A friend who works at a frontier AI lab wants to bet me 20k that AI will solve nine out of every ten open maths problems before 2027 (we're in the process of agreeing upon a list of these problems). I think this is easy money, but he seems insanely confident. Am I going to lose 20k? I can still back out.
>>17069259You should agree to that bet, but you should pay him 100 a month, because 20k might be too much.
>>17069259easiest 20k he'll ever make
>>17069259I don't think it's possible anymore for a problem to be famous and not be solved by AI since every open problem gets solved by incremental progress and when enough increments are made, an AI company will sweep in and claim credit. Not saying what AI is doing isn't impressive, but the idea of any human laying the final piece is no more.
Definitely not before 2027.It's good only at discrete math. It can't do anything else so far. Ho for it.
>>17069259>I can still back out.You are retarded if you don't
>>17069259Personally, I think the chance before 2027 is quite low (assuming you didn't mean "by the end of 2027), but not low enough that I'd comfortably bet 20k. If I was a millionnaire I would take that bet, sure, but it's a major investment decision to me.
