Yeah, I've read that one and I probably should have come up with a way to integrate it. It's certainly more nuanced, and oddly enough it does a better job of arguing for why AI-generated art would be appealing, perhaps because it doesn't take that appeal for granted.

I think it matters just how big a leap the bowls are over what was available before, though. All of a sudden, society goes from crappy implied-to-be-around-2010s-tech-level AI artists to having AI artists that know individual authors better than those authors know themselves. A big part of the problem of bowl art is that you can ask a bowl for a particular author's *future* output *and get the correct answer*. I don't get the impression that that's a near-term problem we're facing: it seems like the software currently coming down the pipeline won't even give *consistent* answers to that question, let alone correct ones. (Where would it even get the data for that accurate of a simulation?)

(I do find it odd that the characters seem to take for granted that bowl simulations aren't sapient, instead of being uncertain and being deeply concerned by that uncertainty.)

((*looks through that page I pulled off a search engine as an example of the concept (because I'm not sure exactly which page about it I read)*

"Literally nobody outside of MIRI or FHI ever talks about this problem."

Yeah, okay, maybe that's why the characters didn't think of it.))