The moments that felt like AGI to Nick Turley were GPT-4 writing poetry and producing working, compiled code — capabilities he hadn't believed were possible for a language model — and later a reasoning model spontaneously swearing in its chain-of-thought after catching its own mistake, a behavior that emerged entirely from the RL process.

59:02you feel it? I've had so many honestly. Um, and it's it's definitely not stopped. Um, a few weeks or so after I joined Open AI, GP24 had finished training >> and I remember trying it out and it actually it didn't impress me at all. Um, nor anyone else that week because it kind of didn't work. And it's because we hadn't figured out how to post train it.

59:31And uh I think seeing it go from kind of wait is this really a thing or was GPT3 kind of it to wow actually this is an entire step change with what felt to me at the time who didn't understand much about AI at all as like just some tweaks or some a little bit of final stretch work >> was profoundly humbling because you can realize that you might it might not look

59:57like we are close >> to really powerful useful AI But we probably are. Um, >> and then the moment that, you know, really, you know, there was two things that GBD4 did that felt like AGI to me. One is it could do poetry and I didn't think it was possible for AI model to do poetry. It just kind of >> fundamental philosophically, it just didn't feel like

60:19>> in scope >> and then the other one was it could produce code that actually worked and compiled. And then my next moment where I stared at the ceiling just in awe was when I realized GPD4 could just simulate an entire computer terminal um um like a full computer with commands etc. And I'm like, wait, how would this be imbued in a in a in a language model? And there's

60:41been so many moments since then, honestly. Like reasoning was a moment. One of the moments was when I think Mark and I were uh uh giving a demo of reasoning uh in front of the um whole company. Um and um this was a moment where we were still trying to kind of find use cases that were hard enough for the AI for the reasoning to, you know,

61:01make it make a difference. We're way past that point, we know. But at the time, you know, I think we were having to do a puzzle in front of everyone and um I think one of the moments that made me totally feel the AGI is like um we were in the middle of the demo and everyone started laughing and I was like, "Wait, what what is funny?" And