Pavitra comments on Open Thread: September 2011 - LessWrong

5 Post author: Pavitra 03 September 2011 07:50PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (441)

You are viewing a single comment's thread. Show more comments above.

Comment author: Pavitra 09 September 2011 07:03:45PM 1 point [-]

Maybe I have something different in mind when I say "status". To the extent that an agent's treatment of an opponent is based on some sort of persistent state, rather than on the opponent's local behavior, I would say that that state is a form of status information. (Thus, one can have a status with a grim trigger, but not with a tit-for-tat.)

In humans (and I conjecture in approximately-winful strategies generally?), status tends to be one-dimensional, with one end eliciting "do nice things for this person" behavior, and the other end eliciting "do mean things to this person" behavior.