x
Why (and why not) Bayesian Updating? — LessWrong