The author makes other mistakes too such as thinking history that was already written is an absolute truth when it's actually something subjective written by the "winners." I lived through that once, the history that I was taught as truth in my school many years ago was proven to be false a few years ago and the lies were based on what today you'd consider hate speech.
I think the moment we can solve detecting hate speech using a machine will be the same a machine can pass the Turing test because it's impossible to define hate speech in objective terms. Meanwhile we have two equally incomplete solutions: allow all speech or let some people subjectively censor things. Neither of those things will satisfy the author of this post (well maybe the latter will).