But all the people putting money into stocks, provide the market makers for the people with better information to make money.
If you look at the actual scenario there, the game was essentially in a stalemate, where the only possible way to win was to force the other player to advance a pawn. Stockfish can't look 30 moves ahead to see that it's possible to do that, so would have just flailed around.
You still need stockfish, because without it, any move you make could be a tactical error which the other players computer would pounce on. But stockfish can't see the greater strategic picture if it's beyond its tactical horizon.
Interesting. Note that Jon Edwards didn't win a single game there via play - he won one game because the opponent inputted the wrong move, and another because the opponent quit the tournament. All other games were draws.
Presumably you could test this directly. Have 100 players play a correspondence game against each of the top 3 chess engines, where you gave them a max amount of compute time they could use (say no more than 2 hours per move) and they just play directly against the computer.
My guess is they would have a slight advantage, at least over stockfish, if only by exploiting known bugs/adversarial situations.
There is a time abyss. Anatomically modern humans have been around for 300k years. That's enough time for several repetitions of our path from early agriculture to computers.
That seems unlikely - our genetic diversity provides evidence of how many humans there were at any point in the past. We would notice if there were billions of humans only a couple of hundreds of thousand years ago. See https://en.m.wikipedia.org/wiki/Prehistoric_demography.
Furthermore what would they have powered their industrial revolution with? How come there's still so much surface level coal and oil? A few hundred thousands of years shouldn't be enough to regenerate that.
And before humans, there could have been other sufficiently brainy species (e.g. Troodon dinosaurs).
I'm guessing that we would still be able to notice that in the distribution of minerals - e.g. maybe all the high iron concentration ores should have been mined, but I don't really know much about this.
It's not dominated - holding all other players constant the two strategies have equal payoffs, so neither dominates the other.
I think this speaks of a move from focusing on building the largest models possible, to optimising smaller models for mundane utility. Since smaller models have a hard ceiling on capabilities, this is good news as it pushes off AGI.
I think this is a reasonable idea.
However I think it's wonderful that lesswrong has a norm of offering monetary bounties for work. This is a norm we should encourage, and I'm worried introducing non-monetary bounties will make monetary bounties less common.
Who says what an AI would want to do is illegal? Not only is it legal to build killer robots and adapt viruses to be more lethal, you can even get government funding for both.
How come you're trusting essentially random internet strangers to pay up significant sum of money if they lose a bet in up to 5 years?