(Downvoters: Did I read his essay incorrectly? The reddit post from 1 year ago and the github post today seem be the same theme of managing the community expectations of Cognitec.)
Rich put it well that Clojure is not closed, it is conservative. He made it to solve the problems he encountered in the industry, which were large multithreaded proprietary systems. It is a business first language (cognitect is a consultancy, after all). The fact that so many hobbyists like me ended up using it was kind of a happy accident.
These two audiences don't have completely opposed interests, but they do have different priorities. Businesses care about stability; they don't care much about whether the language accepts PRs on Github or conducts twitter polls. I accepted that a long time ago, and I'm still using it six years later.
And it is a little bit of that; but it's also a lot more of an intentionally conservative, even boring, extremely practical tool for building real things in a concise manner. In this way it fits into the Java ecosystem well, alongside many other very boring, robust, proven, reliable chunks of code.
Much of my work today is in the Node/NPM/JS ecosystem; although that has its advantages also, some days I really miss the boring reliable conservatism of Java and Clojure!