Jamie Brew’s “predictive text” stuff (the cookbook one, the Strunk and White one I just reblogged) is way funnier and more grammatical than usual Markov chain output, which is what computer-generated text on the internet usually is.  It looks like it has learned vocabulary and grammatical structures from the source text and is generating new text using them, with awareness of things like maintaining an overall topic across multiple sentences.

This definitely isn’t impossible, but given that I can’t seem to find his code anywhere (or even information about what algorithms he’s using), and that “predictive text” is an odd and inaccurate way of describing this kind of program, I am skeptical that he’s actually producing these with a computer.

ETA: someone asks about the code and he says it will be available in a couple of weeks once he learns how to use github, and that he’s “newish to coding.”  This sounds pretty iffy.  (New to coding, hasn’t used github, but can do advanced NLP?)