Post #2380713
2026-05-11 04:52 UTC
A few weeks ago, I postulated several conjectures, and my PhD student thought about them and came up with proofs or counterexamples for each of them. I was overjoyed and excited for our future work together.
I imagine a scenario where instead I had asked a chatbot to prove the theorems for me. Even if it did prove the theorems (extreme doubt), I think getting such a proof wouldn't have made me feel anything. No one's achieved anything.
Replies (2)
-
@ltchen@mathstodon.xyz 2026-05-11 06:05
@liamoc@types.pl Have you seen Gowers’ recent post? To be clear, I’m not saying LLM will solve all questions etc. I agree with you that solving a problem by a human being is more than the mere solution. https://gowers.wordpress.com/2026/05/08/a-recent-experience-with-chatgpt-5-5-pro/ My feeling so far (not yet articulated) might be summarised as “the journey itself is the destination”. Solutions might be generated magically but one has advanced by themselves to make sense of it. Some textbooks include selected solutions to exercises, but does it make solving those exercises fruitless? Certainly not.
-
@TristanNguyen@mathstodon.xyz 2026-05-11 12:47
@liamoc@types.pl Tokenmaxxing mathematics championed by a Fields Medalist. Sad!