The Goal of the Bayesian Conspiracy — LessWrong