This is an awful take, and I hope other readers ignore it, and give the podcast a listen/read. It's a long podcast, and if you skim the transcript you can see

This is an awful take, and I hope other readers ignore it, and give the podcast a listen/read. It's a long podcast, and if you skim the transcript you can see that the discussion doesn't start until much later. I'm a pretty strong Bayesian, and have heard more than my fair share of vague, hand wavy, and stubborn frequentist arguments against Bayesian statistics. When I see an "Against Bayesianism" rant, I'm already biased against it from seeing so many awful arguments thrown out there, mostly to troll Bayesians. This is absolutely not one of those. This is a very thoughtful and clearly articulated discussion of the applications of, what Deutsch easily agrees, is a correct statistical methodology to larger epistemological questions. It is long, so I only had a chance to skim this but it is incredibly obvious that David Deutsch is not "seeking attention", but has very legitimate concerns with the mindless application of Bayesian reasoning to larger epistemological problems in science. I'll certainly be revisiting this later for a closer listen.