The core contrast was between a community which claims to "arrive at confident conclusions rationally, by applying Bayes' Theorem." They're a bunch of non-mathematicians who ONLY use that old theorem, by itself, without any confidence intervals or statistically significant sample sizes. Dirichlet Process is mentioned in contrast, though MUCH has been done in statistics in the 260 years since Bayes' Theorem; that's why I also mentioned SAS. Dirichlet Process is about sampling from potential populations, to adjust the estimate for the population, instead of naively assuming that "the sample distribution is the best guess of the population distribution" which that "Rationalist" community assumes in error.