I’d like to propose a more modest version of Aumann’s agreement theorem, call it Aumann’s Less-Than-Total-Disagreement Theorem, which says that two rational agents shouldn’t both end up with 99.9…% confidence on opposite sides of the same problem.
So, it seems that actually not both of these people need to be so confident -
If one says “I give a 99.9% chance that AGI won’t happen”, the second person doesn’t need the same confidence in the opposite direction, just the same confidence in that the first person shouldn’t be so confident. thus (if the first disagrees with him), we again end up in a situation where two people are very confident in the opposite direction, first person thinks with a 99.9% certainty that he should be that certain about AGI, and the second person thinks with a 99.9% certainty that he shouldn’t.
Though i’m kinda confused now, cause it seems that if BOTH now need to not be so confident, than it seems that once the second person lowers his confidence, the other person can pump it back up, and maybe it goes on like that forever. frankly, i don’t understand Bayesian math enough to answer myself.
(Also, is there a reason there are almost no comments on these posts?)
So, it seems that actually not both of these people need to be so confident -
If one says “I give a 99.9% chance that AGI won’t happen”, the second person doesn’t need the same confidence in the opposite direction, just the same confidence in that the first person shouldn’t be so confident. thus (if the first disagrees with him), we again end up in a situation where two people are very confident in the opposite direction, first person thinks with a 99.9% certainty that he should be that certain about AGI, and the second person thinks with a 99.9% certainty that he shouldn’t.
Though i’m kinda confused now, cause it seems that if BOTH now need to not be so confident, than it seems that once the second person lowers his confidence, the other person can pump it back up, and maybe it goes on like that forever. frankly, i don’t understand Bayesian math enough to answer myself.
(Also, is there a reason there are almost no comments on these posts?)
They are reposts from slatestarcodex.com.