This article does raise interesting points, and Nate Silver does stretch the power of Bayesian statistics a little bit. However, the crux of his argument for Bayesian statistics is that it provides a process by which we can approach truth asymptotically.<p>The beautiful thing about bayesian probability is that the first experiments conditional probability is the next experiment's bayesian prior. We acknowledge the epistemological fact that our initial understanding of the world is flawed, and correct it from there.