people often mistake skepticism for expertise or knowledge

This is a pretty good insight.

that's why you see so many on HN that dismiss new technologies in ML (especially in NLP, since everyone can understand the output

I think also in NLP people see output that is the same as some training data, so think it is copying it. It takes some a little bit more thought to think "ok if I asked 100 experts to try to write how to sort an array in Python" or their code is going to be very similar. This doesn't mean it is copied.