Acausal trade

Abstract

Acausal trade is a concept in decision theory that describes a scenario where two agents can cooperate even though they cannot communicate, influence each other, or even be certain of the other’s existence. This is possible because the agents are highly rational and can predict each other’s behavior, including their decision to cooperate or defect. This concept stems from the Prisoner’s Dilemma, a game theory scenario where two players are incentivized to defect, even though cooperation would be mutually beneficial. Acausal trade offers a potential solution to this dilemma by introducing the element of prediction and superrationality. It suggests that by reasoning about the other agent’s actions and motivations, agents can achieve a cooperative outcome despite the limitations of their environment. The article discusses several examples and objections to acausal trade, including the possibility of agents “cheating” and the question of whether agents can care about entities with whom they cannot interact. It also explores the different mechanisms by which agents can predict each other’s behavior, including the possibility of knowing each other’s mental architectures or simulating each other’s actions. The article concludes that acausal trade is a potential solution to the problem of cooperation in the Prisoner’s Dilemma, and that it is relevant to the development of artificial intelligence. – AI-generated abstract.