The Sustainability Implications of Language Models

I think of this stuff as being, I've always liked the framework that some maybe more agnostic researchers have taken here. There's like the sort of true belief of you, which is all we need to do is scale these up and they'll be completely sentient. And there's kind of almost a view in the middle, which says we've got a few more big things to discover, but these will be part of it. That also touches upon another aspect of the AI ethics aspect. The whole team at Gabriel's story last year was a big deal for good reason.