There's a fundamental hand wave going on in this article.

First, we're led to believe that LLM "AI" is entirely a new, revolutionary thing

Then, it waves off objections by saying that people made objections when compilers were introduced.

I think you can't claim "completely revolutionary" and then not deal with all observations, or at least you can't reject observations and objections based on some historical introduction of technology. If "AI" is that great, then you can deal with all objections.