Logical Uncertainty and Functional Decision Theory — LessWrong