Yann LeCun
Appearances
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's really what, I mean, all the mistakes that humanity makes is because of lack of intelligence, really, or lack of knowledge, which is, you know, related. So making people smarter, we just can only be better. I mean, for the same reason that, you know, public education is a good thing. And books are a good thing. And the internet is also a good thing intrinsically.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And even social networks are a good thing if you run them properly. It's difficult, but, you know, you can. Because, you know, it's... helps the communication of information and knowledge and the transmission of knowledge. So AI is going to make humanity smarter.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And the analogy I've been using is the fact that perhaps an equivalent event in the history of humanity to what might be provided by generalization of AI assistant is the invention of the printing press. It made everybody smarter. The fact that People could have access to books. Books were a lot cheaper than they were before.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so a lot more people had an incentive to learn to read, which wasn't the case before. And people became smarter. It enabled the enlightenment, right? There wouldn't be an enlightenment without the printing press. It enabled... philosophy, rationalism, escape from religious doctrine, democracy, science,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And certainly without this, there wouldn't have been the American Revolution or the French Revolution. And so we'd still be under feudal regimes, perhaps. And so it completely transformed the world because people became smarter and kind of learned about things.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Now, it also created 200 years of essentially religious conflicts in Europe, because the first thing that people read was the Bible and realized that perhaps there was a different interpretation of the Bible than what the priests were telling them. And so that created the Protestant movement and created the Rift.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And in fact, the Catholic Church didn't like the idea of the printing press, but they had no choice. And so it had some bad effects and some good effects. I don't think anyone today would say that the invention of the printing press had an overall negative effect, despite the fact that it created 200 years of religious conflicts in Europe. Now, compare this
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And I thought I was very proud of myself to come up with this analogy, but realized someone else came with the same idea before me. Compare this with what happened in the Ottoman Empire. The Ottoman Empire banned the printing press for 200 years. And he didn't ban it for all languages, only for Arabic.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You could actually print books in Latin or Hebrew or whatever in the Ottoman Empire, just not in Arabic. And I thought it was because the rulers just wanted to preserve the control over the population and the dogma, religious dogma and everything. But after talking with the UAE minister of AI, Omar Al-Olamar, he told me, no, there was another reason.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And the other reason was that it was to preserve the corporation of calligraphers. There's like an art form, which is writing those beautiful Arabic poems or whatever religious text in this thing. And it was a very powerful corporation of scribes, basically, that kind of ran a big chunk of the empire. And we couldn't put them out of business.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So they banned the Beijing press in part to protect that business. Now, what's the analogy for AI today? Who are we protecting by banning AI? Who are the people who are asking that AI be regulated to protect their jobs? And of course, it's a real question of what is going to be the effect of technological transformation like AI on the job market. and the labor market.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And there are economists who are much more expert at this than I am. But when I talk to them, they tell us, you know, we're not going to run out of jobs. This is not going to cause mass unemployment. This is just going to be gradual shift of different professions. The professions are going to be hot 10 or 15 years from now. We have no idea today what they're going to be.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So yeah, that's what a lot of people are working on. So the short answer is no. And the more complex answer is you can use all kinds of tricks to get an LLM to basically digest visual representations of images Or video, or audio for that matter. And a classical way of doing this is you train a vision system in some way.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The same way if we go back 20 years in the past, like who could have thought 20 years ago that like the hottest job, even like five, 10 years ago was mobile app developer. Like smartphones weren't invented.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, it could be. Yeah.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So I share that feeling, okay? I think people are fundamentally good. And in fact, a lot of doomers are doomers because they don't think that people are fundamentally good. And they either don't trust people or they don't trust the institution to do the right thing so that people behave properly.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And we have a number of ways to train vision systems, either supervised, semi-supervised, self-supervised, all kinds of different ways. That will turn any image into a high-level representation. basically a list of tokens that are really similar to the kind of tokens that typical LLM takes as an input. And then you just feed that to the LLM in addition to the text.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And you just expect the LLM to kind of, during training, to kind of be able to use those representations to help make decisions. I mean, there's been work along those lines for quite a long time. And now you see those systems, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, there are LLMs that have some vision extension, but they're basically hacks in the sense that those things are not like trained end-to-end to handle, to really understand the world. They're not trained with video, for example. They don't really understand intuitive physics, at least not at the moment.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We're not going to be able to do this with the type of LLMs that we are working with today. And there's a number of reasons for this. But the main reason is The way LLMs are trained is that you take a piece of text, you remove some of the words in that text, you mask them, you replace them by blank markers, and you train a genetic neural net to predict the words that are missing.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And if you build this neural net in a particular way so that it can only look at words that are to the left of the one it's trying to predict, then what you have is a system that basically is trying to predict the next word in a text, right? So then you can feed it a text, a prompt, and you can ask it to predict the next word. It can never predict the next word exactly.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so what it's going to do is produce a probability distribution over all the possible words in your dictionary. In fact, it doesn't predict words, it predicts tokens that are kind of subword units. And so it's easy to handle the uncertainty in the prediction there, because there is only a finite number of possible words in the dictionary, and you can just compute a distribution over them.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Then what the system does is that it picks a word from that distribution. Of course, there's a higher chance of picking words that have a higher probability within that distribution. So you sample from the distribution to actually produce a word. And then you shift that word into the input. And so that allows the system not to predict the second word, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And once you do this, you shift it into the input, et cetera. That's called autoregressive prediction, which is why those LLMs should be called autoregressive LLMs. But we just call them LLMs. And there is a difference between this kind of process and a process by which before producing a word, when you talk, when you and I talk, You and I are bilingual.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We think about what we're going to say, and it's relatively independent of the language in which we're going to say it. When we talk about, I don't know, let's say a mathematical concept or something, the kind of thinking that we're doing and the answer that we're planning to produce is is not linked to whether we're going to see it in French or Russian or English.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right. It's certainly true for a lot of thinking that we do.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, pretty much.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, it depends what kind of thinking, right? If it's just, if it's like producing puns, I get much better in French than English about that.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
There is an abstract representation of imagining the reaction of a reader to that text.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Or figure out a reaction you want to cause and then figure out how to say it so that it causes that reaction. But that's really close to language. But think about a mathematical concept or imagining something you want to build out of wood or something like this, right? the kind of thinking you're doing is absolutely nothing to do with language, really.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's not like you have necessarily an internal monologue in any particular language. You're imagining mental models of the thing, right? If I ask you to imagine what this water bottle will look like if I rotate it 90 degrees, that has nothing to do with language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so clearly there is a more abstract level of representation in which we do most of our thinking and we plan what we're gonna say if the output is you know, uttered words as opposed to an output being, you know, muscle actions, right? We plan our answer before we produce it. And LLMs don't do that. They just produce one word after the other instinctively, if you want.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's like, it's a bit like the, you know, subconscious actions where you don't, Like you're distracted, you're doing something, you're completely concentrated, and someone comes to you and asks you a question, and you kind of answer the question. You don't have time to think about the answer, but the answer is easy, so you don't need to pay attention. You sort of respond automatically.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's kind of what an LLM does, right? It doesn't think about its answer, really. It retrieves it because it's accumulated a lot of knowledge, so it can retrieve some things, but it's going to just spit out one token after the other without planning the answer.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, but then that assumes that those systems actually possess an eternal world model.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah. So can you build this, first of all, by prediction? Right. And the answer is probably yes. Can you build it by predicting words? And the answer is most probably no, because language is very poor in terms of weak or low bandwidth, if you want. There's just not enough information there.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So building world models means observing the world and understanding why the world is evolving the way it is. And then the extra component of a world model is something that can predict how the world is going to evolve as a consequence of an action you might take, right? So one model really is, here is my idea of the state of the world at time t. Here is an action I might take.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
What is the predicted state of the world at time t plus 1? Now that state of the world does not need to represent everything about the world. It just needs to represent enough that's relevant for this planning of the action, but not necessarily all the details. Now here is the problem. You're not going to be able to do this with generative models.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So a generative model that's trained on video, and we've tried to do this for 10 years. You take a video, show a system a piece of video, and then ask it to predict the reminder of the video. Basically, predict what's going to happen.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Either one frame at a time or a group of frames at a time. But yeah, a large video model, if you want. The idea of doing this has been floating around for a long time, and at FAIR, some of my colleagues and I have been trying to do this for about 10 years.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And you can't really do the same trick as with LLMs, because LLMs, as I said, you can't predict exactly which word is going to follow a sequence of words, but you can predict the distribution of words. Now, if you go to video, what you would have to do is predict the distribution over all possible frames in a video. And we don't really know how to do that properly.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We do not know how to represent distributions over high-dimensional continuous spaces in ways that are useful. And there lies the main issue. And the reason we can do this is because the world is incredibly more complicated and richer in terms of information than text. Text is discrete. Video is highly dimensional and continuous. A lot of details in this.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So if I take a video of this room, and the video is a camera panning around, there is no way I can predict everything that's going to be in the room as I pan around. The system cannot predict what's going to be in the room as the camera is panning. Maybe it's going to predict this is a room where there's a light and there is a wall and things like that.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It can't predict what the painting on the wall looks like or what the texture of the couch looks like. Certainly not the texture of the carpet. So there's no way it can predict all those details. So the way to handle this
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
is one way possibly to handle this, which we've been working for a long time, is to have a model that has what's called a latent variable, and the latent variable is fed to a neural net, and it's supposed to represent all the information about the world that you don't perceive yet, and that you need to augment the system for the prediction to do a good job at predicting pixels, including the fine texture of the
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
the carpet and the couch, and the painting on the wall. That has been a complete failure, essentially. And we've tried lots of things. We tried just straight neural nets, we tried GANs, we tried VAEs, all kinds of regularized autoencoders, we tried many things.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We also tried those kind of methods to learn good representations of images or video that could then be used as input to, for example, an image classification system. And that also has basically failed. All the systems that attempt to predict missing parts of an image or video form a corrupted version of it, basically. So I take an image or a video, corrupt it or transform it in some way,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then try to reconstruct the complete video or image from the corrupted version. And then hope that internally the system will develop good representations of images that you can use for object recognition, segmentation, whatever it is. That has been essentially a complete failure. And it works really well for text. That's the principle that is used for LLMs, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, so the reason this doesn't work is, first of all, I have to tell you exactly what doesn't work because there is something else that does work. So the thing that does not work is training the system to learn representations of images by training it to reconstruct a good image from a corrupted version of it. That's what doesn't work.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And we have a whole slew of techniques for this that are a variant of denoising autoencoders. Something called MAE, developed by some of my colleagues at FAIR, masked autoencoder. So it's basically like the you know, LLMs or things like this, where you train the system by corrupting text, except you corrupt images, you remove patches from it, and you train a gigantic neural net to reconstruct.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The features you get are not good. And you know they're not good because if you now train the same architecture, but you train it supervised, with label data, with textual descriptions of images, et cetera, you do get good representations. And the performance on recognition tasks is much better than if you do this self-supervised pre-training. So the architecture is good. The architecture is good.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The architecture of the encoder is good. But the fact that you train the system to reconstruct images does not lead it to produce, to learn good generic features of images.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Self-supervised by reconstruction. Yeah, by reconstruction. Okay, so what's the alternative? The alternative is joint embedding. What is joint embedding?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, so now instead of training a system to encode the image and then training it to reconstruct the full image from a corrupted version, you take the full image, you take the corrupted or transformed version. You run them both through encoders, which in general are identical, but not necessarily.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then you train a predictor on top of those encoders to predict the representation of the full input from the representation of the corrupted one. So joint embedding, because you're taking the full input and the corrupted version, or transformed version, run them both through encoders, so you get a joint embedding.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then you're saying, can I predict the representation of the full one from the representation of the corrupted one? And I call this a JEPA, so that means joint embedding predictive architecture, because there's joint embedding and there is this predictor that predicts the representation of the good guy from the bad guy. And the big question is, how do you train something like this?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And until five years ago, six years ago, we didn't have particularly good answers for how you train those things, except for one called contrastive learning.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And the idea of contractive learning is you take a pair of images that are, again, an image and a corrupted version or degraded version somehow, or transformed version of the original one, and you train the predicted representation to be the same as that. If you only do this, the system collapses. It basically completely ignores the input and produces representations that are constant.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So the contrastive methods avoid this. And those things have been around since the early 90s. I had a paper on this in 1993. You also show pairs of images that you know are different. And then you push away the representations from each other.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So you say, not only do representations of things that we know are the same, should be the same or should be similar, but representation of things that we know are different should be different. And that prevents the collapse, but it has some limitation. And there's a whole bunch of techniques that have appeared over the last six, seven years that can revive this type of method.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Some of them from FAIR, some of them from Google and other places. But there are limitations to those contrasting methods. What has changed in the last... you know, three, four years, is now we have methods that are non-contrastive. So they don't require those negative contrastive samples of images that we know are different.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You turn them on you with images that are, you know, different versions or different views of the same thing. And you rely on some other tweaks to prevent the system from collapsing. And we have half a dozen different methods for this now.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, so it's a first step. So first of all, what's the difference with generative architectures like LLMs? So LLMs or vision systems that are trained by reconstruction generate the inputs. They generate the original input that is non-corrupted, non-transformed. So you have to predict all the pixels.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And there is a huge amount of resources spent in the system to actually predict all those pixels, all the details. In a JEPA, you're not trying to predict all the pixels. You're only trying to predict an abstract representation of the inputs, right? And that's much easier in many ways.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So what the JEPA system, when it's being trained, is trying to do is extract as much information as possible from the input, but yet only extract information that is relatively easily predictable. Okay. So there's a lot of things in the world that we cannot predict.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
For example, if you have a self-driving car driving down the street or road, there may be trees around the road and it could be a windy day. So the leaves on the tree are kind of moving in kind of semi-chaotic random ways that you can't predict and you don't care. You don't want to predict. So what you want is your encoder to basically eliminate all those details.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It will tell you there's moving leaves, but it's not going to keep the details of exactly what's going on. And so when you do the prediction in representation space, you're not going to have to predict every single pixel of every leaf.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And that not only is a lot simpler, but also it allows the system to essentially learn an abstract representation of the world where what can be modeled and predicted is preserved, and the rest is viewed as noise and eliminated by the encoder. So it kind of lifts the level of abstraction of the representation. If you think about this, this is something we do absolutely all the time.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Whenever we describe a phenomenon, we describe it at a particular level of abstraction. We don't always describe every natural phenomenon in terms of quantum field theory. That would be impossible.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We have multiple levels of abstraction to describe what happens in the world, starting from quantum field theory to atomic theory and molecules and chemistry, materials, and all the way up to concrete objects in the real world and things like that. We can't just only model everything at the lowest level. That's what the idea of JEPA is really about.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Learn abstract representation in a self-supervised manner. You can do it hierarchically as well. That, I think, is an essential component of an intelligent system. In language, we can get away without doing this because language is already, to some level, abstract, and already has eliminated a lot of information that is not predictable.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So we can get away without doing the chanter embedding, without lifting the abstraction level, and by directly predicting words.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right. And the thing is, those self-supervised algorithms that learn by prediction, even in representation space, they learn more concepts if the input data you feed them is more redundant. The more redundancy there is in the data, the more they're able to capture some internal structure of it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so there, there is way more redundancy and structure in perceptual inputs, sensory input, like vision, than there is in text, which is not nearly as redundant. This is back to the question you were asking. a few minutes ago. Language might represent more information really because it's already compressed. You're right about that, but that means it's also less redundant.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so self-supervised learning will not work as well.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, eventually, yes. But I think if we do this too early, we run the risk of being tempted to cheat. And in fact, that's what people are doing at the moment with vision language model. We're basically cheating. We're using language as a crutch to help the deficiencies of our vision systems to kind of learn good representations from images and video. And the problem with this is that
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We might improve our vision language system a bit, I mean, our language models by feeding them images, but we're not going to get to the level of even the intelligence or level of understanding of the world of a cat or a dog, which doesn't have language. They don't have language, and they understand the world much better than any LLM.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
They can plan really complex actions and sort of imagine the result of a bunch of actions, How do we get machines to learn that before we combine that with language? Obviously, if we combine this with language, this is going to be a winner. But before that, we have to focus on how do we get systems to learn how the world works.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's the hope. In fact, the techniques we're using are non-contrastive. So not only is the architecture non-generative, the learning procedures we're using are non-contrastive. We have two sets of techniques. One set is based on distillation, and there's a number of methods that use this principle. One by DeepMind called BYOL, a couple by FAIR, one called VicReg, and another one called IGEPA.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And vcrag, I should say, is not a distillation method, actually, but ijpa and BYOL certainly are. And there's another one also called dino, also produced at FAIR. And the idea of those things is that you take the full input, let's say an image, you run it through an encoder, produces a representation.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then you corrupt that input or transform it, running through essentially what amounts to the same encoder, with some minor differences. And then train a predictor, sometimes the predictor is very simple, sometimes it doesn't exist, but train a predictor to predict a representation of the first uncorrupted input from the corrupted input. But you only train the second branch.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You only train the part of the network that is fed with the corrupted input. The other network you don't train, but since they share the same weight, when you modify the first one, it also modifies the second one. And with various tricks, you can prevent the system from collapsing, with the collapse of the type I was explaining before, where the system basically ignores the input.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So that works very well. The two techniques we've developed at FAIR, Deno and IGEPA, work really well for that.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So there's several scenarios. One scenario is you take an image, you corrupt it by changing the cropping, for example, changing the size a little bit, maybe changing the orientation, blurring it, changing the colors. doing all kinds of horrible things to it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Basic horrible things that sort of degrade the quality a little bit and change the framing, you know, crop the image. And in some cases, in the case of iJet, you don't need to do any of this. You just mask some parts of it, right? You just basically remove some regions, like a big block, essentially. Yeah.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then run through the encoders and train the entire system, encoder and predictor, to predict the representation of the good one from the representation of the corrupted one. So that's the IGEPA. It doesn't need to know that it's an image, for example, because the only thing it needs to know is how to do this masking.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Whereas with Deno, you need to know it's an image because you need to do things like geometry transformation and blurring and things like that that are really image-specific. A more recent version of this that we have is called VJPA. So it's basically the same idea as iJPA, except it's applied to video. So now you take a whole video and you mask a whole chunk of it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And what we mask is actually kind of a temporal tube. So like a whole segment of each frame in the video over the entire video. Mm-hmm.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Throughout the tube, yeah. Typically it's 16 frames or something, and we masked the same region over the entire 16 frames. It's a different one for every video, obviously. And then again, train that system so as to predict the representation of the full video from the partially masked video. That works really well.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's the first system that we have that learns good representations of video so that when you feed those representations to a supervised classifier head, it can tell you what action is taking place in the video with pretty good accuracy. So that's the first time we get something of that quality.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah. We have also preliminary results that seem to indicate that the representation allows our system to tell whether the video is physically possible or completely impossible because some object disappeared or an object suddenly jumped from one location to another or changed shape or something.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Possibly. This is going to take a while before we get to that point, but there are robotic systems that are based on this idea. And what you need for this is a slightly modified version of this, where imagine that you have a complete video.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And what you're doing to this video is that you're either translating it in time towards the future, so you only see the beginning of the video, but you don't see the latter part of it that is in the original one. Or you just mask the second half of the video, for example. And then you train a JEPA system of the type I described to predict the representation of the full video from the shifted one.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But you also feed the predictor with an action. For example, the wheel is turned 10 degrees to the right or something. So if it's a dash cam in a car and you know the angle of the wheel, you should be able to predict to some extent what's going to happen to what you see.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You're not going to be able to predict all the details of objects that appear in the view, obviously, but at an abstract representation level, you can probably predict what's going to happen. So now what you have is...
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
an internal model that says, here is my idea of state of the world at time t, here is an action I'm taking, here is a prediction of the state of the world at time t plus one, t plus delta t, t plus two seconds, whatever it is. If you have a model of this type, you can use it for planning.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So now you can do what LLMs cannot do, which is planning what you're going to do so as to arrive at a particular outcome or satisfy a particular objective. So you can have a number of objectives. I can predict that if I have an object like this and I open my hand, it's going to fall. And if I push it with a particular force on the table, it's going to move.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
If I push the table itself, it's probably not going to move with the same force. So we have this internal model of the world in our mind, which allows us to plan sequences of actions to arrive at a particular goal. And so now if you have this world model, we can imagine a sequence of actions, predict what the outcome of the sequence of action is going to be,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
measure to what extent the final state satisfies a particular objective, like moving the bottle to the left of the table, and then plan a sequence of actions that will minimize this objective at runtime. We're not talking about learning, we're talking about inference time. So this is planning, really. And in optimal control, this is a very classical thing. It's called model predictive control.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You have a model of the system you want to control that can predict the sequence of states corresponding to a sequence of commands. And you're planning a sequence of commands so that, according to your world model, the end state of the system will satisfy an objective that you fix. This is the way...
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Rocket trajectories have been planned since computers have been around, so since the early 60s, essentially.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, so no, you will have to build a specific architecture to allow for hierarchical planning. So hierarchical planning is absolutely necessary if you want to plan complex actions. If I want to go from, let's say, from New York to Paris, it's the example I use all the time, and I'm sitting in my office at NYU, my objective that I need to minimize is my distance to Paris at a high level.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
a very abstract representation of my location, I would have to decompose this into two sub-goals. First one is go to the airport. Second one is catch a plane to Paris. Okay, so my sub-goal is now going to the airport. My objective function is my distance to the airport. How do I go to the airport? Well, I have to go in the street and hail a taxi, which you can do in New York.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, now I have another sub-goal. Go down on the street. What that means, going to the elevator, going down the elevator, walk out the street. How do I go to the elevator? I have to... Stand up from my chair, open the door of my office, go to the elevator, push the button. How do I get up from my chair?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Like, you know, you can imagine going down, all the way down to basically what amounts to millisecond by millisecond muscle control. Okay. And obviously you're not going to plan your entire trip from New York to Paris in terms of millisecond by millisecond muscle control. First, that would be incredibly expensive.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But it will also be completely impossible because you don't know all the conditions of what's going to happen. How long it's going to take to catch a taxi or to go to the airport with traffic. You would have to know exactly the condition of everything to be able to do this planning. And you don't have the information.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So you have to do this hierarchical planning so that you can start acting and then sort of replanning as you go. And nobody really knows how to do this in AI. Nobody knows how to train a system to learn the appropriate multiple levels of representation so that hierarchical planning works.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right, so there's a lot of questions that are actually implied by this, right? So the first thing is, LLMs will be able to answer some of those questions down to some level of abstraction. under the condition that they've been trained with similar scenarios in their training set.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, true. I mean, they will probably produce some answer, except they're not going to be able to really kind of produce millisecond by millisecond muscle control of how you stand up from your chair, right? But down to some level of abstraction where you can describe things by words,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
They might be able to give you a plan, but only under the condition that they've been trained to produce those kind of plans, right? They're not going to be able to plan for situations that they never encountered before. They basically are going to have to regurgitate the template that they've been trained on.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Certainly, LLM would be able to solve that problem if you fine-tune it for it. I can't say that LLM cannot do this. It can do this if you train it for it. There's no question. Down to a certain level, where things can be formulated in terms of words. But if you want to go down to how you climb down the stairs or just stand up from your chair in terms of words, you can't do it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's one of the reasons you need experience of the physical world, which is much higher bandwidth than what you can express in words. In human language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Sure. And, you know, a lot of plans that people know about that are relatively high level are actually learned. Most people don't invent the, you know, plans. They... We have some ability to do this, of course, obviously, but most plans that people use are plans that they've been trained on. They've seen other people use those plans or they've been told how to do things.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You can't invent how you take a person who's never heard of airplanes and tell them, like, how do you go from New York to Paris and... They're probably not going to be able to deconstruct the whole plan unless they've seen examples of that before. So certainly LLMs are going to be able to do this.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But then how you link this from the low level of actions, that needs to be done with things like JEPA that basically lifts the abstraction level of the representation without attempting to reconstruct every detail of the situation. That's why we need JEPAs for it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, there's one thing that autoregressive LLMs, or that LLMs in general, not just the autoregressive one, but including the BERT-style bidirectional ones, are exploiting, and it's self-supervised learning. And I've been a very, very strong advocate of self-supervised learning for many years. So those things are an incredibly impressive demonstration that self-supervised learning actually works.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It didn't start with Bert, but it was really a good demonstration with this. The idea that you take a piece of text, you corrupt it, and then you train some gigantic neural net to reconstruct the parts that are missing, that has produced an enormous amount of benefits. It allowed us to create systems that understand language, systems that can translate hundreds of languages in any direction.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
systems that are multilingual, so it's a single system that can be trained to understand hundreds of languages and translate in any direction, and produce summaries, and then answer questions and produce text.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then there's a special case of it, which is the autoregressive trick, where you constrain the system to not elaborate a representation of the text from looking at the entire text, but only predicting a word from the words that come before. And you do this by constraining the architecture of the network. And that's what you can build an autoregressive LLM from.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So there was a surprise many years ago with what's called decoder-only LLMs. So, you know, systems of this type that are just trying to produce words from the previous one. And the fact that when you scale them up, they tend to really kind of understand more about language. When you train them on lots of data, you make them really big. That was kind of a surprise.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And that surprise occurred quite a while back, like, you know, with work from Google Meta, OpenAI, et cetera, going back to the GPT kind of work, general pre-trained transformers.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, I mean, there were work from various places, but if you want to kind of place it in the GPT timeline, that would be around GPT-2, yeah.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, we're fooled by their fluency, right? We just assume that if a system is fluent in manipulating language, then it has all the characteristics of human intelligence. But that impression is false. We're really fooled by it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Alan Turing would decide that the Turing test is a really bad test. Okay. This is what the AI community has decided many years ago, that the Turing test was a really bad test of intelligence.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Hans Moravec would say the Moravec paradox still applies. Okay. Okay. Okay, we can pass.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, of course, everybody would be impressed. But, you know, it's not a question of being impressed or not. It's the question of knowing what the limit of those systems can do. Again, they are impressive. They can do a lot of useful things. There's a whole industry that is being built around them. They're going to make progress.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But there is a lot of things they cannot do and we have to realize what they cannot do and then figure out how we get there. I'm seeing this from basically 10 years of research on the idea of self-supervised learning.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Actually, that's going back more than 10 years, but the idea of self-supervised learning, so basically capturing the internal structure of a set of inputs without training the system for any particular task, learning representations. You know, the conference I co-founded 14 years ago is called International Conference on Learning Representations.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's the entire issue that deep learning is dealing with, right? And it's been my obsession for almost 40 years now. So learning representation is really the thing. For the longest time, you could only do this with supervised learning.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then we started working on what we used to call unsupervised learning and sort of revived the idea of unsupervised learning in the early 2000s with Yoshua Bengio and Jeff Hinton. Then discovered that supervised learning actually works pretty well if you can collect enough data. And so the whole idea of unsupervised self-supervised learning kind of took a backseat for a bit.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then I kind of tried to revive it in a big way, starting in 2014, basically when we started FAIR. and really pushing for finding new methods to do self-supervised learning, both for text and for images and for video and audio. And some of that work has been incredibly successful.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, the reason why we have a multilingual translation system, you know, things to do content moderation on Meta, for example, on Facebook, that are multilingual, that understand whether a piece of text is hate speech or not or something. is due to that progress using self-supervised learning for NLP, combining this with transformer architectures and blah, blah, blah.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But that's the big success of self-supervised learning. We had similar success in speech recognition, a system called Wave2Vec, which is also a joint embedding architecture, by the way, trained with contrastive learning. And that system also can produce speech recognition systems that are multilingual,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
with mostly unlabeled data and only need a few minutes of labeled data to actually do speech recognition. That's amazing. We have systems now based on those combination of ideas that can do real-time translation of hundreds of languages into each other.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's right. We don't go through text. It goes directly from speech-to-speech using an internal representation of kind of speech units that are discrete. But it's called textless NLP. We used to call it this way. But yeah, so that, I mean, incredible success there. And then, you know, for 10 years, we tried to apply this idea
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
to learning representations of images by training a system to predict videos, learning intuitive physics by training a system to predict what's going to happen in a video, and tried and tried and failed and failed with generative models, with models that predict pixels. We could not get them to learn good representations of images. We could not get them to learn good representations of videos.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And we tried many times. We published lots of papers on it. They kind of sort of worked, but not really great. It started working. We abandoned this idea of predicting every pixel and basically just doing digital embedding and predicting in representation space. That works.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So there's ample evidence that we're not going to be able to learn good representations of the real world using generative model. So I'm telling people, everybody's talking about generative AI. If you're really interested in human-level AI, abandon the idea of generative AI.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right. Well, there's a lot of situations that might be difficult for a purely language-based system to know. Like, okay, you can probably learn from reading texts, the entirety of the publicly available texts in the world, that I cannot get from New York to Paris by snapping my fingers. That's not going to work, right? Yes.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But there's probably more complex scenarios of this type, which an NLM may never have encountered and may not be able to determine whether it's possible or not. So that link from the low level to the high level. The thing is that the high level that language expresses is based on a common experience of the low level, which LLMs currently do not have.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
When we talk to each other, we know we have a common experience of the world. A lot of it is similar.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, I agree with what you just said, which is that to be able to do high-level common sense, to have high-level common sense, you need to have the low-level common sense to build on top of.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And that's not there in LLMs. LLMs are purely trained from text. So then the other statement you made, I would not agree with the fact that implicit in all languages in the world is the underlying reality. There's a lot about underlying reality which is not expressed in language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You don't need to communicate the stuff that is common.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I just think that most of the information of this type that we have accumulated when we were babies is just not present in text, in any description, essentially.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, that's the 16,000 hours of wake time of a four-year-old and 10 to the 15 bytes, you know, going through vision, just vision, right? There is a similar bandwidth, you know, of touch and a little less through audio. And then text doesn't, language doesn't come in until like, you know, a year in life. And by the time you are nine years old, you've learned everything.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
about gravity, you know, about inertia, you know, about gravity, you know, the stability, you know, you know, about the distinction between animate and inanimate objects, you know, by 18 months, you know, about like why people want to do things and you help them if they can't, you know, I mean, there's a lot of things that you learn mostly by observation, really not even through interaction.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
In the first few months of life, babies don't really have any influence on the world. They can only observe, right? And you accumulate a gigantic amount of knowledge just from that. So that's what we're missing from current AI systems.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right. So because of the autoregressive prediction, every time an LLM produces a token or a word, there is some level of probability for that word to take you out of the set of reasonable answers. And if you assume, which is a very strong assumption, that the probability of such error is that those errors are independent across a sequence of tokens being produced.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
What that means is that every time you produce a token, the probability that you stay within the set of correct answers decreases, and it decreases exponentially.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah. And that drift is exponential. It's like errors accumulate, right? So the probability that an answer would be nonsensical increases exponentially with the number of tokens.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, it's basically a struggle against the curse of dimensionality. So the way you can correct for this is that you fine-tune the system by having it produce answers for all kinds of questions that people might come up with. And people are people, so a lot of the questions that they have are very similar to each other, so you can probably cover 80% or whatever.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
of questions that people will ask by collecting data. And then you fine-tune the system to produce good answers for all of those things. And it's probably going to be able to learn that because it's got a lot of capacity to learn. But then there is... you know, the enormous set of prompts that you have not covered during training. And that set is enormous.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Like within the set of all possible prompts, the proportion of prompts that have been used for training is absolutely tiny. It's a tiny, tiny, tiny subset of all possible prompts. And so the system will behave properly on the prompts that has been either trained, pre-trained or fine-tuned.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But then there is an entire space of things that it cannot possibly have been trained on because the number is gigantic. So whatever training the system has been subject to to produce appropriate answers, you can break it by finding out a prompt that will be outside of the set of prompts it's been trained on, or things that are similar, and then it will just spew complete nonsense.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, people have come up with things where you put essentially a random sequence of characters in a prompt, and that's enough to kind of throw the system into a mode where it's going to answer something completely different than it would have answered without this. So that's a way to jailbreak the system, basically go outside of its conditioning, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, some people have done things like you write a sentence in English or you ask a question in English and it produces a perfectly fine answer. And then you just substitute a few words. by the same word in another language. And all of a sudden, the answer is complete nonsense.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So the problem is that there is a long tail. Yes. This is an issue that a lot of people have realized in social networks and stuff like that, which is there's a very, very long tail of things that people will ask. And you can fine-tune the system for the 80% or whatever of the things that most people will ask.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And then this long tail is so large that you're not going to be able to fine-tune the system for all the conditions. And in the end, the system ends up being kind of a giant lookup table, right, essentially, which is not really what you want. You want systems that can reason, certainly that can plan. So the type of reasoning that takes place in LLM is very, very primitive.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And the reason you can tell it's primitive is because the amount of computation that is spent per token produced is constant. So if you ask a question and that question has an answer in a given number of token, the amount of computation devoted to computing that answer can be exactly estimated.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's like, you know, it's the size of the prediction network, you know, with its 36 layers or 92 layers or whatever it is, multiplied by number of tokens, that's it. And so essentially it doesn't matter if the question being asked
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
is simple to answer, complicated to answer, impossible to answer because it's undecidable or something, the amount of computation the system will be able to devote to the answer is constant, or is proportional to the number of tokens produced in the answer, right? This is not the way we work.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The way we reason is that when we're faced with a complex problem or a complex question, we spend more time trying to solve it and answer it, right? Because it's more difficult.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, whether it's difficult or not, the near future will say, because a lot of people are working on reasoning and planning abilities for dialogue systems. I mean, even if we restrict ourselves to language, just having the ability to plan your answer before you answer in terms that are not necessarily linked with the language you're going to use to produce the answer.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So this idea of this mental model that allows you to plan what you're going to say before you say it. That is very important. I think there's going to be a lot of systems over the next few years that are going to have this capability. But the blueprint of those systems would be extremely different from autoregressive LLMs. So
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
it's the same difference as the difference between what psychologists call system one and system two in humans, right? So system one is the type of tasks that you can accomplish without like deliberately, consciously think about how you do them. You just do them, you've done them enough that you can just do it subconsciously, right? Without thinking about them. If you're an experienced driver,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You can drive without really thinking about it, and you can talk to someone at the same time or listen to the radio, right? If you are a very experienced chess player, you can play against a non-experienced chess player without really thinking either. You just recognize the pattern and you play. Right? That's system one.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So all the things that you do instinctively without really having to deliberately plan and think about it. And then there is all the tasks where you need to plan. So if you are... and not to an experienced chess player, or you are experienced when you play against another experienced chess player. You think about all kinds of options, right? You think about it for a while, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You're much better if you have time to think about it than you are if you play Blitz with limited time. So this type of deliberate planning, which uses your internal world model, that's system two. This is what LLMs currently cannot do. So how do we get them to do this? How do we build a system that can do this kind of planning or reasoning that devotes
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
more resources to complex problems than to simple problems. And it's not going to be autoregressive prediction of tokens. It's going to be more something akin to inference of latent variables in what used to be called probabilistic models or graphical models and things of that type. So basically, the principle is like this. The prompt is like observed variables.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And what the model does is that it can measure to what extent an answer is a good answer for a prompt. So think of it as some gigantic neural net, but it's got only one output. And that output is a scalar number, which is, let's say, zero if the answer is a good answer for the question, and a large number if the answer is not a good answer for the question. Imagine you had this model.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
If you had such a model, you could use it to produce good answers. The way you would do is produce the prompt and then search through the space of possible answers for one that minimizes that number. That's called an energy-based model.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, so really what you need to do would be to not search over possible strings of text that minimize that energy. But what you would do is do this in abstract representation space. So in sort of the space of abstract thoughts, you would elaborate a thought, right, using this process of minimizing the output of your model, okay, which is just a scalar. It's an optimization process.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So now the way the system produces its answer is through optimization, by minimizing an objective function, basically. And we're talking about inference, we're not talking about training. The system has been trained already. So now we have an abstract representation of the thought of the answer, representation of the answer. We feed that to, basically, an autoregressive decoder,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
which can be very simple, that turns this into a text that expresses this thought. So that, in my opinion, is the blueprint of future dialogue systems. They will think about their answer, plan their answer by optimization before turning it into text. And that is Turing-complete.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The space of representations. It goes abstract representation. Abstract representation. So you have an abstract representation inside the system. You have a prompt. The prompt goes through an encoder, produces a representation, perhaps goes through a predictor that predicts a representation of the answer, of the proper answer.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But that representation may not be a good answer because there might be some complicated reasoning you need to do, right? So then you have another process that takes the representation of the answers and modifies it so as to minimize a cost function that measures to what extent the answer is a good answer for the question.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Now, we sort of ignore the fact for, I mean, the issue for a moment of how you train that system to measure whether an answer is a good answer for a question.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's an optimization process. You can do this if the entire system is differentiable, that scalar output is the result of running through some neural net, running the answer, the representation of the answer through some neural net. Then by gradient descent, by back-propagating gradients, you can figure out how to modify the representation of the answer so as to minimize that.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So that's still a gradient-based... It's gradient-based inference. So now you have a representation of the answer in abstract space. Now you can turn it into text. And the cool thing about this is that the representation now can be optimized through gradient descent, but also is independent of the language in which you're going to express the answer.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
sensory information right okay but this can can this do something like reasoning which is what we're talking about well not really in a only in a very simple way i mean basically you can think of those things that's doing the kind of optimization i was i was talking about except they optimize in the discrete space which is the space of possible sequences of of tokens
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And they do this optimization in a horribly inefficient way, which is generate a lot of hypotheses and then select the best ones. And that's incredibly wasteful in terms of computation. Because you basically have to run your LLM for every possible generated sequence. And it's incredibly wasteful.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So it's much better to do an optimization in continuous space where you can do gradient descent as opposed to like generate tons of things and then select the best. You just iteratively refine your answer to go towards the best, right? That's much more efficient. But you can only do this in continuous spaces with differentiable functions.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right. So then we're asking the question of, conceptually, how do you train an energy-based model? So an energy-based model is a function with a scalar output, just a number. You give it two inputs, x and y, and it tells you whether y is compatible with x or not. x you observe. Let's say it's a prompt, an image, a video, whatever.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And y is a proposal for an answer, a continuation of the video, you know, whatever. And it tells you whether y is compatible with x. And the way it tells you that y is compatible with x is that the output of that function would be zero if y is compatible with x. It would be a positive number, non-zero, if y is not compatible with x.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, how do you train a system like this at a completely general level is you show it pairs of X and Ys that are compatible, a question and the corresponding answer, and you train the parameters of the big neural net inside to produce zero.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, now that doesn't completely work because the system might decide, well, I'm just going to say zero for everything. So now you have to have a process to make sure that for a wrong y, the energy would be larger than zero. And there you have two options. One is contrastive method.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So contrastive method is you show an x and a bad y, and you tell the system, well, that's, you know, give a high energy to this, like push up the energy, right? Change the weights in the neural net that computes the energy so that it goes up. So that's contrasting methods.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The problem with this is if the space of Y is large, the number of such contrasting samples you're going to have to show is gigantic. But people do this. They do this when you train a system with RLHF. Basically what you're training is what's called a reward model, which is basically an objective function that tells you whether an answer is good or bad. And that's basically exactly what
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
what this is. So we already do this to some extent. We're just not using it for inference, we're just using it for training. There is another set of methods which are non-contrastive, and I prefer those. And those non-contrastive methods basically say, okay, the energy function needs to have low energy on pairs of x, y's that are compatible, that come from your training set.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
How do you make sure that the energy is going to be higher everywhere else? And the way you do this is by having a regularizer, a criterion, a term in your cost function that basically minimizes the volume of space that can take low energy. And the precise way to do this is all kinds of different specific ways to do this, depending on the architecture. But that's the basic principle.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So that if you push down the energy function for particular regions in the XY space, it will automatically go up in other places because there's only a limited volume of space that can take low energy by the construction of the system or by the regularizing function.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, so you can do this with language directly by just, you know, x is a text and y is a continuation of that text. Yes. Or x is a question, y is the answer.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, no, it depends on how the internal structure of the system is built. If the internal structure of the system is built in such a way that inside of the system there is a latent variable, let's call it z, that... you can manipulate so as to minimize the output energy. Then that Z can be viewed as a representation of a good answer that you can translate into a Y that is a good answer.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Very similar way, but you have to have this way of preventing collapse, of ensuring that there is high energy for things you don't train it on. And currently it's very implicit in LLM. It's done in a way that people don't realize is being done, but it is being done. It's due to the fact that when you give a high probability to a word,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
automatically you give low probability to other words because you only have a finite amount of probability to go around right there, to sum to one.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So when you minimize the cross entropy or whatever, when you train your LLM to predict the next word, you're increasing the probability your system will give to the correct word, but you're also decreasing the probability it will give to the incorrect words. Now,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Indirectly, that gives a high probability to sequences of words that are good and low probability to sequences of words that are bad, but it's very indirect. It's not obvious why this actually works at all, because you're not doing it on a joint probability of all the symbols in a sequence.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You're just doing it to factorize that probability in terms of conditional probabilities over successive tokens.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So we've been doing this with Ojepa architectures, basically. The joint embedding. Ojepa. So there, the compatibility between two things is, here's an image or a video, here's a corrupted, shifted, or transformed version of that image or video, or masked. And then the energy of the system is the prediction error of the representation
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
the predicted representation of the good thing versus the actual representation of the good thing. So you run the corrupted image to the system, predict the representation of the good input, uncorrupted, and then compute the prediction error. That's the energy of the system. So this system will tell you this is a good representation
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
If this is a good image and this is a corrupted version, it will give you zero energy if those two things are effectively, one of them is a corrupted version of the other. It gives you a high energy if the two images are completely different.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And we know it does because then we use those representations as input to a classification system.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
for a number of reasons. The first is that there is a number of characteristics of intelligent behavior. For example, the capacity to understand the world, understand the physical world, the ability to remember and retrieve things, persistent memory, the ability to reason, and the ability to plan. Those are four essential characteristics of intelligent systems or entities, humans, animals.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I don't hate reinforcement learning, and I think it should not be abandoned completely, but I think its use should be minimized because it's incredibly inefficient in terms of samples. And so the proper way to train a system is to first have it learn good representations of the world and world models from mostly observation, maybe a little bit of interactions.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, now there's two things you can use. If you've learned a world model, you can use the world model to plan a sequence of actions to arrive at a particular objective. You don't need RL unless the way you measure whether you succeed might be inexact. Your idea of, you know, whether you were going to fall from your bike.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
might be wrong, or whether the person you're fighting with MMA was going to do something and then do something else. So there's two ways you can be wrong. Either your objective function does not reflect the actual objective function you want to optimize, or your world model is inaccurate. So the prediction you were making about what was going to happen in the world is inaccurate.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So if you want to adjust your world model while you are operating in the world or your objective function, that is basically in the realm of RL. This is what RL deals with to some extent, right? So adjust your world model. And the way to adjust your world model, even in advance, is to explore parts of the space where you know that your world model is inaccurate.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's called curiosity, basically, or play, right? When you play, you kind of explore part of the state space that you don't want to do for real because it might be dangerous, but you can adjust your world model without killing yourself, basically. So that's what you want to use RL for. When it comes time to learning a particular task,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You already have all the good representations, you already have your world model, but you need to adjust it for the situation at hand. That's when you use RL.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
What's had the transformational effect is human feedback. There's many ways to use it, and some of it is just purely supervised, actually. It's not really reinforced by learning.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's the HF. And then there is various ways to use human feedback, right? So you can ask humans to rate answers, multiple answers that are produced by a world model. And then what you do is you train an objective function to predict that rating. And then you can use that objective function to predict whether an answer is good.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And you can backpropagate gradient through this to fine-tune your system so that it only produces highly rated answers. That's one way. In RL, that means training what's called a reward model. Basically, a small neural net that estimates to what extent an answer is good.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's very similar to the objective I was talking about earlier for planning, except now it's not used for planning, it's used for fine-tuning your system. I think it would be much more efficient to use it for planning, but currently it's used to fine-tune the parameters of the system. Now, there are several ways to do this. Some of them are supervised.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You just ask a human person, what is a good answer for this? Then you just type the answer. I mean, there's lots of ways that those systems are being adjusted.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
LLMs can do none of those, or they can only do them in a very primitive way. They don't really understand the physical world. They don't really have persistent memory. They can't really reason, and they certainly can't plan. If you expect the system to become intelligent just without having the possibility of doing those things, you're making a mistake.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I actually made that comment on just about every social network I can, and I've made that point multiple times in various forums. Here's my point of view on this. People can complain that AI systems are biased, and they generally are biased by the distribution of the training data that they've been using.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
trend on, that reflects biases in society and that is potentially offensive to some people or potentially not. And some techniques to de-bias then become offensive to some people because of historical incorrectness and things like that. And so you can ask the question, you can ask two questions. The first question is, is it possible to produce an AI system that is not biased?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And the answer is absolutely not. And it's not because of technological challenges, although there are technological challenges to that. It's because Bias is in the eye of the beholder. Different people may have different ideas about what constitutes bias for a lot of things.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, there are facts that are indisputable, but there are a lot of opinions or things that can be expressed in different ways. And so you cannot have an unbiased system. That's just an impossibility. And so what's the... What's the answer to this? And the answer is the same answer that we found in liberal democracy about the press. The press needs to be free and diverse.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We have free speech for a good reason. is because we don't want all of our information to come from a unique source. Because that's opposite to the whole idea of democracy and progress of ideas and even science, right? In science, people have to argue for different opinions and And science makes progress when people disagree and they come up with an answer and, you know, a consensus forms, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And it's true in all democracies around the world. So there is a future. which is already happening, where every single one of our interaction with the digital world will be mediated by AI systems, AI assistants, right? We're going to have smart glasses.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You can already buy them from Meta, the Ray-Ban Meta, where you can talk to them, and they are connected with an LLM, and you can get answers on any question you have. Or you can be looking at a...
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
monument and there is a camera in the in the system that in in the glasses you can ask it like what can you tell me about this uh building on this monument you can be looking at a menu in a foreign language and the thing will translate it for you or you can do real-time translation if you speak different languages so
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
a lot of our interactions with the digital world are going to be mediated by those systems in the near future. Increasingly, the search engines that we're going to use are not going to be search engines. They're going to be dialogue systems that will just ask a question. And it will answer and then point you to perhaps an appropriate reference for it. But here is the thing.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We cannot afford those systems to come from a handful of companies on the west coast of the U.S., Because those systems will constitute the repository of all human knowledge. And we cannot have that be controlled by a small number of people, right? It has to be diverse. For the same reason, the press has to be diverse. So how do we get a diverse set of AI assistants?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That is not to say that autoregressive LLMs are not useful. They're certainly useful. That they're not interesting, that we can't build a whole ecosystem of applications around them. Of course we can, but as it paths towards human-level intelligence, they're missing essential components. And then there is another tidbit or fact that I think is very interesting.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's very expensive and difficult to train a base model, a base LLM at the moment. In the future, it might be something different, but at the moment, that's an LLM. So only a few companies can do this properly. And If some of those top systems are open source, anybody can use them. Anybody can fine tune them.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
If we put in place some systems that allows any group of people, whether they are individual citizens, groups of citizens, government organizations, NGOs, companies, whatever, to take those open source systems
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
systems, AI systems, and fine-tune them for their own purpose on their own data, then we're going to have a very large diversity of different AI systems that are specialized for all of those things, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So I tell you, I talked to the French government quite a bit, and the French government will not accept that the digital diet of all their citizens be controlled by three companies on the West Coast of the U.S. That's just not acceptable, right? It's a danger to democracy, regardless of how well-intentioned those companies are. And it's also a danger to local culture, to values, to language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I was talking with the founder of Infosys in India. He's funding a project to fine-tune Lama 2, the open-source model produced by Meta. so that Lama Thu speaks all 22 official languages in India. It's very important for people in India.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I was talking to a former colleague of mine, Moustapha Sissé, who used to be a scientist at FAIR, and then moved back to Africa, created a research lab for Google in Africa, and now has a new startup called Kera.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And what he's trying to do is basically have LLM that speaks the local languages in Senegal so that people can have access to medical information, because they don't have access to doctors. It's a very small number of doctors per capita in Senegal. You can't have any of this unless you have open-source platforms.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
With open-source platforms, you can have AI systems that are not only diverse in terms of political opinions or things of that type, but in terms of language, culture, value systems, political opinions. technical abilities in various domains. And you can have an industry, an ecosystem of companies that fine tune those open source systems for vertical applications in industry, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You have, I don't know, a publisher has thousands of books and they want To build a system that allows a customer to just ask a question about the content of any of their books, you need to train on their proprietary data. You have a company, we have one within Meta, it's called MetaMate, and it's basically an LLM that can answer any question about internal stuff about the company. Very useful.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
A lot of companies want this. A lot of companies want this not just for their employees, but also for their customers, to take care of their customers. So the only way you're going to have an AI industry, the only way you're going to have AI systems that are not uniquely biased is if you have open source platforms on top of which any group can build specialized systems.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So the inevitable direction of history is that the vast majority of AI systems will be built on top of open source platforms.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Those LLMs are trained on enormous amounts of text, basically the entirety of all publicly available texts on the internet, right? That's typically on the order of 10 to the 13 tokens. Each token is typically two bytes. So that's two 10 to the 13 bytes as training data. It would take you or me 170,000 years to just read through this at eight hours a day.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, so you have several business models, right? The business model that Meta is built around is, is you offer a service and the financing of that service is either through ads or through business customers. So for example, if you have an LLM that can help a mom and pop pizza place, by, you know, talking to the customers through WhatsApp.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so the customers can just order a pizza and the system will just, you know, ask them like, what topping do you want or what size, blah, blah, blah. Um, the business will pay for that. Okay. That's a model. Um, And otherwise, you know, if it's a system that is on the more kind of classical services, it can be ad supported or, you know, there's several models.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But the point is, if you have a big enough potential customer base and you need to build a system anyway for them, it doesn't hurt you to actually distribute it in open source.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, no, the bet is more, we already have a huge user base and customer base. Ah, right. Right, so it's going to be useful to them. Whatever we offer them is going to be useful, and there is a way to derive revenue from this. Sure. And it doesn't hurt that we provide that system or the base model, the foundation model, in open source for others to build applications on top of it, too.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
If those applications turn out to be useful for our customers, we can just buy it from them. It could be that they will improve the platform. In fact, we see this already. I mean, there is literally millions of downloads of Lama 2 already. and thousands of people who have provided ideas about how to make it better.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
This clearly accelerates progress to make the system available to a wide community of people. There are literally thousands of businesses who are building applications with it. Our ability to, Meta's ability to derive revenue from this technology is not impaired by the distribution of it, of base models in open source.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So it seems like an enormous amount of knowledge that those systems can accumulate. But then you realize it's really not that much data. If you talk to developmental psychologists and they tell you a four-year-old has been awake for 16,000 hours in his or her life, and the amount of information that has reached the visual cortex of that child in four years... is about 10 to the 15 bytes.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, I don't think the issue has to do with the political leaning of the people designing those systems. It has to do with the acceptability or political leanings of their customer base or audience, right? So a big company cannot afford to offend too many people. So they're going to make sure that whatever product they put out is safe, whatever that means. And it's very possible to overdo it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And it's also very possible to, it's impossible to do it properly for everyone. You're not going to satisfy everyone. So that's what I said before. You cannot have a system that is unbiased, that is perceived as unbiased by everyone.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's going to be, you know, you push it in one way, one set of people are going to see it as biased, and then you push it the other way, and another set of people is going to see it as biased. And then in addition to this, there's the issue of if you push the system perhaps a little too far in one direction, it's going to be non-factual, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You're going to have, you know, black Nazi soldiers in the image.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right, and can be offensive for some people as well, right? So it's gonna be impossible to kind of produce systems that are unbiased for everyone. So the only solution that I see is diversity.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Basically, yeah. I mean, Mark is right about a number of things that he lists that indeed scare large companies. You know, certainly congressional investigations is one of them, legal liability. making things that get people to hurt themselves or hurt others.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Big companies are really careful about not producing things of this type, because they don't want to hurt anyone, first of all, and then second, they want to preserve their business.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's essentially impossible for systems like this that can inevitably formulate political opinions and opinions about various things that may be political or not, but that people may disagree about, about, you know, moral issues and, you know, things about like questions about religion and things like that, right? Or cultural issues.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
issues that people from different communities would disagree with in the first place so there's only kind of a relatively small number of things that people will sort of agree on you know basic principles but beyond that if you if you want those systems to be useful they will necessarily have to offend a number of people inevitably
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's right. Open source enables diversity.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And you can compute this by estimating that the optical nerve carry about 20 megabytes per second, roughly. And so 10 to the 15 bytes for a four-year-old versus two times 10 to the 13 bytes for 170,000 years worth of reading What that tells you is that through sensory input, we see a lot more information than we do through language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah. I mean, there are some limits to what, you know, the same way there are limits to free speech, there has to be some limit to the kind of stuff that those systems might be authorized to do. to produce some guardrails.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So, I mean, that's one thing I've been interested in, which is in the type of architecture that we were discussing before, where the output of a system is a result of an inference to satisfy an objective. That objective can include guardrails. And we can put guardrails in open source systems. I mean, if we eventually have systems that are built with this blueprint.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We can put guardrails in those systems that guarantee that there is sort of a minimum set of guardrails that make the system non-dangerous and non-toxic, etc. You know, basic things that everybody would agree on. And then, you know, the fine-tuning that people will add or the additional guardrails that people will add will kind of cater to their community, whatever it is.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Right, so the increasing number of studies on this seems to point to the fact that it doesn't help. So having an LLM doesn't help you design or build a bioweapon or a chemical weapon if you already have access to a search engine and a library. And so the sort of increased information you get or the ease with which you get it doesn't really help you. That's the first thing. The second thing is,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's one thing to have a list of instructions of how to make a chemical weapon, for example, or a bioweapon. It's another thing to actually build it. And it's much harder than you might think, and an LLM will not help you with that. In fact, nobody in the world, not even countries, use bioweapons because most of the time they have no idea how to protect their own populations against it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So it's too dangerous, actually, to ever use. And it's, in fact, banned by... international treaties chemical weapons is different it's also banned by treaties but but it's the same problem it's difficult to use in situations that doesn't turn against the perpetrators but we could ask Elon Musk like I can I can give you a very precise list of instructions of how you build a rocket engine and
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And even if you have a team of 15 engineers that are really experienced building it, you're still going to have to blow up a dozen of them before you get one that works. And it's the same with chemical weapons or bio-weapons or things like this. It requires expertise in the real world that an airline is not going to help you with.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, exactly. A lot of biologists have posted on this actually in response to those things saying like, do you realize how hard it is to actually do the lab work? Like, you know, this is not trivial.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And that despite our intuition, most of what we learn and most of our knowledge is through our observation and interaction with the real world, not through language. Everything that we learn in the first few years of life and certainly everything that animals learn has nothing to do with language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, a number of things. So there's going to be various versions of LLAMA that are improvements of previous LLAMAs. Bigger, better, multimodal, things like that. And then in future generations, systems that are capable of planning, that really understand how the world works. Maybe are trained from video, so they have some world model.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Maybe, you know, capable of the type of reasoning and planning I was talking about earlier. Like how long is that going to take? Like when is the research that is going in that direction going to sort of feed into the product line, if you want, of Lama? I don't know. I can't tell you. And there is, you know, a few breakthroughs that we have to basically go through before we can get there.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But you'll be able to monitor our progress because we publish our research, right? So, you know, last week we published the Vijepa work, which is sort of a first step towards training systems for video. And then the next step is going to be World models based on this type of idea, training from video.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
There's similar work at DeepMind also and taking place people and also at UC Berkeley on world models from video. A lot of people are working on this. I think a lot of good ideas are appearing. My bet is that those systems are going to be JEPA-like, they're not going to be generative models. And we'll see what the future will tell.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
There's really good work at a gentleman called Danezhar Hafner, who is not DeepMind, who's worked on kind of models of this type that learn representations and then use them for planning or learning tasks by reinforcement training. And a lot of work at Berkeley by Peter Abbeel, Sergei Levine, a bunch of other people of that type.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I'm collaborating with, actually, in the context of some grants with my NYU hat. And then collaborations also through Meta, because the lab at Berkeley is associated with Meta in some way, with FAIR. So I think it's very exciting. I think... I'm super excited about... I haven't been that excited about the direction of machine learning and AI since 10 years ago when Fairway started.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Before that, 30 years ago, we were working on... 35 on convolutional nets and the early days of neural nets. So I'm super excited because I see a path towards... potentially human-level intelligence with systems that can understand the world, remember, plan, reason. There is some set of ideas to make progress there that might have a chance of working. And I'm really excited about this.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
What I like is that somewhat we get onto a good direction and perhaps succeed before my brain turns to a white sauce or before I need to retire.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, I used to be a hardware guy many years ago. Decades ago.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Changed a little bit.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, certainly scale is necessary, but not sufficient. Absolutely. So we certainly need computation. I mean, we're still far in terms of computer power. from what we would need to match the compute power of the human brain. This may occur in the next couple of decades, but we're still some ways away. And certainly in terms of power efficiency, we're really far.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So a lot of progress to make in hardware. And right now, a lot of the progress is not, I mean, there's a bit coming from silicon technology, but a lot of it coming from architectural innovation. And quite a bit coming from more efficient ways of implementing the architectures that have become popular, basically a combination of transformers and convnets, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So there's still some ways to go until... we're going to saturate, we're going to have to come up with new principles, new fabrication technology, new basic components, perhaps based on different principles than classical digital CMOS.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, if you want to make it ubiquitous, yeah, certainly. Because we're going to have to reduce the power consumption. A GPU today is half a kilowatt to a kilowatt. Human brain is about 25 watts. And a GPU is way below the power of the human brain. You need something like 100,000 or a million to match it. So we are off by a huge factor here.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So first of all, it's not going to be an event. The idea somehow, which is popularized by science fiction and Hollywood, that somehow somebody is going to discover the secret to AGI or human-level AI or AMI, whatever you want to call it, and then turn on a machine and then we have AGI. That's just not going to happen. It's not going to be an event. It's going to be gradual progress.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Are we going to have systems that can learn from video how the world works and learn good representations? Yeah. Before we get them to the scale and performance that we observe in humans, it's going to take quite a while. It's not going to happen in one day. Are we going to get systems that can have large amount of associative memory so they can remember stuff?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, but same, it's not going to happen tomorrow. I mean, there is some basic techniques that need to be developed. We have a lot of them, but to get this to work together with a full system is another story. Are we going to have systems that can reason and plan, perhaps along the lines of objective-driven AI architectures that I described before?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
yeah but like before we get this to work you know properly it's going to take a while so and before we get all those things to work together and then on top of this have systems that can learn like hierarchical planning hierarchical representations systems that can be configured for a lot of different situation at hands the way the human brain can you know all of this is going to take you know at least a decade and probably much more because there are
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
a lot of problems that we're not seeing right now, that we have not encountered. And so we don't know if there is an easy solution within this framework. So, you know, it's not just around the corner. I mean, I've been hearing people for the last 12, 15 years claiming that AGI is just around the corner and being systematically wrong. And I knew they were wrong when they were saying it.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I call their bullshit.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I don't think it's just Moravec's paradox. Moravec's paradox is a consequence of realizing that the world is not as easy as we think. First of all, intelligence is not a linear thing that you can measure with a single number. Can you say that humans are smarter than orangutans? In some ways, yes. But in some ways, orangutans are smarter than humans in a lot of domains.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That allows them to survive in the forest, for example.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, IQ can measure approximately something for humans. Mm-hmm. Because humans come in relatively uniform form, right? But it only measures one type of ability that may be relevant for some tasks, but not others. But then if you're talking about other intelligent entities for which the you know, the basic things that are easy to them is very different, then it doesn't mean anything.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So intelligence is a collection of skills and an ability to acquire new skills efficiently, right? And the collection of skills that a particular intelligent entity possess or is capable of learning quickly is different from the collection of skills of another one.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And because it's a multidimensional thing, the set of skills is a high dimensional space, you can't measure, you cannot compare two things as to whether one is more intelligent than the other. It's multidimensional.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Okay, so AI doomers imagine all kinds of catastrophe scenarios of how AI could escape or control and basically kill us all. And that relies on a whole bunch of assumptions that are mostly false. So the first assumption is that the emergence of superintelligence is going to be an event. That at some point we're going to figure out the secret and we'll turn on a machine that is superintelligent.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And because we've never done it before, it's going to take over the world and kill us all. That is false. It's not going to be an event. We're going to have systems that are like... as smart as a cat, have all the characteristics of human-level intelligence, but their level of intelligence would be like a cat or a parrot, maybe, or something. And then we're going to work our way up
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
to kind of make those things more intelligent. And as we make them more intelligent, we're also going to put some guardrails in them and learn how to kind of put some guardrails so they behave properly. And we're not going to do this with just one. It's not going to be one effort. It's going to be lots of different people doing this.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And some of them are going to succeed at making intelligent systems that are controllable and safe and have the right guardrails. And if some other goes rogue, then we can use the good ones to go against the rogue ones. So it's going to be my smart AI police against your rogue AI. So it's not going to be like we're going to be exposed to a single rogue AI that's going to kill us all.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's just not happening. Now, there is another fallacy, which is the fact that because the system is intelligent, it necessarily wants to take over.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
There are several arguments that make people scared of this, which I think are completely false as well. One of them is, in nature, it seems to be that the more intelligent species are the ones that end up dominating the other, and even extinguishing the others, sometimes by design, sometimes just by mistake.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so there is sort of thinking by which you say, well, if AI systems are more intelligent than us, surely they're going to eliminate us, if not by design, simply because they don't care about us. And that's just preposterous for a number of reasons. First reason is they're not going to be a species. They're not going to be a species that competes with us.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So it's a big debate among philosophers and also cognitive scientists, like whether intelligence needs to be grounded in reality. I'm clearly in the camp that, yes, intelligence cannot appear without some grounding in some reality. It doesn't need to be Physical reality could be simulated, but the environment is just much richer than what you can express in language.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
They're not going to have the desire to dominate because the desire to dominate is something that has to be hardwired into an intelligent system. It is hardwired in humans. It is hardwired in baboons, in chimpanzees, in wolves, not in orangutans. The species in which this desire to dominate or submit or attain status in other ways is is specific to social species.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Non-social species like orangutans don't have it, right? And they are as smart as we are, almost, right?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, there's all kinds of incentive to make AI systems submissive to humans, right? I mean, this is the way we're going to build them, right? And so then people say, oh, but look at LLMs. LLMs are not controllable. And they're right. LLMs are not controllable. But object-driven AI, so systems that derive their answers by optimization of an objective, means they have to optimize this objective.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And that objective can include guardrails. One guardrail is... Obey humans. Another gut reality is don't obey humans if it's hurting other humans.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yes. Maybe in a book.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, of course. So this is not a simple problem, right? I mean, designing those guardrails so that the system behaves properly is not going to be a simple issue for which there is a silver bullet, for which you have a mathematical proof that the system can be safe.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's going to be a very progressive, iterative design system where we put those guardrails in such a way that the system behaves properly. And sometimes they're going to do something that was unexpected because the guardrail wasn't right, and we're going to correct them so that they do it right.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The idea somehow that we can't get it slightly wrong because if we get it slightly wrong, we all die is ridiculous. We're just going to go progressively. And it's just going to be, the analogy I've used many times is turbojet design. How did we figure out how to make turbojets so unbelievably reliable, right? I mean, those are like, you know, incredibly complex things
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
pieces of hardware that run at really high temperatures for 20 hours at a time sometimes. We can fly halfway around the world on a two-engine jetliner at near the speed of sound. How incredible is this? It's just unbelievable. Did we do this because we invented a general principle of how to make turbojets safe? No.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It took decades to kind of fine-tune the design of those systems so that they were safe. Is there a separate group within General Electric or Snekma or whatever that is specialized in turbojet safety? No, the design is all about safety because a better turbojet is also a safer turbojet. So a more reliable one. It's the same for AI. Do you need specific provisions to make AI safe?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, you need to make better AI systems and they will be safe because they are designed to be more useful and more controllable.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Language is a very approximate representation of our percepts and our mental models. There's a lot of tasks that we accomplish where we manipulate a mental model of the situation at hand, and that has nothing to do with language. Everything that's physical, mechanical, whatever, when we build something, when we accomplish a task, a model task of grabbing something, etc.,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So that AI system designed by Vladimir Putin or whatever, or his minions, you know, is going to be... Trying to talk to every American to convince them to vote for whoever pleases Putin or whatever, or rile people up against each other as they've been trying to do. They're not going to be talking to you. They're going to be talking to your AI assistant.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
which is going to be as smart as theirs, right? That AI, because as I said, in the future, every single one of your interaction with the digital world will be mediated by your AI assistant. So the first thing you're going to ask is, is this a scam? Like, is this thing like telling me the truth? It's not even going to be able to get to you because it's only going to talk to your AI system.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Your AI system is not even going to... It's going to be like a spam filter, right? You're not even seeing the email, the spam email, right? It's automatically put in a folder that you never see. It's going to be the same thing. That AI system that tries to convince you of something is going to be talking to your AI system, which is going to be at least as smart as it is.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And it's going to say, this is spam. It's not even going to bring it to your attention.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That's the history of the world. History of the world is, you know, whenever there is a progress someplace, there is a countermeasure. And, you know, it's a cat and mouse game.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No. As I said, it's not going to be an event. It's going to be continuous progress. And whenever one breakthrough occurs, it's going to be widely disseminated really quickly, probably first within industry. I mean, this is not a domain where government or military organizations are particularly innovative, and they're in fact way behind.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so this is going to come from industry and this kind of information disseminates extremely quickly. We've seen this over the last few years, right? Where you have a new, like, you know, even take AlphaGo, this was reproduced within three months, even without like particularly detailed information.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, but even if there is, just the fact that you know that something is possible makes you realize that it's worth investing the time to actually do it. You may be the second person to do it, but you'll do it. And same for all the innovations of self-supervised learning, transformers, decoder-only architectures, LLMs.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, those things, you don't need to know exactly the details of how they work to know that it's possible because it's deployed and then it's getting reproduced. And then people who work for those companies move. They go from one company to another and the information disseminates.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
What makes the success of the US tech industry and Silicon Valley in particular is exactly that, is because information circulates really, really quickly. and this, you know, disseminates very quickly. And so, you know, the whole region sort of is ahead because of that circulation of information.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We plan our action sequences, and we do this by essentially imagining the result of the outcome of a sequence of actions that we might imagine. And that requires mental models that don't have much to do with language. And that's, I would argue, most of our knowledge is derived from that interaction with the physical world.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, there is a natural fear of new technology and the impact it can have on society. And people have kind of instinctive reaction to the world they know being threatened by major transformations that are either cultural phenomena or technological revolutions. And they fear for their culture, they fear for their job, they fear for the future of their children and their way of life.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So any change is feared. And you see this, you know, along history, like any technological revolution or cultural phenomenon was always accompanied by, you know, groups or reaction in the media that basically attributed all the problems, the current problems of society to that particular change, right? Electricity was going to kill everyone.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
At some point, you know, the train was going to be a horrible thing because, you know, you can't breathe past 50 kilometers an hour. And so there's a wonderful website called the Pessimist Archive, which has all those newspaper clips of all the horrible things people imagine would arrive because of either technological innovation or a cultural phenomenon. You know,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
These are wonderful examples of jazz or comic books being blamed for unemployment or young people not wanting to work anymore and things like that. That has existed for centuries. It's knee-jerk reactions. The question is, do we embrace change or do we resist it? And what are the real dangers as opposed to the imagined ones?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So a lot of my colleagues who are more interested in things like computer vision are really on that camp that AI needs to be embodied, essentially. And then other people coming from the NLP side or maybe some other motivation don't necessarily agree with that. And philosophers are split as well. And the complexity of the world is hard to imagine. It's hard to
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, that's exactly why we need open source platforms.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah. Again, I think the answer to this is open source platforms and then enabling a widely diverse set of people to build AI assistants that represent the diversity of cultures, opinions, languages, and value systems across the world, so that you're not bound to just be brainwashed by a particular way of thinking because of a single AI entity.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So, I mean, I think it's a really, really important question for society. And the problem I'm seeing is that, which is why I've been so vocal and sometimes a little sardonic about it. Never stop. Never stop, Jan.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
It's because I see the danger of this concentration of power through proprietary AI systems as a much bigger danger than everything else. That if we really want diversity of opinion, AI systems that in this future where we'll all be interacting through AI systems.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
We need those to be diverse for the preservation of diversity of ideas and creeds and political opinions and whatever, and the preservation of democracy. And what works against this is people who think that for reasons of security, we should keep AI systems under lock and key because it's too dangerous to put it in the hands of everybody because it could be used by terrorists or something.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That would lead to potentially A very bad future in which all of our information diet is controlled by a small number of companies through proprietary systems.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Isn't that what democracy and free speech is all about? I think so. Do you trust institutions to do the right thing? Do you trust...
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
people to do the right thing and yeah there's bad people who are going to do bad things but they're not going to have superior technology to the good people so then it's going to be my good ai against your bad ai right i mean it's the examples that we were just talking about of you know maybe some rogue country will build you know some ai system that's going to try to convince everybody to
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
go into a civil war or something or elect a favorable ruler. But then they will have to go past our AI systems.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And doesn't put any articles in their sentences.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Not soon, but it's going to happen. The next decade, I think, is going to be really interesting in robots. The emergence of the robotics industry has been in the waiting for 10, 20 years without really emerging other than for pre-programmed behavior and stuff like that. And the main issue is Again, the Moravec paradox, how do we get the system to understand how the world works and plan actions?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
represent all the complexities that we take completely for granted in the real world that we don't even imagine require intelligence, right? This is the old Moravec paradox from the pioneer of robotics, Hans Moravec, who said, you know, how is it that with computers it seems to be easy to do high-level complex tasks like playing chess and solving integrals and doing things like that, whereas
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so we can do it for really specialized tasks. And the way Boston Dynamics goes about it is basically with a lot of handcrafted dynamical models and careful planning in advance, which is very classical robotics with a lot of innovation, a little bit of perception. But it's still not, they can't build a domestic robot.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And we're still some distance away from completely autonomous level five driving. And we're certainly very far away from having level five autonomous driving by a system that can train itself by driving 20 hours like any 17 year old. So until we have, again, world models, systems that can train themselves to understand how the world works,
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
uh, we're not going to, we're not going to have significant progress in robotics. So a lot of the people working on robotic hardware at the moment are, are betting or banking on the fact that AI is going to make sufficient progress towards that.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I mean, there's, you know, cleaning up, cleaning the house, clearing up the table after a meal, washing the dishes, you know, all those tasks, you know, cooking. I mean, all the tasks that, you know, in principle could be automated, but are actually incredibly sophisticated, really complicated.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
That sort of works. Like you can sort of do this now. Navigation is fine.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Yeah, it's not going to be, you know, necessarily. I mean, we have demos actually because, you know, there is a so-called embodied AI group at FAIR. And, you know, they've been not building their own robots, but using commercial robots. Yeah.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And you can tell a robot dog, like, you know, go to the fridge and they can actually open the fridge and they can probably pick up a can in the fridge and stuff like that and bring it to you. So it can navigate, it can grab objects as long as it's been trained to recognize them, which, you know, vision systems work pretty well nowadays.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
But it's not like a completely general robot that would be sophisticated enough to do things like clearing up the dinner table.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
The thing we take for granted that we do every day, like, I don't know, learning to drive a car or, you know, grabbing an object. We can't do it with computers. And, you know, we have LLMs that can pass the bar exam. So they must be smart. But then they can't learn to drive in 20 hours like any 17-year-old.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Well, I mean, I hope things can work as planned. I mean, again, we've been kind of working on this idea of self-supervised learning from video for 10 years and only made significant progress in the last two or three years.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So basically, I've listed them already. This idea of how do you train a world model by observation? Mm-hmm. And you don't have to train necessarily on gigantic datasets. I mean, it could turn out to be necessary to actually train on large datasets to have emergent properties like we have with LLMs. But I think there's a lot of good ideas that can be done without necessarily scaling up.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Then there is how you do planning with a learned world model. If the world the system evolves in is not the physical world, but it's the world of... Let's say the Internet or some world where an action consists in doing a search in a search engine or interrogating a database or running a simulation or calling a calculator or solving a differential equation.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
How do you get a system to actually plan a sequence of actions to give the solution to a problem? So the question of planning is not just a question of planning physical actions. It could be planning actions to use tools for a dialogue system or for any kind of intelligent system. And there's some work on this, but not a huge amount.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
Some work at FAIR, one called Toolformer, which was a couple years ago, and some more recent work on planning. But I don't think we have a good solution for any of that. Then there is the question of hierarchical planning. So the example I mentioned of planning a trip from New York to Paris, that's hierarchical. But almost every action that we take involves hierarchical planning in some sense.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And we really have absolutely no idea how to do this. There's zero demonstration of hierarchical planning in AI. where the various levels of representations that are necessary have been learned. We can do two-level hierarchical planning when we design the two levels. For example, you have a dog-like robot. You want it to go from the living room to the kitchen.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
You can plan a path that avoids the obstacle. And then you can send this to a lower-level planner that figures out how to move the legs to kind of follow that trajectory. So that works, but that two-level planning is designed by hand. We specify what the proper levels of abstraction, the representation at each level of abstraction have to be. How do you learn this?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
How do you learn that hierarchical representation of action plans? With Cognites and deep learning, we can train the system to learn hierarchical representations of percepts. What is the equivalent when what you're trying to represent are action plans?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
For example.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
They can't learn to clear out the dinner table and fill up the dishwasher like any 10-year-old can learn in one shot. Why is that? Like, you know, what are we missing? What type of learning or reasoning architecture or whatever are we missing that basically prevent us from, you know, having level five self-driving cars and domestic robots?
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
No, but even doing something fairly simple like a household task, like cooking or something.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
I love that question. We can make humanity smarter with AI. Okay. I mean, AI basically will amplify human intelligence. It's as if every one of us will have a staff of smart AI assistants. They might be smarter than us. They'll do our bidding. perhaps execute tasks in ways that are much better than we could do ourselves because they'd be smarter than us.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
And so it's like everyone would be the boss of a staff of super smart virtual people. So we shouldn't feel threatened by this any more than we should feel threatened by being the manager of a group of people, some of whom are more intelligent than us. I certainly have a lot of experience with this, of having people working with me who are smarter than me. That's actually a wonderful thing.
Lex Fridman Podcast
#416 – Yann Lecun: Meta AI, Open Source, Limits of LLMs, AGI & the Future of AI
So having machines that are smarter than us that assist us in our work, all of our tasks, our daily lives, whether it's professional or personal, I think would be an absolutely wonderful thing. Because intelligence is the most, is the commodity that is most in demand.