A decision theory idea I just had, which may or may not grow into something interesting.
Sometime ago I proposed to evaluate logical counterfactuals by their proof length. At the September workshop we managed to develop that idea into a full candidate solution to the problem of logical counterfactuals. Another long-standing open problem is “who moves first” in timeless negotiations. Could that problem also be solved by proof lengths? For example, do we feel that a “defecting rock” is impossible to manipulate because there are short proofs about it?
A decision theory idea I just had, which may or may not grow into something interesting.
Sometime ago I proposed to evaluate logical counterfactuals by their proof length. At the September workshop we managed to develop that idea into a full candidate solution to the problem of logical counterfactuals. Another long-standing open problem is “who moves first” in timeless negotiations. Could that problem also be solved by proof lengths? For example, do we feel that a “defecting rock” is impossible to manipulate because there are short proofs about it?
Hm. I feel like “impossible to manipulate” just means that you can prove that it will never cooperate when the opponent will defect.
But yeah, if we equate “acting first” with acting in ignorance of the other person’s move, then we get something interesting.