Sundar Pichai: prediction

Transcript context

…What’s a skinny building doing in my neighborhood? It’s a street light, not a building. You see glimpses, that’s why use the word AJI because then you see stuff which obviously we are far from AGI too, so you have both experiences simultaneously happening to you. I’ll answer your question, but I’ll also throw out this. I almost feel the term doesn’t matter, what I know is by 2030 there’ll be such dramatic progress. We’ll be dealing with the consequences of that progress, both the positive externalities and the negative externalities that come with it in a big way by 2030. So that I strongly feel. Whatever, we may be arguing about the term or maybe Gemini can answer what that moment is in time in 2030, but I think the progress will be dramatic. So that I believe in. Will the AI think it has reached AGI by 2030? I would say we will just fall short of that timeline, so I think it’ll take a bit longer. It’s amazing, in the early days of Google DeepMind in 2010, they talked about a 20-year timeframe to achieve AGI, which is kind of fascinating to see, but for me, the whole thing, seeing what Google Brain did in 2012, and when we acquired DeepMind in 2014, right close to where we are sitting, in 2012, Jeff Dean showed the image of when the neural networks could recognize a picture of a cat and identify it. This is the early versions of Brain. And so we all talked about couple decades. I don’t think we’ll quite get there by 2030, so my sense is it’s slightly after that, but I would stress it doesn’t matter what that definition is because you will have mind-blowing progress on many dimensions. Maybe AI can create videos. We have to figure out as a society, we need some system by which we all agree that this is AI-generated and we have to disclose it in a certain way because how do you distinguish reality otherwise? Yeah, there’s so many interesting things you said. So first of all, just looking back at this recent, now feels like distant, history with Google Brain, I mean that was before TensorFlow, before TensorFlow was made public, and open-sourced. So the tooling matters too. Combined with GitHub, ability to share code. Then you have the ideas of a potential transformers and the diffusion now and then there might be a new idea that seems simple in retrospect but will change everything, and that could be the post-training, the inference time innovations. And I think shadcn Tweeted that Google is just one great UI from completely winning the AI race, meaning UI is a huge part of it. How that intelligence, I think the [inaudible 00:42:45] Project likes to talk about this right now, it’s an LLM, but when is it going to become a system where you’re talking about shipping systems versus shipping a particular model? Yeah, that matters too, how the system manifests itself and how it presents itself to the world. That really, really matters…

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