Incentive compatibility and the Revelation Principle — LessWrong