Yan laon the head of meta's AI division just did a 3our interview with Lex fredman in…
- User
Yan laon the head of meta's AI division just did a 3our interview with Lex fredman in which he covered robots AGI and so much more I went through the whole interview I grabbed the bits that I thought were most interesting and I encourage you to watch the whole video I'll drop a link to that in the description below in this first clip Yan talks about how autor regressive llms basically the large language models that we all know today like open AI chaty PT won't get us to AGI Yan is a huge proponent of Open Source but he doesn't think our current technology will allow us to achieve AGI let's watch 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 characteristic of intelligent systems or entities humans animals LMS 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 you know without having the possibility of doing those things uh you're making a mistake and then next following up on the premise that our current Technologies won't get us to AGI Yan talks about the amount of data necessary to train large language models as compared to what humans use to you know essentially train our own minds and he says that the amount of data used to train large language models is actually not that much compared to what humans need and this is where we start to get into synthetic data and I've been preparing my thoughts about synthetic data quite a bit because the fact of the matter is I don't think humans are going to Pro I don't think humans are going to produce enough data to actually reach AGI using large language models even if our current architecture could allow us to get there so synthetic data meaning meaning data that has been created by other artificial intelligence is one of those necessary ingredients to reaching AG let's take a look at this video 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 byit 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 8 hours a day uh so it seems like an enormous amount of knowledge right that those systems can accumulate but then you realize it's really not that much data if you you talk to developmental psychologist and they tell you a four-year-old has been awake for 16,000 hours in his or 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 and you can compute this by estimating that the optical nerve carry about 20 meab megabytes per second roughly and so so 10 to the 15 bytes for a 4-year-old versus 2 * 10 13 bytes for 170,000 years worth of reading what it tells you is that through sensory input we see a lot more information than we than we do through language 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 and then again continuing with the large language models are not going to allow us to reach AGI he basically says that just using language is not enough to model the world to have a world model even though language is highly compressed he describes how when people are talking and thinking about things they're using much more than just language we have a world model inside our head and we're conceptualizing ideas and perceptions before we even convert those into language it's very interesting that think about but because of this he says we're not going to reach AGI with language alone we need some other kind of technology to help us augment the language part let's look at that video it's a big debate among uh philosophers and also cognitive scientists like whether intelligence needs to be grounded in reality uh I'm clearly in the camp that uh yes uh intelligence cannot appear without some grounding in uh some reality it doesn't need to be you know physical reality could be simulated but um but the environment is just much richer than what you can express in language language is a very approximate representation of our percepts and our mental models right I mean there there's a lot of tasks that we accomplish where we manipulate uh a mental model of uh of the situation at hand and that has nothing to do with language and next kind of proving the point that large language models in their current form aren't going to reach AGI Yan goes on to describe why large language models are really good at some things like creative writing like coding passing the bar exam and other things like that but really bad at other things like self-driving and the ability to Simply pick up a cup and for some reason large language models are really good at some things and really bad at others so let's take a look you know we have llms that can pass 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 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 what are we missing what what type of learning or or reasoning architecture or whatever are we missing that basically prevent us from from you know having level five sing Cars and domestic robots and then finally kind of closing the loop on AGI and what's possible with our current large language model architecture he goes on to talk about how kind of world model reasoning is Out Of Reach for large language models as they are today he continues to explain that when humans are talking about things and conceptualizing things in the real world we're doing so much more than just speaking in our minds so he's really doubling down on this point which I find fascinating let's take a look at this clip we're not going to be able to do this with the type of llms that we are uh working with today and there's a number of reasons for this but uh the main reason is the way llm llms are trained is that you you take a piece of text remove some of the words in that text you mess them you replace by replace by blank markers and you train a gnetic neural net to predict the words that are missing 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 is 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 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 subw units and so it's easy to handle the uncertainty in the prediction there because there's only a finite number of possible words in the dictionary and you can just compute a distribution over them then what you what the system does is that it 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 that 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 and once you do this you shift it into the input Etc that's called Auto regressive prediction and which is why those llms should be called Auto regressive llms 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 bilinguals 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 we talk about like I don't know let's say mathematical concept or something the kind of thinking that we're doing and the answer that we're planning to produce is not linked to whether we're going to see it in French or Russian or English and next we continue our discussion about language as a world model now Yan is in one camp that believes language alone cannot be a world model and I want to know what you think do you agree or disagree when you see something like Sora which uses Transformers which uses current architecture that is found in large language models it is pretty darn good at simulating the world but it's not perfect and if it's ever going to get there is still a question and he continues to hammer the point that large language models are not enough to reach a world model but this actually reminds me of a couple things there's the movie arrival and if you haven't seen it it's a fantastic science fiction movie but the gist is aliens come down from space and they speak a different language and they have an expert linguist try to figure out what that language is and over time they figure it out and the language itself is what the aliens are there to give Humanity as kind of the great gift and all of a sudden because this expert langu now understands this new language she starts to gain all of this power and so yeah it's super interesting to think about that language actually shapes perception and that is something I believe in if we don't have a word for something then most likely we can't describe it and maybe we can't even perceive it as it really is so I believe language is incredibly important but I'll probably defer to Yan thinking that language alone is not enough to have a world model you're saying your thinking is same in French as it is in English well it it depends what kind of thinking right if it's just uh if it's like producing puns I get much better in French than English about that no but so is there an abstract representation of puns like is your humor an abstract like when you tweet and your tweets are sometimes a little bit spicy what's is there an abstract representation in your brain of a tweet before maps on to English there is an asct representation of uh Imagining the reaction of a reader to that uh text or you start with laughter and then figure out how to make that happen or figure out like a reaction you want to cause and and then figure out how to say it right so that it causes that reaction but that's like really close to language but think about like a ma mathematical concept or um you know imagining you know something you want to build out of wood or something like this right the kind of thinking you're doing has absolutely nothing to do with language really like it's not like you have necessarily like an internal monologue in any particular language you you're you know imagining mental models of of the thing right I mean if I if I ask you to like imagine what this uh water bottle will look like if I rotate it 90° that has nothing to do with language so clearly there is you know a more abstract level of representation in which we we do most of our thinking and we plan what we're going to say if the output is is you know uttered words as opposed to an output being uh you know muscle actions we we plan our answer before we produce it and LMS don't do that they just produce one word after the other instinctively if you want and in this next clip Yan talks about if we're able to have a world model based on prediction alone and again the way that large language models work is that it's essentially just predicting the next token in a sentence the next word in a sentence and he actually says yes just based on prediction alone that part of the technology we will be able to achieve AGI we will be able to achieve a world model but just based on language prediction alone he doesn't think so and this starts to get into simulation Theory so do you need to understand every single atom do you need to predict how every single atom is going to move to truly create a world model it turns out maybe not and this is actually something a Twitter user had replied to me which I had never really thought about with simulation Theory I always always thought that you had to predict every single atom but that's not necessarily true you only have to predict enough that the perceiver the person who is viewing this reality is convinced that it looks real so just enough but not everything so let's take a look at this clip so can you build this first of all by prediction 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 so building World models means observing the world and understanding why the world is evolving the way the way it is and then uh the 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 World model really is here is my idea of the state of the world at time T here is an action I might take what is the predicted state of the world at time t plus one now that state of the world doesn't does not need to represent everything about the world it just needs to represent enough that's relevant for this planning of 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 so a generative model has 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 either one friend at a time or a group of friends at a time but yeah uh a large video model if you want the idea of of doing this has been floating around for a long time and at at Fair colleagues and I have been trying to do this for about 10 years and you can't you can't really do the same trick as with llms because you know llms as I said you can't predict exactly which word is going to follow a sequence of words you can predict a distribution of words now if you go to video what you would have to do is predict a distribution over all possible frames in a video and we don't really know how to do that properly we do not know how to represent distribution over High dimensional continuous spaces in ways that are useful and that's that 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 than text text is discrete video is high dimensional and continuous a lot of details in this okay in this next clip he breaks down why video prediction basically frame by frame prediction or even an entire video prediction is really hard to do and this interview came out after Sora so as you're listening to this think about what Sora was able to achieve and compare it to what Yan is saying here so if I take a video of this room and the video is you know 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 this is a room where there's a light and there is a wall and things like that I 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 I can predict all those details 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 an Neal net and it's supposed to represent all the information about the world that you don't perceive yet that you need to augment the the system for the prediction to do a good job at predicting pixels including the fine texture of the of the carpet and the and the couch and and the painting on the wall that has been a complete failure and we've tried lots of things we tried uh just straight neural Nets we tried Gans we tried Vees all kinds of regularized Auto encoders we tried many things and in this next clip he goes into the details about why video prediction doesn't work with our current architecture our current technology and of course Yan recently released jeppa which is what he believes is going to be able to augment large language models and Transformers to be able to allow to actually have a world model with video so let's see what he says about that so the reason this doesn't work first of all I have to tell you exactly what doesn't work because there is something else that does work uh so the thing that does not work is training a 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 and we have a whole slew of techniques for this uh that are you know variant of dining Auto incoder something called Mae developed by some of my colleagues at Fair Max Doo encoder so it's basically like the you know llms or or or or things like this where you train a system by corrupting text except you corrupt images you remove Patches from it and you train gantic neet to reconstruct 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 supervis with label data with text textual descriptions of images you do get good representations and the performance on recognition task is much better than if you do this self-supervised pre trining the architecture of the encoder is good okay but the fact that you train the system to reconstruct images does not lead it to produce to learn good generic features of images when you train a self-supervised way self supervised by reconstruction and now let's watch what he has to say about jepa and again jepa was just released by meta under Yan's leadership with his entire team and it's a very interesting project it gives the ability to augment large language models with the ability to predict video and I'm still wrapping my mind around it so I'm going to let him explain it the alternative is joint embedding so now instead of training a system to encode the image and then training it to reconstruct the 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 and then you 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 you're taking the the full input and the corrupted version run them both through encoders you get a joint and then you're you're saying can I predict the representation of the full one from the representation of the corrupted one and I call this a JEA so that means joint Ting predictive architecture because J embedding and there is this predictor that predicts the representation of the good guy from from the bad guy and the big question is how do you train something like this and until 5 years ago or six years ago we didn't have particularly good answers and now he's going to talk about about what is the difference between large language models and jeppa and why them working together might be the solution to reach a gii and a world model first of all uh what's the difference with generative architectures like llms so llms or Vision systems that are trained by reconstruction generate the inputs right they generate the original input that is non-corrupted non-transformed so you have to predict all the pixels and there is a huge amount of resources spent in the system to actually predict all those pixels all the details uh in a JEA you're not trying to predict all the pixels you're only trying to predict an abstract representation of of the inputs and that's much easier in many ways so what the Jad 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 so there's a lot of things in the world that we cannot predict like for example if you have a cell driving car driving down the street there may be uh trees around the around the road and it could be a windy day so the 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 so what you want is your anoder to basically eliminate all those details 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 a relief and that not only is a lot simpler but also it allows the system to essentially learn an abstract representation of of the world where you know what can be model then 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 whenever we describe a phenomenon we describe it at a particular level of abstraction and we don't always describe every natural phenomenon in terms of quantum field Theory right that would be impossible right so we have multiple levels of abstraction to describe what happens in the world starting from Quantum field Theory to like atomic theory and ules you know in chemistry you know all the way up to you know kind of concrete objects in the real world and things like that so the we we can't model everything at the lowest level that's what the idea of JEA is really on is really about learn abstract representation in a self-supervised uh Manner and you know you can do it hierarchically as well so that I think is an essential component of an intelligent system and 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 so we can get away without doing the J Bing without you know lifting the abstraction level and by directly two words now in this next clip Yan talks about hierarchical planning and what that essentially means is there is a ton of planning that needs to happen for basic actions that we take every single day so to describe hierarchical planning let's use an example let's say I want to drive to the store so first I have to get up and walk to my car so how do I do that well first I have to actually stand up well how do I do that well my brain has to tell my legs to use certain muscles to push me up and then how do I do that and it's this recursive algorithm that is required and every single step from something as basic as just walking to the car is actually thousands or really just infinite steps if you really think about it and so how do you model that how do you actually model that with current architecture and so that's what he starts to talk about here let's watch so hierarchical planning is absolutely necessary if you want to plan complex actions uh if I want to go from let's say from New York to Paris this the example I use all the time and I'm sitting uh in my office at NYU my objective that I need to minimize is my distance to Paris at a high level a very as track representation of my uh my location I will 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 where I have to go in the street and H a taxi which you can do in New York okay now I have another sub goal go down on the street uh well that means uh going to the elevator going down the elevator walk out the street how do I go to the elevator I have to uh stand up from my chair open the door of my office go to the elevator push push the button how do I get up from my chair like you know you can imagine going down all the way down to basically what amounts to millisecond by millisecond muscle control 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 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 I mean you you would have to know exactly the condition of everything to be able to do this planning and you don't have the information you you have to do this hierarchical planning so that you can start acting and then sort of replanning as you go nobody really knows how to do this in AI nobody knows how to train a system to learn the appro multiple levels of representation so that hierarchical planning works and now Lex is going to press Yan about scaling because what we've seen with the ability to scale large language models is truly amazing we've gone from GPT 1 2 3 now we're at four almost at five and the increase in parameters the increase in GPU processing power has Unleashed incredible capabilities from these large language models and now you have Sora and Sora proved if you increase compute you can get much better results so Lex really tries to press Leon about hey if we keep scaling up can't we reach super intelligence with our current Technologies and he also asked what Alan Turing would think about current Ai and if you're not familiar with Alan Turing he created the Turing test amongst many other things and the Turing test is simply a test to see if artificial intelligence is truly intelligent and all it really needs to do is convince a human that it is human and with our current technology that's very possible especially if you just do it over text but even with voice it's possible and so Yan says something really funny about it I'm not going to spoil that joke take a look does it make sense to you that they're able to form enough of a representation in the world to be damn convincing essentially passing the original touring test with flying colors well we're fooled by their fluency right we just assume that if a system is fluent manipulating language then it has all the characteristics of human intelligence but that impression is false we're really fooled by it what do you think Alan tan would say it without understanding anything just hanging out with it an on Turing would decide that a turning test is a really bad test okay this is what the AI Community has decided many years ago that the train test was a really bad test of intelligence next Yan discusses hallucinations why do they happen why are they such a big limitation for current large language models and it is a huge limitation it's probably probably one of the biggest limitations to it being widely adopted in the corporate setting there are a lot of techniques to prevent hallucinations but let's find out why they actually occur so because of the auto regressive prediction every time at an produces a token or 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 M what that means is that every time you produce a token the probability that you rest you you stay within the the set of correct answer decreases and it decreases exponentially if there's a nonzero probability of making a mistake which there appears to be then there's going to be a kind of drift yeah and that drift is exponential it's like errors accumulate so so the probability that an answer will be nonsensical increases exponentially with the number of tokens all right this next clip is very interesting Yan talks about why he doesn't like reinforcement learning or at least why reinforcement learning is very good but only for a very narrow use case let's watch why do you still hate uh reinforcement learning I don't hate reinforcement learning and I think I think it should not be abandoned completely but I think it its use should be minimized because it's incredibly inefficient in terms of samples and so the 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 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 ourl unless the way you measure whether you succeed might be inexact your idea of whether you're going to fall from your bike might be wrong or whether the person you're fighting with MMA is going to do something and do something else so there's two ways you can be wrong either your your objective function does not reflect the actual objective function you want to optimize or your world model is inaccurate the prediction you were making about what was going to happen in the world is inaccurate if you want to adjust your world model while you are operating the world or your objective function that is basically in the realm of RL this is what RL deal 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 for your world model where you know that your world model is inaccurate that's called curiosity basically or play when you play you kind of explore part of the stay space that you know you don't want to do in for real because it might be dangerous but but you can adjust your world model without killing yourself basically so that's what you want to use RL for when when when it comes time to learning a particular task you already have all the good representations you already have your world model but you want you need to adjust it for the situation at hand that's when you use RL now to the part that I'm truly excited about Yan talks about why open source is so important he references Gemini 1.5s wokeness he references llm biases and why opsource models are the answer to all of it and I tend to agree with much of what he says so let's take a look a lot of people have been very critical of the recently released Google's Gemini 1.5 for essentially in my words I could say super woke woke in the uh negative connotation of that word uh there is some almost hilariously absurd things that it does like it modifies history generating images of a u black George Washington or perhaps more seriously something that you commented on Twitter which is refusing to comment on or generate images of uh or even descriptions of tianan square or the the Tank Man one of the most sort of legendary protest images in history and of course these images are highly censored by the Chinese government and therefore everybody started asking questions of what is the process of Designing these llms what is what is what is the role of censorship in these so you uh commented on Twitter saying that open source is the answer can you explain I I actually made a comment on just about every Social Network I can and I've I I've uh I've made that point multiple times in in various forums um here's my 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 trained on that reflects biases in society and that is potentially offensive to some people or potentially not and and some techniques to debias then become offensive to some people because of you know 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 and the answer is absolutely not and it's not because of technological challenges although they are technological challenges to that it's because bias is in the eye of the beholder different ideas about what constitutes bias you know for a lot of uh a lot of things I mean there are facts that are you know indisputable but there are a lot of opinions or 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 the answer is the same answer that we found in Liberal democracy about the press the press needs to be free and uh diverse we have free speech for a good reason is because we don't want all of our information to come from a unique Source cuz that's opposite to the whole idea of democracy and progress of ideas and even science right in in science people have to argue for different opinions and science makes progress when people disagree and they come up with an answer and you know a consensus forms right and it's true in all democracies around the world so there is a a future is already happening where every single one of our interaction with the digital world will be mediated by AI systems we're going to have smart glasses you can already buy them from MAA the ran MAA where you know you can talk to them and they are connected with an llm and you you can get answers on any question you have or you can be looking at a monument and there is a camera in the in the system that in in the glasses you can ask it like can what can you tell me about this uh building or this Monument you can be looking at a menu in a foreign language and it thing will translate it for you or can do realtime translation if we speak different languages so a lot of our interaction with the digital world are going to be mediated by those systems in the near future you know increasingly the search engines that we're going to use are not going to be search engines they're going to be dialog systems we just ask a question and it will answer and then point you to perhaps appropriate reference for it but here is the thing we cannot afford those systems to come from a handful of companies on the west coast of the US 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 assistance um it's very expensive and difficult to train a based model right a based llm at the moment you know in the future might be something different but at the moment that's an llm so only a few companies can do this properly if some of those top systems are open source anybody can use them anybody can fine-tune them if we put in place some systems that allows any group of people whether they are individual citizens groups of citizens government organizations NOS companies whatever to take those open- Source 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 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 citizen be controlled by three companies on the west coast of the US that's just not acceptable 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 I was talking with the founder of infosis in India he's funding a project to fine-tune Lama 2 the open source model produced by by meta so that Lama 2 speaks all 22 official languages in India I mean you can't have any of this unless you have open source platforms so 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 you 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 just ask a question about any the content of any of their books you need to train on their proprietary data you have a company we have one within meta is called Metate and it's basically an llm that can answer any question about internal stuff about about the company very useful a lot of companies want this right 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 the direction of of inevitable direction of history is that the vast majority of AI systems will be built on top of Open Source platforms and on the topic of Open Source Lex asks him okay if it's open source how do you actually run a business based on open source what are the economics of Open Source and this has been proven over the years we have a lot of very successful open- Source projects that have made a lot of money many of the standards that we use on the internet today many of the standards that we use with databases with code architecture everything a lot of that is open source and there are a lot of benefits with open source and a lot of companies have made a lot of money with open source just look at Android for Google just as an example off the top of my head so let's see what Yan says about the economics of meta's open- source contributions you have several business models right the business model that uh meta is built around is you offer a service and the the financing of that service is uh either through ads or through business customers so for example if you have an llm that uh you know can help a mom and pop pizza place by you know talking to the customers through WhatsApp and so the customers can order a pizza and the system will just you know ask them like what topping do you want or websites blah blah blah the business will pay for that and otherwise you know if it's a system that is on the more kind of classical Services it can be uh ad supported or you know there's several models but the point is if you have a big enough potential customer base and you need to build that system anyway for them it doesn't hurt you to actually distribute it an open source and then Lex continues to press him about the open source economics why doesn't another company just come along take that open source project and build competition and that probably will happen and that's probably good but at the end of the day meta is still meta they have an entire multi-billion human user base to sell their products and services to so let's see what he says the the bet is is more we have we already have a huge user base and customer base 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 and it doesn't hurt that you know we provide that system or the base the base model right the foundation model in open source for others to build applications on top of it too 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 you know literally millions of downloads of uh Lama 2 and thousands of people who have you know provided ideas about how to make it better so you know this this clearly accelerates progress to make the system available to a wide community of people and there's 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 based models in open source now we continue on the whole biases front many employees at tech companies tend to lean left and that is a point that Lex makes and I tend to agree with and so aren't those biases going to be built into large language models and again Yan points to open source being the answer 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 the their customer base 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 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 is perceived as unbiased by everyone 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 you're going to have you know black Nazi soldiers in yes we should we should mention image generation of of uh black Nazi soldiers which is not factually accurate right and can be offensive for some people as well right so you know it's going to be impossible to kind of produce systems that are unbiased for everyone so the only solution that I see is is diversity and diversity in the full meaning of that word diversity of in every possible way and continuing on the open source front now they're going to talk about Free Speech guard rails on AI censorship biases and everything I love this conversation and hearing Yan frame open source in this light makes a ton of sense to me let's watch 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 to produce you know some guardrail 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 the system is the result of an inference to satisfy an objective that objective can include guard rails and we can put guard rails in open source systems I mean if we eventually have systems that are built with this blueprint we can put guard rails in those systems that guarantee that there is sort of a minimum set of guard rails that make the system non- dangerous and non-toxic Etc you know basic things that everybody would agree on and then you know the the fine tuning that people will add or the additional guardwell that people will add will kind of cater to their Community whatever it is next if you wanted a little preview of when llama 3 is coming and what it's going to be about this is the clip for you let's watch there's going to be like various versions of of llama that are you know improvements of previous Lamas bigger better multimodal things like that and then in future generation systems that are capable of planning that really understand how the world Works maybe are trained from video so they have some World model 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 doing 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 you'll be able to monitor our progress because we publish our research right so last week we published the Via work which is sort of a first step towards Training Systems from video and then the next step is going to be World models based on kind of this type of idea training training from video there similar work at a Deb also and taking place people and also at UC Berkeley on world models video a lot of people are working on this I think a lot of good ideas are coming are appearing my bet is that those systems are going to be Jeet likee they're not going to be gen generative models and uh we'll see what the future will tell there's really good work at a gentleman called danar Hafner who is not Deep Mind who who's worked on kind of models of this type that learn representations and then use them for planning or learning tasks by reinforcement learning now going back a few topics they're going to talk about the power and efficiency in the human mind to process all of this data and what we get from it compared to large language models and it's really not even a comparison large language models require a ton of data a ton of processing power a ton of energy to train and to use versus a human brain which is incredibly efficient we're still far in terms of computer power from you know what we would need to match the compute power of the human brain you know this may occur in the next couple decades but but we're still some ways away and certainly in terms of power efficiency we're really far so a lot of progress to make in uh in in in Hardware right now a lot of progress is is is not I mean there 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 you know implementing the architectures that have become popular basically combination of Transformers and comets so you know there's still some ways to go until we're going to saturate we're going to have to come up with like new new principles new fabrication technology new uh basic components perhaps you know based on sort of different principles and those classical digital semas and next they're going to start talking about AGI when is it coming and Yan has made this point over and over again a lot of people think AGI is just going to be some switch that is flipped one day and it's just going to be this inflection point that AGI just has this quote unquote hard takeoff but he doesn't think that's the case he thinks it's going to be Progressive let's see what he has to say so first of all it's not going to be an event right the idea somehow which you know is popularized by science fiction and Hollywood that you know somehow somebody is going to discover the secret the secret to AI or human level AI or Ami whatever you want to call it and then you know turn on a machine and then we have a GI that's just not going to happen it's not going to be an event it's going to be gradual progress are we going to have systems that can learn from video how the world works and learn good World presentation yeah before we get them to the scale and performance that we observe in humans it's going to take quite a while are we going to get systems that can have large amount of associative memory so they can they can remember stuff 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 like you know to get this to work together with full system is another story how we going to have systems that can reason and plan perhaps along the lines of objective driven AI architectures that I I described before yeah but like before we get this to work you know properly it's going to take a while 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 representation systems that can be configured for a lot of different situation at hands the way the human brain can all of this is going to take you know at least a decade and probably much more because there are 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 a easy solution within this framework work you know it's it's not just around the corner I mean I've I've been hearing people for the last 12 15 years claiming that you know edgi is just around the corner and being systematically wrong and I knew they were wrong when they were saying it I call their and next he talks about AI to see the wide spectrum of people's beliefs about AGI and AI Doom 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 super intelligence is going to be an event that at some point we're going to have we're going to figure out the secret and we'll turn on a machine that is super intelligent and because we' never done it before it's going 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 has all the have all the characteristics of you know human level intelligence but their level of intelligence would be like a cat or a pirrot maybe or something and then we're going to work our way up to kind of make those things more intelligent and as we make them more intelligent we're also going to put some guard rails in them and learn how to kind of put some guard rails 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 and some of them are going to succeed at making intelligent systems that are uh controllable and safe and have the right guard rails and if some other goes Rogue then we can use the the good ones to go against the Rogue ones uh so it's going to be my you know smart AI police against your Rogue AI it's not going to be like you know we're going to be exposed to like a single Rogue AI that's going to kill us all that's just not not happening now there is another fallacy which is the fact that because the system is intelligent it necessarily wants to take over and there is several arguments that make people scar of this which I I think are completely false as well so one of them is you know in nature it seems to be that the more intelligent species other one that end up dominating the other distinguishing the others sometimes by Design sometimes just by mistake and so you know there is sort of uh 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 for a number of reason um first reason is they're not going to be a species they're not going to be a species that competes with us 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 orang Tong the species in which this desire to dominate or submit or or attain status in other ways is is specific to social species non-social species like our Tong don't have it right and they are as smart as we are almost and then next they tackle the comparison between AI AGI and nuclear systems and there's been a lot of comparisons between nuclear and Ai and I don't necessarily agree that's an Apples to Apples comparison and Yan tends to agree and he explains why but not only that he explains why AI is essentially going to be our filter to the rest of the internet and to other AI models an AI model is not going to be able to directly communicate with us we're not even going to see it he actually brings brings up the analogy of a spam filter for email we get a ton of spam and most of which we never see and we never need to see so it's interesting to think about but it's also a little worrying to think that an AI is going to be my filter to all of my information diet so that AI system designed by Vladimir Putin or whatever or his uh minions you know is going to be uh talking to trying to talk to every American to uh convince them to vote for you know whoever whoever is Puttin or ride people up against each other they're not going to be talking to you they're going to be talking to your AI assistant which is going to be as smart as their that 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 tell me TR like it's not even going to be able to get to you because it's only going to talk to your a assistant and your a assistant is not not even going to it's going to be like a filter 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 and next he talks about robots and I've been making a lot of videos about robots lately so let's see what Yan has to say about robots like the next decade I think is going to be really interesting in robots like the the emergence of the robotics industry has been in the waiting for 10 20 years without really emerging other than for like you know kind of pre-program behavior and stuff like that and the main issue is again the Maric Paradox like you know how do we get those system to understand how the world works and and kind of you know plan actions and so we can do it for really specialized tasks the way Bon Dynamics goes about it is you know basically with a lot of um handcrafted dynamical models and careful planning uh in advance which is very classical robotics with a lot of innovation a little bit of perception but it's still not like they can't build a domestic robot we're still some distance away from completely autonomous level five driving and we're certainly very far away from having uh you know level five autonomous driving by A system that can train Itself by driving 20 hours like any 17-year-old 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 and last Lex asks a good question what gives you hope about Humanity over the next few decades and Yan gives some good answers there I really love hearing this I love ending it on of a positive note we can make Humanity Smarter with AI 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 uh tasks in ways that are much better than we could do ourselves because they'll be smarter than us and so it's like everyone would be the the boss of a staff of super smart virtual people so we shouldn't feel threatened by 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 that's actually a wonderful thing so uh having machines that are smarter than us that assist us in our all of our tasks our daily lives whether it's professional or personal I think would be absolutely wonderful thing because intelligence is the most is the commodity that is most in demand that's really what I mean all the mistakes are it makes is because of lack of intelligence really or lack of knowledge which is you know related making people smarter which 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 and even social networks are a good thing if you run them properly it's difficult but you know you can helps the communication of information and knowledge and the transmission of knowledge so AI is going to make Humanity smarter so that's it I encourage you to check out the video in full it is very long hopefully this cut down version helped and if you like this video please consider giving a like And subscribe and I'll see you in the next one
- Generated with https://kome.ai
- Assistant
It appears you have provided a comprehensive summary and analysis of an extensive and in-depth interview between Yan LeCun, the head of Meta's AI division, and Lex Fridman. The discussion covered a wide array of topics related to artificial intelligence, including the limitations of autonomous regressive language models (like OpenAI's ChatGPT) in achieving AGI (Artificial General Intelligence), the necessity of open source in AI development, biases within AI systems, the economics of open-source AI, and the future of AI in terms of robots, AGI, and its impact on humanity.
Yan LeCun emphasized several key points throughout the interview:
-
Current Limitations of AI: LeCun argued that current AI technologies, particularly large language models, lack essential capabilities required for AGI, such as understanding the physical world, persistent memory, reasoning, and planning.
-
Data and Synthetic Data: He mentioned that the amount of data required to train AI models is immense, and synthetic data generated by AI could be necessary to reach AGI due to the limitations in human-generated data.
-
Language as a World Model: LeCun expressed skepticism that language alone could serve as a sufficient model for understanding the world, highlighting the need for a more comprehensive approach to achieving AGI.
-
Open Source as the Solution: Throughout the interview, LeCun strongly advocated for open source as a crucial factor in democratizing AI, preventing biases, and ensuring a diverse range of AI applications and services.
-
AI Doomsday and AGI: LeCun dispelled common doomsday theories related to AI, suggesting that the emergence of AGI would be a gradual process, with significant time to address potential risks and instill safety measures.
-
Robots and AI's Future: He discussed the current state of robotics and the promising future of AI in making humanity smarter by amplifying human intelligence through AI assistants.
-
Economics of Open Source and AI Biases: LeCun addressed the economic models behind open-source AI contributions and the inherent challenges in eliminating biases from AI systems, reinforcing the importance of open-source models to create a balanced, unbiased AI ecosystem.
The interview provides deep insights into LeCun's perspectives on the challenges and opportunities in the AI field, his advocacy for open-source AI, and his optimistic views on the potential for AI to benefit humanity. It also underscores the complexity of issues surrounding AGI, including ethical considerations, the need for diverse data and models, and the importance of progressing cautiously and responsibly.
This summary encapsulates the main points discussed in the interview, providing a concise overview of LeCun's thoughts on the current state and future directions of AI research and development.
-