There’s an argument for cooperating with any agent in a class of quasi-rational actors, although I don’t know how exactly to define that class. Basically, if you predict that the other agent will reason in the same way as you, then you should cooperate.
(This reminds me of Kant’s argument for the basis of morality—all rational beings should reason identically, so the true morality must be something that all rational beings can arrive at independently. I don’t think his argument quite works, but I believe there’s a similar argument for cooperating on the prisoner’s dilemma that does work.)
There’s an argument for cooperating with any agent in a class of quasi-rational actors, although I don’t know how exactly to define that class. Basically, if you predict that the other agent will reason in the same way as you, then you should cooperate.
(This reminds me of Kant’s argument for the basis of morality—all rational beings should reason identically, so the true morality must be something that all rational beings can arrive at independently. I don’t think his argument quite works, but I believe there’s a similar argument for cooperating on the prisoner’s dilemma that does work.)