[mxcrossb]

> Well … OK I guess? But, pretty much exactly as pointed out last time, not only is the proof only one line, but the nature of the proof makes clear exactly how unoriginal this is to mathematics

This is what I was wondering about while reading the article. If the AI only generates formula for which proofs involve only a few trivial steps back to something that is known, then it doesn’t feel useful. But I feel like the question “what makes a good conjecture?” in its own right makes for a very interesting discussion.

[prof-dr-ir]

Mathematical physicist Robbert Dijkgraaf has got you covered:

https://www.quantamagazine.org/the-subtle-art-of-the-mathema...

[bawolff]

Wouldn't a good conjecture be anything that's interesting if true. General bonus points for if intuitively it seems like it should be obviously true (or false) but yet is hard to prove or if proving it is true would allow you to prove lots of other interesting statements.

[alisonkisk]

Sure. Define "interesting".

What's interesting to me is probably in a standard textbook already.

[bawolff]

If it is in a standard textbook, than its almost certainly interesting (although probably not a conjecture unless its a pretty advanced textbook)