Exactly. This is part of the reason I will win the bet, i.e. it is the reason the first super intelligent AI will be programmed without attention to Friendliness.
Unfortunately being right isn’t sufficient for winning a bet. You also have to not have been torn apart and used as base minerals for achieving whatever goals a uFAI happens to have.
This implies that you have some other strategy for winning a bet despite strong UFAI existing. At very least this requires that some entity with which you are able to make a bet now will be around to execute behaviours that conditionally reward you after the event has occurred. This seems difficult to arrange.
Actually, all he needs is someone who believes the first AI will be friendly and who think they’ll have a use for money after a FAI exists. Then they could make an apocalypse bet where Unknowns gets paid now and then pays their opponent back if and when the first AI is built and it turns out to be friendly.
Unfortunately being right isn’t sufficient for winning a bet. You also have to not have been torn apart and used as base minerals for achieving whatever goals a uFAI happens to have.
True, that’s why it’s only a part of the reason.
This implies that you have some other strategy for winning a bet despite strong UFAI existing. At very least this requires that some entity with which you are able to make a bet now will be around to execute behaviours that conditionally reward you after the event has occurred. This seems difficult to arrange.
Actually, all he needs is someone who believes the first AI will be friendly and who think they’ll have a use for money after a FAI exists. Then they could make an apocalypse bet where Unknowns gets paid now and then pays their opponent back if and when the first AI is built and it turns out to be friendly.