Iterated prisoner's dilemma allows a sort of "communication". As the number of iterations grows, the cost of losing each individual round becomes negligible in the long run and agents can learn to use their decisions (COOPERATE or DEFECT) as a binary communication channel. So instead of saying "let's cooperate" over some side-channel, an agent indicates its intention to cooperate by simply cooperating.
In iterated prisoners dilemma and other similar games, the "API" with which agents interact with the world is extremely simple. The statement of the problem is also very simple. The agent itself can be any computable algorithm for deciding to cooperate or defect based on the past history of game rounds. I find it interesting to see agents learn recognizable behaviours like "communication" or "trade" when they aren't explicitly programmed to do those things.
In iterated prisoners dilemma and other similar games, the "API" with which agents interact with the world is extremely simple. The statement of the problem is also very simple. The agent itself can be any computable algorithm for deciding to cooperate or defect based on the past history of game rounds. I find it interesting to see agents learn recognizable behaviours like "communication" or "trade" when they aren't explicitly programmed to do those things.