How Bayes' theorem is consistent with Solomonoff induction — LessWrong