Actually, I agree with you. What I criticize is not the fact that we are using DL, but the fact that many people are confusing it with real AI. ML/DL serve a very specific purpose very well. AGI, instead, is just an abstract concept for now.
What I mean in the sentence you highlighted reflects my hunger for knowing more about the human mind. I don't want us to settle for today's DL thinking that it's as far as it gets.
For me, DL is useful. AGI is interesting. Right now the other way around isn't true for neither.