Hi Trevor, thank you for your response!
Actually, I knew very soon that AI was not what I expected before entering the field, but why would I leave immediately? I wanted to learn as much as I could.
ML=DL=AI isn't something I believe, it's what's often conveyed by the mass media. I disagree with it (I think it's clear in the article).
I agree with what you say between novices and experts and the generalized nature of the issue. However, when there are important interests behind, this situation can be aggravated. AI is not just another field. This happens everywhere, but it doesn't happen everywhere to the same degree.
I agree that not many people can do true AI (actually, I don't think I can). My first sentence in that section is "it's hyperbole." It's an exaggeration, but within the tech world, AI seems to be everywhere.
About reinforcement learning, I don't think it's what "most closely resembles the grandiose promises of creating generally intelligent agents." And I think most experts wouldn't agree either. Probably self-supervised learning is what's closer (without taking into consideration developmental robotics and that type of sub-fields). The problem with RL is that the environments in which systems are trained are very simplified versions of reality. RL is not different from ML, is a paradigm within ML.
I agree with your last paragraph. If those young learners aren't disillusioned, then they won't change their minds by reading my article. If they knew what AI's state of the art is all along, they already know what I've written here.
I enjoyed your comment, and always appreciate a nice debate, Trevor. Hope to see you around again!