This is an excellent description by Cal Newport, who I hadn't heard of until now. He is a MIT-trained computer science professor at Georgetown University & Bestselling Author. I'm always impressed by New Yorker content.

I would add a couple of things. William S. Burroughs and Brion Gysin came up with some interesting results with their cut up technique invented by Gysin, with random sentence fragments juxtaposed and massaged by them into sensible sounding sentences.

https://www.briongysin.com/cut-ups/

Also, I think the kind of mind it has is a subjective experience in part, because artificial intelligence is like artificial grass and if you have no experience with real grass or real intelligence you tend to think the artificial kind is more real than it is. I think we tend to anthropomorphize our intelligent programs, more so if they are smarter than us we have a more difficult time detecting their machine flaws.

So one person might say "that thing is really bright and has a mind of it's own" while a computer scientist might say of the same thing "I can see it's machinations failing in some cases in stupid ways that indicate to me it's just a dumb program"