AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
The Problem With Bayesian Inference
Departures from Bayesian inference are not irrational as the behavioral economists would have us believe in a non-frequentist world. And yet there are still ways of judging which narratives we should have more confidence in than others using some sort of correct process. So to gain confidence is not of zero, one sort of choice. There's always degrees of confidence, which might think we're back to base and kind of the Bayesian logic is fine. But they are never actually true once and forever. That is, there'll be a context in which most scientific findings don't work.