Demis Hassabis: preference

Transcript context

…Is there any potential for the AGI to eventually come from a pure RL approach? The way we’re talking about it, it sounds like the LLM will form the right prior and then this sort of tree search will go on top of that. Or is it a possibility that it comes completely out of the dark? Theoretically, I think there’s no reason why you couldn’t go full AlphaZero-like on it. There are some people here at Google DeepMind and in the RL community who work on that, fully assuming no priors, no data, and just building all knowledge from scratch. I think that’s valuable because those ideas and those algorithms should also work when you have some knowledge too. Having said that, I think by far the quickest way to get to AGI, and the most plausible way, is to use all the knowledge that’s existing in the world right now that we’ve collected from things like the Web. We have these scalable algorithms, like transformers, that are capable of ingesting all of that information. So I don’t see why you wouldn’t start with a model as a kind of prior, or to build on it and to make predictions that help bootstrap your learning. I just think it doesn’t make sense not to make use of that. So my betting would be that the final AGI system will have these large multimodal models as part of the overall solution, but they probably won’t be enough on their own. You’ll need this additional planning search on top. This sounds like the answer to the question I’m about to ask. As somebody who’s been in this field for a long time and seen different trends come and go, what do you think the strong version of the scaling hypothesis gets right and what does it get wrong? The idea that you just throw enough compute at a wide enough distribution of data and you get intelligence.…

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