Unraveling the impact and emergent features of powerful AI-based systems
Interesting topic and so rich references!
what do you think about the view about The sudden increase in power of large model is more like a phase transition than an emergence? I mean , of course there are some kinds of emergencies in concept, but because of the vague definitions of emergence, it seems phase transition is a better description and We might have to figure out what's breaking symmetry in the model, which seems like a more definite well question.
btw, I noticed you mentioned the work of Rosas and Erik Hole in your review, I wanna know what's your personal opinion with these existing theories about emergence.
I know they are not easy to answer here. In fact, the true purpose of my comment is to thank you for your updates! I'am always following closely because it is all interesting topic for me!