For example, I just decided that the symmetric Prisoner’s Dilemma should be introduced quite late in the text, not near the beginning as I thought. The reason is that it’s tricky to formalize, even if you assume that participants are robots.
“Robot, I have chosen you and another robot at random, now you must play the PD against each other. My infallible predictor device says you will either both cooperate or both defect. What do you do?”—The answer depends on how the robots were chosen, and what the owner would do if the predictor didn’t predict symmetry. It’s surprisingly hard to patch up.
Can you give me some examples of those exercises and loopholes you have seen?
For example, I just decided that the symmetric Prisoner’s Dilemma should be introduced quite late in the text, not near the beginning as I thought. The reason is that it’s tricky to formalize, even if you assume that participants are robots.
“Robot, I have chosen you and another robot at random, now you must play the PD against each other. My infallible predictor device says you will either both cooperate or both defect. What do you do?”—The answer depends on how the robots were chosen, and what the owner would do if the predictor didn’t predict symmetry. It’s surprisingly hard to patch up.