Decision Theory: Value in Time — LessWrong