Axel Højmark: prediction

Transcript context

…Mythos is the 1st 1 where it seems to have insane power to do unexpected things. Mhmm. And, like, did you think basically that we're pretty close to to the limit here? Like, you know, yes. Because at 1st I thought, yeah, it was bad. I I it it safety washing. You know, they they shouldn't they shouldn't have stopped it from coming out. And now I've actually had access to it. I'm actually thinking, oh, shit. Maybe they had a point. Where where is this gonna go? Yeah. I mean, I think Dario has this nice single sentence that we're at the end of the exponential, which I really like, in the sense that essentially what's gonna happen or, like, what seems to be what all the labs are planning for is taking these models and using those models to help build the next generation of models, and then that generation of model is more capable and can help build the next next generation, and so we're gonna get this, like, we may not have that much time before we have AIs that are smarter than all of humanity combined, for example. So we Yeah. We have a lot of safety work to do. Yeah. I mean, I I really like what you guys are doing. So you you you now have a quantitative way of estimating the reward seeking behavior in in a production model, and and you're advocating that the Frontier Labs use this as part of their red teaming and their system cards and whatnot. That that that's all great. But I I suppose the problem is we live in a globalized world, and China is still building all this technology. Like, realistically, how what would it take for people to stop?…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.