I would only quibble with the failure to focus on variation across subfields in economics. This is important not for blame assignment, but because the existing equilibria differ across fields as do the possibilities for credible empirical work. There can be no "credibility revolution" in the study of business cycles, for example, because we will never have plausible exogenous variation in them. That does not mean that we should give up, but it does mean that macro is necessarily different from, say, labor economics or the micro part of development economics (which, at present, is essentially all of development economics, though that was not the case in the past).
I also liked the response by Ashwin Parameswaran.
My original response is here, which includes a link to the Thoma piece on professors and practitioners.