Meta's Chief AI Scientist Yann LeCun Makes the Case for Open Source

Kara sits down for a live interview with Yann LeCun, an “early AI prophet” and the brains behind the largest open-source large language model in the world. The two discuss the potential dangers that come with open-source models, the massive amounts of money pouring into AI research, and the pros and cons of AI regulation. They also dive into LeCun’s surprisingly spicy social media feeds — unlike a lot of tech employees who toe the HR line, Yann isn’t afraid to say what he thinks of Elon Musk or President-elect Donald Trump. This interview was recorded live at the Johns Hopkins University Bloomberg Center in Washington, DC as part of their Discovery Series. Questions? Comments? Email us at [email protected] or find us on Instagram and TikTok @onwithkaraswisher Learn more about your ad choices. Visit podcastchoices.com/adchoices

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:11Kara Swisher?

  Hi, everyone. From New York Magazine and the Vox Media Podcast Network, this is On with Kara Swisher, and I'm Kara Swisher. Today, we've got a special episode for you, my conversation with Jan LeCoq, chief AI scientist at Meta. This was recorded live as part of a series of interviews on the future of AI I'm conducting in collaboration with Johns Hopkins University Bloomberg Center. And Jan is really the perfect person for this. He's known as one of the godfathers of AI. Some even call him an early AI prophet. He's been pushing the idea that computers could develop skills using artificial neural networks since the 1980s, and that's the basis for many of today's most powerful AI systems. Jan joined what was then known as Facebook as director

• 0:53Kara Swisher?

  of AI research in 2013, and he currently oversees one of the best-funded AI research organizations anywhere. He's also been a longtime professor at New York University and received the 2018 Turing Award, which is often called the Nobel Prize of Computing, together with Geoffrey Hinton and Yoshua Bengio for their breakthroughs on deep neural networks that have become critical components of computing. Jan is also a firebrand. He's pretty outspoken politically. He's not a fan of President-elect Donald Trump or Elon Musk, and lets you know it on social media, and is also not without controversy in his own field. When others, including Hinton and Bengio, started warning about the potential dangers of unmitigated AI research and calling for government regulation,

• 1:38Kara Swisher?

  Jan called it BS. In fact, he said that regulating AI R&D would have apocalyptic consequences. I want to talk to him about this dispute. We'll also get into what Meta is doing in this space right now, where he sees the potential and risks for all the new generative AI agents coming on the market, and perhaps most importantly, how close we are to artificial general intelligence, or AGI.

• 2:12Speaker 01

  Black Friday is around the corner, your biggest chance to grow sales this year. Don't wait. Realize is Taboola's advertising platform for businesses, letting you launch campaigns on top premium websites beyond search and social. Add Realize to your media mix and scale this Q4. And as an exclusive offer for podcast listeners, new customers get $700 credit back. Just search Realize Podcast on Google or use the promo code PODCAST26. Don't miss out on this limited time offer. Terms apply.

• 2:51Kara Swisher?

  Welcome. Thank you for joining me here at the new Johns Hopkins University Bloomberg Center for this special live conversation. You're obviously known as one of the godfathers of AI because of your foundational work on neural networks. There's a few people like that been around, which is the basis for today's most powerful AI systems. For people who don't know, AI has been with us for a while. It's just reached a moment. You know, we're here with a new administration coming in. And I have to tell you, you are the most entertaining person on social media that's a wonk that I've ever met. You're also quite outspoken as a scientist, as a person, I think as a citizen is what you're talking about. And I promised the meta PR people that I wouldn't get them fired today. But

• 3:33Kara Swisher?

  you, you're an astonishing person. I just want to like, I'm going to read a few. And I want you to talk about why you do this. I don't see a lot of people in tech do this, except for Elon Musk. But you actually, I like. So you write, "Trump is a threat to democracy. Elon is his loudest advocate. You won't get me to stop fighting enemies of democracy." Elon didn't just buy Twitter, he bought a propaganda machine to influence how you think. Those are the nice ones. As I've said multiple times about Elon, I like his cars, his rockets, satellite network. I disagree with his stance on AI existential risk. I don't like his constant hype. I positively hate his newfound vengeful conspiracist paranoid

• 4:15Kara Swisher?

  far-right politics. I'm nicer to him than you are, and that's the thing. And you talk about this a lot. And you're, you've been pretty Not supportive of Donald Trump too, I'm not going to read them all, but they're tough, tougher than I've ever been. So I want to talk about that. You've gotten in open disputes with Elon, you've called President Trump a pathological liar, and Mark was just in Mar-a-Lago enjoying a lovely meal on the terrace there. Talk about your relationship with the upcoming administration and how you're gonna, are you gonna have to start to not do this, or do you give a fuck?

• 4:50Speaker 02

  Well, I mean, I worry about many things, but, or I'm interested in a lot of questions. I'm sort of, you know, politically, pretty clearly a classic liberal, right, which on the European political spectrum puts me right in the center, not in the US.

• 5:14Kara Swisher?

  No.

• 5:15Speaker 02

  And what got me riled up with Elon was when he started sort of attacking the institutions of higher learning, of science, and scientists like Anthony Fauci and things like this. And, you know, I'm a scientist, I'm a professor, as well as an executive at Meta. And I have a very independent voice, and I really appreciate the fact that at Meta, I can have an independent voice, you know, I'm not doing corporate speech, as you can tell. So that tells you something about, I think, how the company is run. And it's also reflected in the fact

• 5:58Speaker 02

  that in the research lab that I started at Meta, we publish everything we do, we distribute our code in open source, you know, we're sort of very open about things and about our opinions as well. So that's the, you know, that's the story, but...

• 6:13Kara Swisher?

  So now he's at the red hot center of things. How are you gonna cope with that going forward?

• 6:18Speaker 02

  Well, I mean, I met Elon a bunch of times, he's, you know, he can be reasonable. I mean, you have to work with people, right? Regardless of disagreements about, you know, political or philosophical opinions, at some point you have to work with people, and that's what's gonna happen between... I don't do policy at Meta. I'm a

• 6:37Kara Swisher?

  You don't.

• 6:38Speaker 02

  I work on fundamental research.

• 6:39Kara Swisher?

  Right.

• 6:39Speaker 02

  I don't do content policy. I don't do any of that. I talk to a lot of governments around the world, but mostly about AI technology and how it impacts, you know, their policies. But, but, you know, I'm not, I don't have any influence on, on sort of the relationship between Meta and the political system.

• 7:01Kara Swisher?

  And I'm curious why you then didn't, why you went to a place like Meta versus in the old days, you would have been at a big research university or somewhere else. How do you look at your power then? What is your influence? I mean, you're sort of saying, I'm just a simple scientist making, you know, making things.

• 7:17Speaker 02

  OK, I'm also an academic, right? I'm a professor at NYU. And I, I've kept my position at NYU. When Mark Zuckerberg approached me 11 years ago, almost to the day, he, you know, he asked me to kind of create basically a research lab in AI for, for Meta because he had this vision that this was going to have a big impact and a big importance. He was right. And I told him, I only have three conditions. I don't move from New York. I don't quit my job at NYU. And all the research that we're going to do, we're going to do it in the open. We're going to publish everything we do. And we're going to open source the code. And his answer was, yes, yes. And the third thing is, you don't have to worry about this. It's in the DNA of the company.

• 8:02Speaker 02

  We already open source all of our platform code. And so, yeah, there's no problem with that. And this is not an answer I would have had anywhere else that would have had the resources to create a research lab. And I had the opportunity there. I was given the opportunity basically to create a research organization in industry from scratch and basically shape it the way I thought was, was proper. I had some experience with this because I started my career at Bell Labs. So I had some experience with sort of, you know, how you do real ambitious research in industry. So I thought that was the most exciting challenge.

• 8:43Kara Swisher?

  So Trump recently named David Sacks as the AI and crypto czar. For those who don't know, Sacks is an investor and part of the PayPal mafia, also a longtime friend of Elon Musk. He's shifted his politics pretty dramatically. Talk about the, is there a need for that right now in Washington as someone who's doing this research? Or do you not care whatsoever? It doesn't matter to you? Is it important that the government do something like this?

• 9:09Speaker 02

  Oh, absolutely.

• 9:09Kara Swisher?

  And tell us why?

• 9:11Speaker 02

  Well, there's a number of different things. The first thing is to not make regulations that make open source AI platforms illegal, because I think it's essential to the, for the, the, the future of not just the progress of technology, but also the way people use them, like to, to make it widely disseminated and everything. So that's the first thing. The second thing is, and, and by the way, there is no problem regulating products that are based on AI. That's, that's perfectly fine. I'm not anti-regulation or anything. The second thing is the academic world is falling behind and has a high time contributing to, to the progress of AI because of the lack of computing resources.

• 9:54Kara Swisher?

  Correct.

• 9:54Speaker 02

  And so I think there should be resources allocated by the government to give computing resources to academics.

• 10:02Kara Swisher?

  To academics. Now, this is, as you said, it's shifted rather dramatically because academics is where a lot of the research, the early computing research was done, and now it's moved away from that. Andrew Ferguson has been tapped to head the FTC. Former Fox News anchor Pete Hegseth is nominated defense secretary. Ferguson seems to want to roll back any attempts to be a regulator. Is this important for government to be more active in this area?

• 10:27Speaker 02

  It's certainly important for the government to be more informed and educated about it. But.

• 10:36Speaker 02

  I mean, active certainly for the reasons that I said before, because there's probably, you know, an industrial policy to have all the all the chips that, you know, enable AI at the moment are all fabricated in Taiwan, designed by a single company. You know, there's probably something to do there to sort of maybe make the landscape a little more competitive or.

• 11:03Kara Swisher?

  For chips, for example.

• 11:04Speaker 02

  For chips, for example. And there's another question, I think, that's really crucial also, and that has consequences not just for the US government, but governments around the world, which is that, you know, AI is quickly going to become a kind of universal knowledge platform, basically, you know, the sort of repository of all human knowledge. But that can only happen with free and open source platforms that are trained on data from around the world. You can't do this within the walls of a single company on the west coast of the US. You can't have a system speak all 700 languages of India, or however many there are. So so eventually, those platforms will have to be trained in a distributed fashion with

• 11:49Speaker 02

  lots of contributors from around the world, and they will need to be open.

• 11:54Kara Swisher?

  So I know you worry about premature regulations, stifling innovation, but you signed an open letter to President Biden against his AI executive order. Talk about why you did that more broadly, what role you think the government should play exactly.

• 12:07Speaker 02

  So I think there were, you know, plenty of completely reasonable things in that executive order, similarly in the EU AI Act, like for protection of privacy and things like this, which make complete sense. What really sort of I disagreed with, both in the EU AI Act in its original form and in the executive order, is that there was a limit established where if you train a model with more than 10 to the 24th, 10 to the 25th flop, you have to basically get a license from the government. Or yet authorization of some kind, based again on the idea that AI is intrinsically dangerous, that you're above a certain level of sophistication is intrinsically dangerous.

• 12:52Speaker 02

  And I'm, I completely disagree with this, this, this approach. I mean, there are, you know, important questions about AI safeties, safety that need to be discussed, but a limit on computation just makes absolutely no sense.

• 13:05Kara Swisher?

  Makes no sense. Recently, many big tech companies rolled out either LLM updates or new AI agents or AI features. I want to get an overview of what you're doing at Meta right now. It's a little different. You released Llama 3.3 is the latest update that powers Meta. I talk about what it does, and I'm going to ask you to compare it to other models out there and be honest. Like, how good is it compared to, and how do you look at that?

• 13:32Speaker 02

  As scientists, I need to be honest.

• 13:34Kara Swisher?

  Yeah.

• 13:35Speaker 02

  I mean, the main difference between Llama and most of the other models is that it's, it's free and open.

• 13:41Kara Swisher?

  Right, open source.

• 13:42Speaker 02

  So technically, it's not.

• 13:43Kara Swisher?

  Explain to people who may not understand what that means.

• 13:47Speaker 02

  Okay, so open source software is software that comes to you with the source code. So you can modify it, compile it yourself, you can use it for free. And in most licenses, if you make some improvement to it and you want to use it in a product, you have to release your improvement as well in the form of source code. So that allows, you know, platform style software to progress really quickly. And it's been astonishingly successful as, as a way to distribute platform software over the years. The entire internet runs on open source software. Most computers in the world run on Linux.

• 14:28Speaker 02

  Almost all computers in the world run on Linux, in your car, in your, you know, Wi-Fi router. So that's incredibly successful, and the reason is it's a platform, people need to be able to modify it, make it safer, more secure, et cetera, make it run, make it run on various hardware. That, that's what happens. And it's, it's not because, it's not bad design, it's just the, the market forces naturally push the industry to pick open source platforms, open source code when it's, when it's a platform. Now, for AI, the question of whether something is open source is complicated, because when you build an AI system, first of all, you have to collect training data. Second, you have to train what's called a foundation model on that training data.

• 15:12Speaker 02

  Okay, and that the training code for that and the data generally is not distributed. So Meta, for example, does not distribute the training data nor the training code, or most of it, for its for for the Llama models, for example. Okay, then you can distribute the trained foundation model, and so that's what Llama is. It's it's and it it comes with open source code, which allows you to run the the system and also fine tune it anywhere you want. You don't have to pay Meta, you don't have to ask questions, you don't have to ask Meta, you can do this. There are some limits to this that are due to the legal landscape, essentially.

• 15:52Kara Swisher?

  So why? Why is that better? You make the argument that all the others are not. They're closed systems, they develop their own.

• 15:58Speaker 02

  There are a few other open.

• 16:00Kara Swisher?

  Right, but the big ones are.

• 16:01Speaker 02

  The big ones are are closed. Yeah, the ones from OpenAI, Anthropic and Google are closed.

• 16:08Kara Swisher?

  Why did they choose that from your perspective?

• 16:11Speaker 02

  Well, quite possibly to, you know, get a commercial advantage. Like, if you want to derive revenue directly from from a product of this type, and you think you are ahead technologically, or you think you can be ahead technologically,

• 16:29Speaker 02

  and and and your main source of revenue is going to come from those from those services, then then maybe there is an argument for keeping it closed. But this is not the case for Meta. You know, for Meta, AI AI tools are, you know, part of kind of a whole set of.

• 16:46Kara Swisher?

  Experiences.

• 16:47Speaker 02

  Of experiences, which are all funded by, you know, advertising, right?

• 16:51Kara Swisher?

  Right.

• 16:51Speaker 02

  And and so that's not the main source of of revenue. On the other hand, what we think is that the platform will progress faster. In fact, we've seen this with.

• 17:01Kara Swisher?

  Be more innovative, because it's.

• 17:03Speaker 02

  More innovative, like, there's a lot of innovations that we would not have had the idea of, or we didn't have the bandwidth to do, that people have done because they had the Llama system in their hands, and they were able to experiment with it and sort of come up with new ideas.

• 17:15Kara Swisher?

  So one of the criticisms is that you were behind, and this was your way to get ahead. Is that? How do you address that? Ever heard that from your competitors?

• 17:23Speaker 02

  So this is an interesting history to all of this, right? So first of all, you have to realize that everyone in the industry, except Google. To build AI system uses open source software platform called PyTorch, which is mostly developed, which was originally developed at Meta. Meta transferred the ownership of it to the Linux Foundation. So now it's not owned by Meta anymore. But you know, OpenAI, Anthropic, everybody uses PyTorch. So without Meta, there would not be ChatGPT and Claude and all of those things. Or not, not to this, you know, same extent that they are, they are today.

• 18:03Speaker 02

  There has been developments, the, the underlying techniques that are used in, in tools like, like ChatGPT were invented in various places. OpenAI made some contributions back when they were not secretive. Google certainly made some.

• 18:17Kara Swisher?

  I like how you just put that in there, when they were not secretive.

• 18:20Speaker 02

  When they were not secretive, because it became secretive, right? They, they kind of climbed up in the last, in the last three years or so. Google climbed up too, to some extent, not completely, but, but they did. And Anthropic has never been open. So they, they, they sort of tried to push the technology in secret. I think we are perhaps at Meta, we're a pretty large research organization. And we also have a, a applied research and advanced development organization called GenAI. The research organization is called FAIR, that used to mean Facebook AI Research. Now that means fundamental AI research.

• 18:53Speaker 02

  And it's about 500 people. And what we're working on is really sort of the, the next generation AI system. So beyond LLMs, beyond large language models, beyond chatbots. There was this idea by some people in the past that you, you take LLMs like the, you know, ChatGPT, Meta, AI, Gemini of the world, and you just scale them up, train them on more data with more compute and somehow sort of human level intelligence will emerge from it. And, and everybody believes in this concept.

• 19:23Kara Swisher?

  Right. We've reached the end and there's no more data.

• 19:26Speaker 02

  Right. And it's pretty clear that we are reaching a kind of a, a ceiling in the performance of those systems, because we basically run out of natural data, like all the text that's publicly available on the internet. Is currently being used to train all those LLMs. And we can get much more than that. So people are kind of generating synthetic data and things like this. But, you know, we're not going to improve this by a factor of 10 or 100, right? So it's hitting a saturation. And what we're working on is basically the next generation AI system that is not based on just predicting the next word. So an LLM is called a large language model because it's basically trained to just predict the next word in a text.

• 20:08Speaker 02

  Right, you collect typically something like 20 trillion words, something of that order. That's all the publicly available text on the internet with some filtering. And you train some gigantic neural net with, you know, billions or hundreds of billions of tunable parameters in it to just predict the next word. Given a sequence of a few thousand words, can you predict the next word that will occur? You can never do this exactly. But what those systems do is that they predict basically a probability distribution of words, so which you can use to then generate text. Now, there is no guarantee that whatever sequence of words is produced makes sense. You know, doesn't generate co-fabulations or make stuff up, right?

• 20:52Speaker 02

  So what a lot of the industry has been working on is basically fine-tuning those systems, training them with humans in the loop to train them to do particular tasks and not produce nonsense. And also to kind of interrogate a database or a search engine where they don't actually know the answer. And so you have to have systems that can actually detect whether they know the answer or not. And then perhaps generate multiple answers and then pick which ones are good. But ultimately, this is not how future AI systems will work.

• 21:24Kara Swisher?

  So talk about that. Last week Meta released Meta Motivo. It's made to make digital avatars that seem more lifelike, as I understand. I feel like it's Mark trying to bring the Metaverse and make it happen again. But talk about how it's, what it is. I don't quite understand it because there's a lot of money you're all investing in all these things.

• 21:45Speaker 02

  Yeah, a lot of money.

• 21:46Kara Swisher?

  To make something that people would want to buy, right? Not just to make better advertising. You've got to have a bigger goal than that.

• 21:52Speaker 02

  Okay, I'll let you in on the secret. I'm wearing smart glasses right now.

• 21:56Kara Swisher?

  Yes, I have a pair myself.

• 21:57Speaker 02

  It's got, right, it's pretty cool, right? It's got cameras.

• 21:59Kara Swisher?

  Yeah.

• 22:00Speaker 02

  If you smile, it can take a picture of you guys.

• 22:01Kara Swisher?

  Yeah, yeah. This is how far we've come.

• 22:06Kara Swisher?

  I had one of the first pairs of Google Glass, but it's a low bar from that, but go ahead.

• 22:11Speaker 02

  Now, here's the thing. Eventually, we'll be working around, you know, we're talking five, 10 years from now. We'll be working around with smart glasses, perhaps other smart devices. And they will have AI assistance in them. This one has one. I can talk to meta AI through this, right?

• 22:28Speaker 02

  And, you know, those things will be sort of assisting us, you know, in our daily lives. And we need those systems to have essentially human-like intelligence, human-level intelligence, or perhaps even superhuman intelligence in many ways.

• 22:42Speaker 02

  And now, you know, how do we get to that point? And we're very far from that point. Like, you know, some people are kind of making us believe that we're really close to what they call AGI, artificial general intelligence. We're actually very far from it. I mean, when I say very far, it's not centuries. It may not be decades, but it's several years.

• 23:02Speaker 02

  And the way you can tell is that

• 23:06Speaker 02

  the type of task, right? We have LLMs that can pass the bar exam or, you know, pass, you know, some college exam or whatever. But, you know, where is our domestic robot that, you know, cleans the house and clears up the dinner table and fills up the dishwasher? We don't have that. And it's not because we can't build the robots. We just cannot make them smart enough. We can't get them to understand the physical world. Turns out, the physical world is much harder for AI systems to understand that language. Language is simple. I mean, it's kind of counterintuitive for humans to think that, you know, we think language is the pinnacle of intelligence. It's actually simple because it's, you know, just a sequence of discrete symbols. We can handle that. The real world, we don't.

• 23:50Speaker 02

  So what we're working on basically are kind of new architectures, new systems that understand the physical world and learn to understand the physical world the way babies and young animals do it by basically observing the world and acting in it. And those systems will eventually be able to plan sequences of actions so as to fulfill a particular goal. And that's what we call agentic, right? So an agentic system is a system that can plan a sequence of actions to arrive at a particular result. Right now, the agentic systems that everybody talks about don't actually do this planning. They can cheat a little bit. They can learn templates of plans.

• 24:26Kara Swisher?

  Right, but they can't do this. You're also working on, the information is reported Meta is developing AI search engine. So, well, I assume you want to best Google search. Is that true? And do you think that's important?

• 24:39Speaker 02

  Well, a component of an intelligent assistant that you want to talk to

• 24:44Kara Swisher?

  Is search.

• 24:45Speaker 02

  Obviously is search. You want to search for facts, right? And link to the sources of that fact, so that the person you talk to kind of trusts the results. So, search engine is a component of an overall, you know, complete AI system.

• 25:00Kara Swisher?

  And an end run around the Google system, presumably.

• 25:03Speaker 02

  Well, I mean, the goal is not necessarily to compete with Google directly, but to, you know, serve people who want an AI system.

• 25:11Kara Swisher?

  So what do you imagine it's going to be for? Because most people perceive that Meta was lagging in the AI race, especially with all the hype around ChatGPT. But Mark Zuckerberg just said it had nearly 600 million monthly active users and on track to be the most used AI globally by the end of the year. It's very different from what people are doing on ChatGPT, which is a standalone app or with search. So what is it for, for you, besides to make advertising more efficient? I know Mark has talked about that, but from your perspective and Meta's perspective, what is it for for Meta? What does it mean for Meta?

• 25:47Speaker 02

  It is that vision of the future where everyone will have an AI assistant with them at all times. And it's going to completely, I mean, it's a new computing platform, right? I mean, before we used to call this a metaverse, but, I mean, those glasses eventually will have displays, you know, augmented reality displays. I mean, there's already demonstrations of this with the Orion project that was shown recently. We can build them cheap enough right now, so we can sell them yet, but eventually they'll be there. So that's the, it's that vision, that long-term vision.

• 26:21Kara Swisher?

  So to be our helper, our agent.

• 26:22Speaker 02

  It'll be our helper, our daily helper. I mean, it's like everyone will work around with, you know, a virtual assistant, which is like a human assistant, basically. Or eventually like a staff of really smart people, maybe smarter people than you, working for you.

• 26:34Kara Swisher?

  That's great, but right now Meta is forecasting to spend between 38 billion and 40 billion. Google says it's going to spend more than 51 billion it's spent this year. Analysts predict Microsoft's spend will come close to 90 billion.

• 26:49Kara Swisher?

  Too much spending. Mark Benioff recently told me it was a race to the bottom. Are you worried about being outspent? And to get me a smarter assistant doesn't seem to be a great business, but I don't know. I didn't take the job at Facebook when I was offered it in the early days, so don't ask me. But go ahead.

• 27:06Speaker 02

  Well, it's a long-term investment. I mean, you need the infrastructure to be able to run those AI assistants at a reasonable speed for a growing number of people. As you said, there is 600 million people using Meta AI right now. By the way, there's another interesting number. The open source engine Lama, on top of which Meta AI is built, but which is open source, has been downloaded 650 million times. That's an astonishing number. I don't know who are all these people, by the way, but that's an astonishing number. There are 85,000 projects that have been derived from Lama that are publicly available

• 27:48Speaker 02

  all open source, mostly in parts of the world. A lot of those projects are basically training Lama, for example, to speak a bunch of languages from Senegal or from India or.

• 28:02Kara Swisher?

  So you don't think this money is ill-spent?

• 28:04Speaker 02

  No, I don't think so, because there's going to be a very large population who will use those AI systems on a daily basis within a year or two and then growing. And then those systems are more useful if they're more powerful, and the more powerful they are, the more expensive they are computationally. So this investment is investment in infrastructure.

• 28:24Kara Swisher?

  In infrastructure, what's happening by private companies. Now, you said the concentration of proprietary AI models in the hands of just a few companies was a huge danger. Obviously, there's also been critics of the open source model. They worry about bad actors could use them to spread misinformation, cyber warfare, bioterrorism. Talk about the difference. Does Meta have a role in preventing that happening, given you're handing these tools, these powerful tools, in an open source method?

• 28:54Speaker 02

  Okay, so this was a huge debate.

• 28:56Kara Swisher?

  It was.

• 28:57Speaker 02

  You know, in the, you know, just fairly, until fairly recently, you know, the early 2023, when we started distributing Lama, the first Lama was not open source. You had to ask permission and you had to show that you were a researcher. And it's because, you know, the legal landscape was uncertain and we don't, we didn't know what people were going to do with it. So, so it wasn't open source. But then, all of us at Meta received a lot of requests from industry saying like, you have to open source the next version because this is going to create a whole industry. It's going to enable like a lot of, you know, startups and kind of new products and new, and new things. And so we had a big internal discussion for several months internally.

• 29:41Speaker 02

  A weekly discussion, two hours with 40 people from Mark Zuckerberg down. Okay, very serious discussions about this, about safety, about legal landscape, about all kinds of questions. And then at some point, the decision was made was by, by Mark to, to say, okay, we're going to open source Lama 2. Tell me how to do it. And that was done in kind of summer 2023. And since then, it's basically completely jumpstarted a whole industry.

• 30:09Kara Swisher?

  Why is it more safe than these proprietary models that are controlled by the companies?

• 30:14Speaker 02

  Because there are more eyeballs on it. And so there are more people kind of fine tuning them for all kinds of things. And so there was a question as to, you know, maybe a lot of badly intentioned people would put their hands on it and then we were, we'll use them for nefarious purposes.

• 30:28Kara Swisher?

  Well, Chinese researchers developed an AI model for military use with an older version of Ben's Lama model as a backbone.

• 30:33Speaker 02

  It's actually kind of a, a very kind of... Minor bad things and he could have used one of the many excellent open source Chinese models, that's really good, which is on par with the best. So the Chinese have good research, good engineers, they open source a lot of their own models. You know, this is not like-

• 30:58Kara Swisher?

  So you don't think that's Meta's responsibility? You put it out there, the tools, and then what people do with it?

• 31:03Speaker 02

  No, it is to some extent, of course. So there is a big effort in the Llama team, in the Gen AI organization, to red team all the systems that we put out so that we ensure that they, you know, are, at least when it comes out, are minimally toxic and things like that, right? And mostly safe. That's a really important effort, actually. We even initially gave Llama to, you know, a bunch of hackers at DevCon and sort of had, you know, asked them, like, try to do something bad with it. And the result is, we haven't been aware of anything really bad done with any of the

• 31:46Speaker 02

  models that we've been distributing over the last almost two years.

• 31:49Kara Swisher?

  Yet would be the word I would put that behind that.

• 31:51Speaker 02

  Well, yet, but you know, it would have happened already. I mean, there have been, you know, the public doesn't realize this because they think it just appeared with ChatGPT, but there have been LLMs, open source LLMs, available for many years before that. And I don't know if you remember this, but when OpenAI came up with GPT-2, they said, "Oh, we're not going to open source it because, you know, it's very dangerous, so people could do really bad things, you know, they could flood the internet with disinformation and blah, blah, blah, so we're not going to open source it." I made fun of them because, I mean, it was kind of ridiculous at the time, the capability of the system really was not that bad. And so, I mean, you have to-

• 32:35Speaker 02

  Except the fact that those things have been available for several years and nothing really bad has happened. There was some, you know, a bit of worry that people would use this for disinformation, you know, in the run up of the elections in the US and all kinds of things like this, you know, cyber attacks and things. None of that really has happened.

• 32:54Kara Swisher?

  It's still good to be worried about such things.

• 32:56Speaker 02

  Well, I mean, you have to be, you know, watchful.

• 32:59Kara Swisher?

  Right.

• 33:00Speaker 02

  And do what you can to prevent those things from happening.

• 33:03Kara Swisher?

  Right.

• 33:04Speaker 02

  But the point is, you know, you don't need any of those AI systems for disseminating disinformation, as Twitter has shown us.

• 33:09Kara Swisher?

  Okay, good there, good. I like how you get your little digs in. I'm watching it very carefully. You did an Elon one, the secretive drama queens of OpenAI, I got that. So you also get a lot of flak online recently for saying that cultural institutions, libraries, foundations should make their content available for training by free and open AI foundation models like Llama, presumably. You were responding to a new dataset that Harvard released, made up of over a million books, but those are public domain works, not works by living authors, artists, academics. Talk about the concerns and the flak you got about these AI models vacuuming up all of our cultural knowledge from the creators, writers, researchers, without getting any credit. I mean, internet companies are known for scraping.

• 33:53Kara Swisher?

  I think Walt called, I believe it was, it was Face, when it used to be called Facebook Rapacious Information Thieves, but he may have been talking about Google. So talk to me about that, the controversy that happened with that.

• 34:04Speaker 02

  Okay. Outside of all of those kind of legal questions, if you have this vision that, you know, AI is going to be the repository of all human knowledge, then all human knowledge has to be available to train those models, right? And most of it is either not digitized or digitized but not available publicly. And it's not necessarily copyrighted material. It could be, you know, the entire content of the French National Library, a lot of which is digitized but not available for, for, for, for, for training. So not, this is, I was not necessarily talking about copyrighted work in that, in that case. It's more like, you know, if you are in

• 34:44Speaker 02

  So I'm from, my family, my father's family is from Brittany, okay? The western part of France, right? The traditional language spoken there, which was spoken until my great-grandfather, was, is Breton. Breton is disappearing. There is something like 30,000 people speaking it on a daily basis, which is very small. If you want future LLMs to speak Breton, there needs to be enough training data in Breton. Where are you gonna get that? You're gonna have cultural nonprofits, you know, kind of collecting all those stuff that they have, maybe governments helping, things like that. And they're gonna say like, you know, "Use my data, like I want your system to speak Breton." Now, they may not want to just hand that data just like that to, you know,

• 35:28Speaker 02

  big companies on the west coast of the US. But a future that I envisioned, this is not company policy, right? This is my, uh, my view. Is that the best way to get to that level is by kind of training an AI system, a common AI system, repository of human knowledge, in a distributed fashion, so that there would be several data centers around the world using local data to contribute to training a global system. You don't have to copy the data.

• 36:00Kara Swisher?

  But who runs that global system?

• 36:03Speaker 02

  Who writes Linux? Okay?

• 36:06Kara Swisher?

  Right, so that should exist for all of humanity.

• 36:09Speaker 02

  Yeah, I mean, who pays for Wikipedia, right?

• 36:12Kara Swisher?

  I pay $7 a month, but go ahead.

• 36:14Speaker 02

  Good idea, or the Internet Archive, right?

• 36:16Kara Swisher?

  Yeah.

• 36:17Speaker 02

  So for Linux, in the case actually, Linux is mostly supported by employees of companies who tell them, you know, to actually distribute their contributions. And you can have kind of a similar system where, you know, everyone contributes to this kind of global model.

• 36:36Kara Swisher?

  That's AI for everybody else.

• 36:37Speaker 02

  Which is AI for AI, you know, L L M in the short term.

• 36:41Kara Swisher?

  For things that aren't necessarily monetizable.

• 36:42Speaker 02

  Yeah, well, you monetize on top of it, right? I mean, Linux, you don't pay for Linux. But if you buy a widget that, you know, runs Linux like an Android phone or a car that has, you know, Linux in its touchscreen, you pay for the widget that you buy. So it's going to be the same thing with AI.

• 36:59Kara Swisher?

  That people could do that.

• 37:00Speaker 02

  The basic foundation model is going to be open and free.

• 37:02Kara Swisher?

  It does feel like that it's a coalescing of small amount of powers running everything. It does at this point. And that that vision is a lovely one, but it's not occurring.

• 37:12Speaker 02

  Well, it's, in my opinion, is actually inevitable.

• 37:15Kara Swisher?

  You've been in a public debate, you like to debate, with other godfathers of AI. Your Turing Award co-winners, Geoffrey Hinton and I think it's Joshua Benjio.

• 37:24Speaker 02

  Yep.

• 37:25Kara Swisher?

  They've both been ringing alarm bells, warning about the potential dangers of AI, quite dramatically, I would say. They've called for stricter government regulation oversight, including R and D. You've called their warnings complete BS. I don't think you minced words there. Talk to me about why that's complete BS. And one of the things you disagreed was one of the first attempts at AI regulation here in the US, California bill SB 1047. Hinton and Benjio both endorsed it. You lobbied against it. You wrote, regulating R and D would have apocalyptic consequences on the AI system. Very dramatic of you, sir. You said the illusion of existential risk is being pushed by a handful of, quote, "delusional think tanks." These two aren't delusional, I don't believe.

• 38:09Kara Swisher?

  Hinton just won the Nobel Prize for his work. Talk about that in particular. And by the way, Governor Newsom vetoed the bill, but is working with people like Stanford professor Fei-Fei Li to overhaul it. Talk about why you called it complete BS. You're very strong on this one.

• 38:24Speaker 02

  Very vocal about that, yes. So Jeff and Joshua are both good friends. We've been friends for decades.

• 38:32Kara Swisher?

  Yeah.

• 38:32Speaker 02

  You know, I did my postdoc in 1987, 88 with Jeff Hinton. So we've known each other for a very long time, for 40 years now. Same with Yashra, I knew, I met him the first time, he was a master's student and I was a postdoc. So, and we've been kind of working together, we won this prize together because we work together at a sort of reviving interest in what we now call deep learning and which is a root of a lot of AI technology today. So, we agree on many things. We disagree on a few things, and that's one of them.

• 39:04Kara Swisher?

  The existential threat to the human race, yeah.

• 39:07Speaker 02

  So, exactly. So, Jeff.

• 39:11Kara Swisher?

  They're like, ah, no. They're like, oh, yeah, they're coming for us.

• 39:15Speaker 02

  I mean, Jeff believes that current LLMs have subjective experience. I completely disagree with this. I think he's completely wrong about that. We disagreed on technical things before. It was kind of less public, it was more kind of technical, but it's not the first time we disagree.

• 39:31Speaker 02

  I just think he's wrong. We're still good friends. Yashra comes from a slightly different point of view. He's more worried, he's worried a little bit about this, but he's more worried about bad people doing bad things with AI systems.

• 39:45Kara Swisher?

  Yeah, I'm with him.

• 39:46Speaker 02

  Developing like bioweapons or chemical weapons or things like this. I think frankly, this is, those dangers have been, you know, formulated for several years now and have been like incredibly inflated to the point of being kind of distorted so much that that really, they don't make, they don't make any sense.

• 40:06Kara Swisher?

  Delusional is the word you use.

• 40:07Speaker 02

  Well, I don't, I don't call them delusional. I call some of the other people who are more extreme and are pushing for, you know, regulation like SB 1047. Yes, delusional. I mean, some people will tell you in the face, you know, a year ago,

• 40:23Speaker 02

  you ask them like, how long is it going to take for AI to kill us all? And they say like five months. And obviously they were wrong.

• 40:30Kara Swisher?

  So, this is what you're talking about, it's over AGI, artificial general intelligence, and how close we are. I would like you to explain it for people. When they hear it, they think about the plot of Terminator, iRobot, or something like that. So, Hinton and Bengio think the timeline for AGI could be more like five years and that we are not prepared. You said several years, if not a decade.

• 40:53Kara Swisher?

  You know, if you're wrong, you're going to be real wrong when it does kill us. So, talk about why, you know, you'll be like, oh, we're not dead yet, and then we're dead. So, talk about why you're not worried.

• 41:04Speaker 02

  So, first of all, there's no question that at some point in the future, we're going to have AI systems that are smarter than us, okay? It's going to happen. Is it five years, ten years, twenty years? It's really hard to tell. In our kind of, or at least my personal vision of it, the earliest it could happen is about five years, six years. But probably more like ten and probably longer because it's probably harder than we think. And it's almost always harder than we think. There is this history over, you know, the several decades of AI of people sort of completely underestimating how hard it is. And again, you know, we don't have automatic robots. We don't have, you know, level five sort of in cars. We don't, there's a lot of things that we don't know how to do with AI systems today.

• 41:46Speaker 02

  And so until we figure out kind of a new set of techniques to get there, we're not even on a path towards human level intelligence. So, once we, you know, a few years on that, once we have kind of a blueprint

• 42:00Speaker 02

  and some kind of believable demonstration that we might have a path towards human level AI. I don't like to call it AGI because human intelligence is very specialized, actually. So we think we have general intelligence, we don't. So, once we have a blueprint, we're gonna have a good way to think about how to make it safe. It's kind of like, you know, if you kind of back, you know, backpedal to the 1920s and someone is telling you, you know, in a few decades, we're gonna be flying millions of people across the Atlantic at near the speed of sound. You know, and someone would say like, oh my God, are you gonna make this safe? The turbojet was not invented yet.

• 42:44Speaker 02

  How can you make turbojets safe if you haven't invented a turbojet? We are in this situation today. So, you know, making AI safe means designing those AI systems in ways in ways that are safe. But until we have a design, we're not gonna be able to make them safe. So the question, you know, makes no sense.

• 43:01Kara Swisher?

  You don't seem worried that AI would ever want to dominate humans. You said this, you've said that current AI is dumber than a house cat. Whether AI is sentient or not doesn't seem to matter if it feels real, right? And so what, how do you, if it's dumb, or it doesn't want to dominate us, or it doesn't want to kill us, what would be restrictions on AI and maybe AI R and D that you would seem reasonable, if any? I think if none is what you're saying to me.

• 43:28Speaker 02

  Well, none on R and D. I mean, clearly, if you want to, you know, put out a domestic robot, and that robot can cook for you, you probably want to like hardwire some rules so that when there is people around the robot, and the robot has a knife in his hand, you know, he's not going to shut his arm around or something.

• 43:44Kara Swisher?

  Right.

• 43:45Speaker 02

  So, you know, kind of, you know, those are guardrails. So, the current, the design of current AI systems, to some extent, is intricately unsafe. You could say it this way. A lot of people, I admit, are going to hate me for saying this, but, but they're kind of hard to control. You, you basically have to train them to behave properly. What you, what you want, and this is something I've proposed, is another type of architecture which I call objective-driven, where the AI system basically is there to fulfill an objective and cannot do anything but fulfill this objective, subject to a number of guardrails which are just other objectives. And that will guarantee that whatever output the system produces,

• 44:30Speaker 02

  whatever action it takes, satisfy those guardrails and objectives and are safe. Now, the next question is, how do we design those objectives? And a lot of people are saying, "Oh, we've never done this before. This is completely new. We're going to have to kind of invent a new science." No, actually, we're pretty familiar with this. It's called making laws. We do this with people. We, we establish laws, and the laws basically change the cost of taking actions.

• 44:55Kara Swisher?

  Right.

• 44:56Speaker 02

  And so we've been shaping the behavior of people by making laws. And we're going to do the same for AI systems. The difference is that people can choose to not Respect the law, whereas the AI system by construction will have to.

• 45:11Kara Swisher?

  Now, both these people, Hinton and Benjou, endorsed a letter signed by current and former OpenAI employees, calling employees at AI companies have the right to warn about serious risks by the technologies and ordinary whistleblowers wouldn't protect them. You didn't endorse it. At the same time, we've seen some regulation in the EU. They differentiate between high-risk AI systems and more general-purpose models. They have bans on certain applications that, quote, "threaten citizens' rights." Facial images, I suppose this robot who wants to knife you. What is the model here to make it safer, to make people... You're suggesting we wait and see when bad things happen before putting up guardrails. Let's wait till there's some murder happening, or not. I can't tell.

• 45:54Speaker 02

  No, that's not what I'm suggesting. I mean, you know, measures like banning massive face recognition in public places, that's a good thing. Like, you know, nobody would really think that's a bad thing, except if you are an authoritarian government.

• 46:09Kara Swisher?

  Some people think it's a great thing.

• 46:11Speaker 02

  Yeah, it already exists in some countries, actually. But that, you know, that's a good thing, right? So, and there are measures like this that make complete sense, but they are at the product level. You know, also, like, you know, changing the face of someone on, you know, some embarrassing video and stuff like that. I mean, it's kind of already illegal, more or less. The fact that we have the tools to do it doesn't make it less illegal. There may be a need for, like, specific rules against that, but I have no problem with that. You know, I have a problem with this idea that, you know, AI is intrinsically dangerous and you need to regulate R and D. And the reason I think it's counterproductive is, you know, in a future in which you would have those open-source platforms I was talking

• 46:56Speaker 02

  about, which I think are necessary for things like democracy in the future, then those rules would be counterproductive. They would, you know, basically make open source too risky for any company to distribute, and so would...

• 47:12Kara Swisher?

  So that these private companies will control everything.

• 47:14Speaker 02

  That's right. A small number of private companies on the west coast of the US would control everything. Now, talk to any government outside the US, and tell them about this future where everyone's digital diet would be mediated by AI assistance.

• 47:30Speaker 02

  And tell them that this will come from three companies on the west coast of the US. And they say like, that's completely unacceptable.

• 47:35Kara Swisher?

  Right.

• 47:35Speaker 02

  Like, this is the death of our democracy. Like, how will people get a diversity of opinions? Right? If it all comes from three companies on the west coast of the US? We'll all have the same culture, we'll all speak the same language. Like, this is completely unacceptable. So what they want are open platforms that then can be fine tuned for any culture, value system, center of interest, whatever, so that users around the world have a choice. They don't have to use like three assistants, they can use, you know, the...

• 48:08Kara Swisher?

  So you're worried about domination by OpenAI, Microsoft, Google, possibly Amazon?

• 48:14Speaker 02

  Anthropic.

• 48:15Kara Swisher?

  Anthropic, which is Amazon, really. So, last two questions. You were awarded the 2024 VIN Future Prize. There's so many prizes in your area. I never get any prizes. For transformational contributions to deep learning. In your acceptance speech, you said, "AI does not learn like humans or animals", which take in a massive amount of visual observation from the physical world. But you've been working to make this happen. You've been talking about it a while. Where do you imagine it being in years? Will it be like humans or animals? Or where?

• 48:45Speaker 02

  Well, so yeah, I mean, there is a point at which we're going to have systems that learn a little bit like humans and animals and can learn new skills and new tasks as efficiently as humans and animals, which is frankly astonishingly fast. Like, we can't reproduce this with machines. Right? We have, you know, companies like Tesla and others have hundreds of thousands or millions of hours of cars being driven by people. They could use this to train AI systems, which they do. They're still not as good as humans. We don't have, yeah, we can't buy a car that actually drives itself or a robot taxi unless we cheat. Like, Waymo can do it, but it's a lot of tricks to it. And again, we can't buy, you know, a domestic robot because we can't make them smart enough. The reason for this is very simple.

• 49:31Speaker 02

  As I said before, we train LLMs and chatbots on all the publicly available text and so more. That's about 20 trillion words, right? A four-year-old has seen essentially the same amount of data visually than the biggest LLM has seen through text. That text would take any of us several hundred thousand years to read through. Okay, so what it tells you is that we're never going to get to a human-level AI by just training on text. We have to train on sensory input, which is basically an unlimited supply. 16,000 hours of video is 30 minutes of YouTube uploads. Okay, we have way more video data than we can, than we know what to do with.

• 50:16Speaker 02

  So the big challenge for the next few years in AI to make progress to the next level is get systems to understand how the world works by basically watching the world go by, watching video, and then interacting in the world.

• 50:30Speaker 02

  And this is not solved, but you know, there's a good chance that progress will be made, like significant progress will be made over the next five years, which is why you see all of those companies starting to build humanoid robots. They can't make them smart enough yet, but they're counting on the fact that AI is going to make sufficient progress over the next five years, that by the time that those things can be sold in the public, that they, you know, the AI will be powerful enough.

• 50:53Kara Swisher?

  Right, now I'm getting the glasses. I understand what you're up to now, finally. I actually believe in a four-year-old more than I believe in most of Silicon Valley, I'll be honest with you. I met people like you, as I was saying, this is my very last question, I'm very quick, so we've got to go. Who are like this, it's going to change learning, it's going to change this, it's going to make everything better, everyone's going to get along. And as you cite all the time, and I respect you for that, is there's hate, there's dysfunction, there's loneliness, self-esteem among girls, danger to people who are often in danger, controlled by billionaires of our government. Why do I trust you this time?

• 51:33Speaker 02

  Me?

• 51:34Kara Swisher?

  You, just you.

• 51:36Speaker 02

  Okay, I'm not a billionaire.

• 51:38Kara Swisher?

  What?

• 51:39Speaker 02

  I'm not a billionaire, that's the first thing.

• 51:43Speaker 02

  I'm doing okay though.

• 51:45Kara Swisher?

  I'm guessing you are.

• 51:49Speaker 02

  Okay, I'm first and foremost a scientist.

• 51:54Speaker 02

  And I would not sort of, you know,

• 51:59Speaker 02

  be able to look at myself in the mirror unless I had some level of integrity, scientific integrity at least. I might be wrong. So you can trust that I'm not lying to you and that I'm not, you know, motivated by nefarious motives like, you know, greed or something like this.

• 52:19Speaker 02

  But I might be wrong. I might very well be wrong. In fact, that's kind of the whole process of science is that you have to accept the fact that you might be wrong and the, you know, elaborating the correct ideas comes from the collision of multiple ideas and people who disagree. So, but like, you know, look at the evidence. So

• 52:43Speaker 02

  we look at the evidence from, you know, the people who said that AI was going to destroy society because we're going to be inundated with disinformation or generated hate speech or things like this. We're just not seeing this at all. We're not seeing it. We've not seen it. I mean, people produce hate speech. People produce disinformation and they try to disseminate it or, you know, every way they can. A lot of people are trying to disseminate hate speech on Facebook and it's against the content policy at Facebook to do this. Now, the best protection we have against this is AI systems. We couldn't do this in 2017, for example. 2017 AI technology

• 53:23Speaker 02

  was not good enough to allow Facebook and Instagram to detect hate speech in every language in the world.

• 53:31Speaker 02

  And what happened in between is progress in AI. Okay, so AI is not the tool that people use to, you know, produce hate speech or disinformation or whatever. It's actually the best countermeasure against it. So what you need is just more powerful AI in the hands of the good guys than in the hands of the bad guys.

• 53:50Kara Swisher?

  I'm worried about the bad guys, but that's a great answer. Thank you so much. I really appreciate it.

• 54:04Kara Swisher?

  On with Kara Swisher is produced by Christian Castro-Rosell, Kateri Yokum, Jolie Myers, Megan Burney, and Kalen Lynch. Mashat Kirwan is Vox Media's executive producer of audio. Special thanks to Corinne Ruff and Kate Furby. Our engineers are Rick Kwan, Fernando Aruda, and Aaliyah Jackson. And our theme music is by Trackademics. If you're already following the show, you get a free pair of Meta glasses. If not, watch out for that stabby robot. Go wherever you listen to podcasts, search for On with Kara Swisher, and hit follow. Thanks for listening to On with Kara Swisher from New York Magazine, the Vox Media Podcast Network, and us. We'll be back on Monday with more.