Hmm
Did musicians stop making music when computers got good? Did chess players stop playing human chess when computers got good? Did runners stop running themselves when cars were invented? Did programmers stop programming when computers got good? Did dishwashers stop washing dishes when dishwashers got good? What is the right comparison? I don't know, I'm just asking questions.But in fact I am no longer asking questions: For the time being, I'll remove the list, as I personally do mostly feel like a question solved fully by a machine is a question wasted. As far as I understand, it is not uncommon nowadays for people to scour the internet for math questions to solve. Good for them, but I don't really want my list to become cannon fodder.
I suppose my current view is that at some point it would make sense to separate different categories of mathematical work based on the amount of computer involvement. I propose a classification:
- Fully computer-solved math. Presumably soon the only source of actual frontier math.
- The slippery slope between.
- Artisan math. I'm thinking good old-fashioned work in discussions, on a blackboard, any algorithmic assistance from programs you wrote, and special-purpose neural networks that you trained, possibly using LLMs to help you learn to do these things.
- Free solo. A controlled exam where you sit 3 years in a room, solving one really hard question with no aid, where it's not known whether a solution exists.
Oh and if you want my list, just ask.