They didn't try to scoop another group; they thought the other group had already solved it, so maybe their latest model could take a shot too... and the model solved the full problem when the other group had not! They found out after the fact that the other group had only solved an important sub-problem.

This was a low-key hilarious replay of George Dantzig and his homework problems: https://en.wikipedia.org/wiki/George_Dantzig

The controversy was whether they had plagiarised that other work on the sub-problem, which they categorically denied after an investigation. And yes, a few days to investigate something like this is reasonable for a company as big as OpenAI. Having seen how data infra is set up when petabytes of data are flowing about, there are thousands of entwined data pipelines to figure out. Not quite as easy as running a query on a sqlite DB!

> Believing that OpenAI's claims have any substance cannot be explained by naivety alone.

Yes, they could be explained by a GitHub repo full of proofs :-)

Or are you suggesting there were hundreds of researchers who just happened to be close to solving hundreds of these long standing open problems using Codex, and OpenAI swooped in plagiarized them all? ;-)

Buckmaster himself said essentially all the progress they made was from July 15th onwards (with a new model on the problem) with no real progress prior to then.

I think it’s really cope to claim this was a human result being stolen.