI claims computes faster Fourier transforms (249 points, 19 days ago, 170 comments) https://news.ycombinator.com/item?id=50065932
Simons Foundation launches $15M effort to help mathematicians navigate AI era (249 points, 16 days ago, 201 comments) https://news.ycombinator.com/item?id=50104654
"I'm on a war footing": Terry Tao and the crisis in math (230 points, 15 days ago, 541 comments) https://news.ycombinator.com/item?id=50118165
Why Math Dumped OpenAI (92 points, 15 days ago, 14 comments) https://news.ycombinator.com/item?id=50128468
The math that OpenAI keeps for itself (54 points, 14 days ago, 19 comments) https://news.ycombinator.com/item?id=50134211
OAI-math-revised: a community hub for revising OpenAI's math write-ups (53 points, 7 days ago, 43 comments) https://terrytao.wordpress.com/2026/10/24/oai-math-revised-a...
AI use leads UC Berkeley math professor to see 'no point' in grading homework (240 points, 6 days ago, 342 comments) https://news.ycombinator.com/item?id=50233304
The Exponential Time Hypothesis Is False [pdf] (802 points, 5 days ago, 301 comments) https://news.ycombinator.com/item?id=50251661
1. Not train models to get better on certain verifiable tasks like math proofs?
2. Train models to get better, produce groundbreaking proofs and keep the results private?
3. ?
* Announce that the model has solutions to these problems, and work with mathematicians on how to release them.
* Do whatever Anthropic is doing. If you have to train them, keep it completely private.
The idea that OpenAI couldn't do anything differently is ridiculous. A few days ago they released a slop solution to ETH -- we now know it was an attempt to outrun Anthropic, which independently found the solution, but was busy tidying it up with mathematicians, so humans could consume it. How silly of them! That's why Sam Altman is such a winner.
The vast majority of the results they have released so far have no practical value beyond marketing. And that marketing aspect didn't work too well, because they didn't understand the target audience.
1. OpenAI do model training on verifiable tasks that push the boundaries.
2. The work from the previous step produces groundbreaking results.
3. OpenAI can keep the unpolished results private, publish them or branch off a mathematics research arm which produces human-level quality of output. I think they should be free to not do that last one.
And it's been less than a month, it would be shocking for any of it to have practical value, even when we take into account the fact that most new mathematical proofs have no practical value whatsoever for a very long time. Take the previous few hundred proofs produced by humans recently and show me their practical value.
The novel thing is the production of proofs in a way that isn't coupled with the simultaneous expansion of human understanding. Mathematicians are massively affected in how they work and what their careers will look like but they're not victims that OpenAI has somehow maliciously targeted like some of the discourse seem to imply.
I've heard claims that OpenAI has spent hundreds of millions in pure mathematics research this year. It's difficult to understand what they are trying to achieve. They can't possibly expect results with sufficient practival value in the timescale the company cares about. And it doesn't work as a marketing campaign, as they failed the marketing part, either due to ignorance or arrogance.