[simplemathmind]

I did not read the pdf, but I think that LLM and Lean could be useful tools for mathematiceans to prove or refute theorems, but the creative idea that sparks knowledge and new theorems lies in the human, the others are tools that can help to reduce time and effort needed and so, indirectly, they can foster and enhance creativity. It also could mitigate some reasoning that require mechanical prove of many details. Anyway, what I just said seems simple and clear, and in no way would it be worth to be published in a high ranking math journal.

[generationP]

Right, and the paper you didn't read contains a lot more than that.

[simplemathmind]

I apologize in that case, can you give a brief summary of what is the most important point of that pdf?, I don't know why but my gut feeling is to refuse to read something just based on people reputation. I know that Tao is a very bright mathematician but I also think that he doesn't know much about computer science or computer languages (my evidence is very slim here: once I noted that Tao was very happy with a very simple program, a trivial one, so I inferred from that he still has to learn a lot about programming). For now, it seems clear that the best he can do (for him and us) is to devote his time to math that is his best skill. But it could happen that he could use his best world IQ and math skills to learn how to use LLMs and Lean in a never seem before way to obtain something really valuable.

Today I don't think there is any evidence that such thing is going to happen. On one hand, in general, intelligent people are the first to learn how to use new tools in new or better ways, tools that are useful for what they are good at and are devote at. On the other hand, following that path detracts energy from the core of math that requires intuition and creativity and not so much mechanical proofs. On a third hand, there is always the money question that we can not see, that is because is in the third hand.