I think that AlphaProof and other Theorem proving AI will become super-human in the next 15 years. They already can solve various IMO problems and there is no reason to think they do not scale.

@Grothenfla I mean, I certainly wouldn't count out the possibility. But nuclear power and self-driving cars didn't scale yet, and I don't think that solving all of mathematics is just like solving various IMO problems. We have robots that can clean rooms but they can't pick up everything in your desk and sort it. So it's not obvious to me that this will happen in the next 15 years. I'm not saying it definitely won't happen. I'm a NO here.

P vs NP alone makes this functionally impossible

@traders I've put some NO limit orders.

@sbares

@firstuserhere Mana where my mouth is.

So people don't think a superintelligence will be able so solve all the Millennium problems? Or are the markets just mispriced.

@TimothyCurrie I think your linked market is laughably overpriced and I have guesses on why.

@justifieduseofFallibilism Super intelligence is an extremely vague term and there's no resolution criteria.