The difference is that these "classically trained" people are wrong. I don't think there is any historical equivalent to what we are experiencing now with LLM generated slop.