Bartosz Milewski on Nostr: Interesting problems might be the ones whose formulation and/or proofs can be ...

Interesting problems might be the ones whose formulation and/or proofs can be efficiently compressed. Famously, the Fermat's last theorem is very compact. So is Riemann's hypothesis. The Langland's program attempts to compress number theory and geometry into one package. And so on. So a good AI prompt might be to search for problems with low Kolmogorov's complexity.