Blaise Agüera y Arcas: evaluation

Transcript context

…There seems to be a bit of attention because the GoFi folks, they had some very interesting ideas. I mean, I'm a big fan of Fodor and Polition, for example, and they spoke about this strong compositionality. We have semantics and intentions. You know, it's possible to build these cognitive representations. But we have the issue that we can't really design them to represent the world in a high fidelity way. We have the semantics divergence. And then you're pointing to this this very interesting constructive thing. And I think a constructive form of AI and compositionality solves a lot of problems because of this path dependence problem and this canalization that we're talking about. That when you build intelligence brick by brick, you can build artifacts of incredible sophistication. But unfortunately, we can't design the artifacts to do exactly what we want. We have we can gently steer them in a certain direction. And even with with Tristan, I I feel that even though he's talking about the what of intelligence is is prediction and and adaptivity, I think the implementation matters. I think adaptivity means structure learning. I think there's something about having a substrate which actually does this form of of composition that you're talking about. That seems to be like a mechanistic necessary condition for intelligence. Right. Yes. I I think in many ways what we're talking about is sort of the the tension between analog and digital ways of thinking or bottom up and top down ways of thinking. Yeah. So for instance, let's talk about how you would recognize a bicycle. You know, in the good old fashioned AI world, you would say, you've got a circle detector, you know, and and a, you know, and a line detector that will detect, you know, the the lines that make up the frame of the bike and and so on, and you'll write you'll handwrite code for all of those things. And and of course, the problem is that, you know, there are many ways of looking at a bike where you're not you're not gonna see the the wheels at once or maybe the bike is of a weird design, you know, there are those funny bikes that have shoes instead of wheels, you know. When you look at 1 of those in a gestalt sort of way, you recognize a bike immediately even if all of the rules are broken as it were. That's really important because when when you're when you're looking as an intelligent being at the world, you have to cluster, you have to you have to find regularities in the world that, you know, whose shapes are not well defined by by a set of rules. You know, they're not they're not just sort of carved up by hyperplanes, they're blobby. And and so, you know, intelligence requires methods that are very neural net like, you know, that look more like continuous function approximators, and that that's why gradient descent is a good idea, for instance. You know, learning these things via smooth functions is a good idea, and learning them or or not learning them, but trying to encode them with rules never never worked out well. Now, on the other hand, DNA is is discrete. Right? There there are there are, you know, 4 symbols and you you order them in a certain way and and, you know, that's it. Doesn't mean that there's no randomness in the way, you know, the way proteins are folded and so on, but composition at the level of DNA really does have to do with, you know, chopping up programs essentially made up discrete symbols and inserting, you know, bits of code and and so on. So, you know, when you're looking from the bottom up, it's a very very quantized world. But when you start to, you know, look at at, you know, giant complex things like us, you know, from from a high level, you have to begin from a more more continuous perspective. I think you've hinted that there are natural convergent patterns in in computation. I mean, can can we sort of get a convex hull of your philosophy?…

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