Dentro de la era de IA de Mark Zuckerberg | The Circuit

Do you know the Sake Renaissance? Sake Renaissance. Don't you know it? No. The Renaissance of Sake. I understand. And how is it? It's that you have a new style with the chains and jackets. Everything you do becomes a meme. At least they're good. It seems that you accept jokes a little more. I think I'm more mature and I care a little less. It's not that important. I just try to be myself while I feel comfortable. Mark Zuckerberg, or rather the face of Facebook, is famous for 20 years. On the offensive, he bought Instagram, WhatsApp and Oculus. And on the defensive, he constantly faces millions of criticism about the social, political and business effects of his empire in expansion. However, Zuckerberg seems to be reinventing himself and his company. Meta's latest move is a bet on an open source AI in contrast to closed-source competitors such as OpenAI and Google. Zuckerberg predicts that his new AI models will have a profound impact on the progress of technology and business. And perhaps in the world. I've come to the Meta headquarters to meet the real Sake. And the future of the AI that he believes is the path to follow. If we go back to 2003, an article of yours came out on Harvard Crimson that I found. Yes. You talked about open source. Yes. It was a long time ago. Really? You've been thinking about it for a long time. Yes, and it is a large sector of the technology industry. Without open source, he would not have been able to develop the initial version of Facebook. He was a student and did not have access to much capital. There are certain types of hacker mentalities. For example, what they do is take the code and use it for whatever they need, which is more efficient in costs. And it is possible to start something like that in just a student residence. Some people see you as an unlikely promoter of open source. You're laughing. Honestly, I don't know why. Or actually, yes, I do. I don't understand. You understand the unlikely. Yes. Actually, at Meta we are great defenders of open source. And we have been proposing it for a long time. And it has been an excellent formula for us. Let's go back to the past. To the way we designed the servers and the way we designed our data centers. This open source computing project has standardized a large part of the infrastructure. I think it will also happen with the AI. I think that by making your open source AI, you are marking a position in what would be an attempt to build the bases of the AI that we will see in the future. How much of this is part of a strategic way to control or own the next technology? Much of our growth in the last 10 or 15 years was due to the fact that we developed our applications through platforms controlled by our competitors. It is quite discouraging to develop something that you think will be good. And then a company like Apple suddenly tells you, you can't launch it because they want to lock you up. Since it considers you a competition. We are already a sufficiently large company and we have solved aspects of the next generation. I would like us not only to develop the next platforms, but also to have greater control over them. For example, I believe that artificial intelligence is a fundamental piece. And I think that augmented and virtual reality is another. The way to be able to control our own destiny in this and ensure that we have access to the most advanced AI is through its development. And making it become a standard in the industry. But in reality it is strengthened by being able to share it and through the ecosystem that surrounds it. You continue to improve the AI of Meta in all your products. But also as an independent chatbot. Why use Meta instead of chatGPT? It has many functions in which it is better. And one of our novelties is that soon we will launch the ability to imagine things. And you only need to write something in real time. I always try it with my daughters. And while a person writes or enters the query, the images are generated. Then we see the images at the same time we press the keyboard. It's really amazing. I had a goal. And this was that by the end of 2024, Meta AI would become the most used assistant in the world. And that was the goal I had at the time of launching Meta AI. A few months ago now. And I think we are about to get it. There are many people using it. For the end of the year? I think we will achieve it before the end of the year. And you will launch Yama 3.1, the set of models, including the largest open source model with 405 billion parameters. What does that improvement allow? I really think the main thing people will do. Since it is open source, it is to use it as a master to train smaller models for different applications. Probably all emerging companies, other companies, or even governments that try to do different things, need to develop personalized models for whatever they are doing. And it is very difficult to do that with the closed systems available, such as OpenAI or Gemini or some Google function. And this comes to a central part of our philosophy. And it is that we do not believe that there will be a single artificial intelligence that dominates everyone else. Our vision is that there will be millions, or even thousands of millions of different models on the market. There will be many gods. Do you think so? I don't think they are gods. But to a certain extent, I think that if an organization thinks it will create an unrepeatable superintelligence, then I do feel that its intentions are to create something like a god. And for this approach I have two feelings. Not only does it seem erroneous to me, but also not very striking. There are about 200 million influencers on our platforms. They try to build their community and people want to interact all day. So I want to do something so that the influencers train a version of them easily and that they adapt it as they wish. The result would be almost like a kind of artistic artifact that they then present to their community virtually, which allows them to interact with all their followers. But the benefit is that they have control over how that interaction happens. I think it would be great and that there could be millions or eventually hundreds of millions of models. So let's look at the general strategy and how it will work. To understand better. So will we have influencers generated by AI with subtitles and avatars generated by AI talking to avatars? I think we will have all that. Will you create the first social network generated by AI? I think it will be part of the process. But one of the most interesting use cases of meta AI today, which is among the four main, is the simulation of difficult social interactions that people will have, either in a professional context such as asking one of their bosses for a raise, or in a private context such as having a difficult conversation with your partner or some of your friends. I could have simulated this conversation with you today. So I hope it's not stressful for you. It's more than good. But it's a very good social tool where no judgments are made. The AI does not criticize and there are no social repercussions for the questions asked. But what evidence do you have that people want to live in this virtual world of avatars? Or that it's good? I think people are looking to connect. To be honest, I consider that all the other issues end up being left over. But technology allows more optimization and efficiency. At first Facebook was text and not much more. And then there was a kind of technological evolution, in which smartphones appeared and we used them mainly to take pictures. And then mobile networks experienced a significant improvement. Then we could share videos that transmit much more and watching them is more entertaining. I don't think it's the final stretch. I think it will be more immersive. You renamed your company as Meta and you still invest a lot in it. We are as advanced as you thought we would be a few years ago. Is there any lesson about the urgency to make the change? Much of the reason I did it, is because in reality I think we were getting stuck and we were considered just a company of social media applications. In reality, I was thinking that the metaverse is something very long-term. I think some things have come out better than I thought. Some have taken longer than I thought. I think glasses are the best example of something that has come out better than I expected. They have elegance and good quality. They have the ideal format for the IA. We didn't know that the IA would be important when starting that project. We thought it would matter later. But if you had asked me five years ago, I would have said that augmented reality would arrive before the IA and not the other way around. So a part of this is a bit of preparing to surf the waves when they arrive. Sequoia calls the IA the question of the $600 billion for all the investment that must be made. But when does it start to bear fruit? Is it a bubble? And if it's not, when do you start to see the money? I think bubbles are interesting because many ended up being very valuable. And it happened over time. So it's a matter of time as you ask. Even in the bubble of the .com, all that fiber was placed and ended up being very valuable. But it wasn't as fast as people thought it would be. The reality is that I don't know if all this will also happen here since it is difficult to predict. But I think the IA will be fundamental. I think there is a possibility that many companies are investing too much and that later we look back and think, maybe we spent a few billion dollars more than necessary. But on the other hand, I also think that the companies that invest are making a rational decision because the drawback of staying behind is that it is out of position for the most important technology of the next 10 to 15 years. You said that your goal is to reach the General Artificial Intelligence or IAG. How would you define the IAG? Do you think you will achieve it first? That's a good question. We have already started working on the development of Llama 4 and our goal is to completely close the gap with the other models. So I don't know. I'm not sure if we will achieve the IAG first. I think there will be advances soon. It is difficult to just predict it linearly. And then the complex question is to define what it is. To be honest, I don't know that there is a single specific definition for this. I believe that intelligence is multivariate. It is not as if there were a number that easily defines your intelligence. There is probably a modality or a specific aspect of intelligence, such as deciphering the faces and emotions of people. And that is something that interests me. So I think we will try to integrate it later. What we are trying to do is allow the whole community to create all these different artificial intelligences for everything they want to do. And that's how I think it ends up being something good for the world. It's very peaceful. Yes, we enjoy nature a lot. Is this like your escape oasis from the real or virtual world? From the real world. And from your virtual world? I don't know. I really like being up high, and also in the surroundings. But it's just nice to spend time here during the summer, take the kids to the lake and teach them different sports. I think it's fine. I think the most important thing is simply to learn to think critically and also to learn values since we were little. One of our daughters is very creative. I say that because she wrote a scheme for a novel that she was interested in writing. And now her novel already has 40 pages. And we spend time together, and something we usually do together is use the meta-IA to generate the images. Her novel is about mermaid crystals. I know kids who would read it. Yes, I think it's very possible that it will be very successful. Every time I want to hire new people, I am guided by the following philosophy. If people have shown that they can go deeper and do something really well, then they have probably gained experience in the art of learning something and raising it to the next level, which is generally useful for other things. It seems like you do everything 100%. How do you get the time? Because you are an executive director, a father, and you also do sports and other activities. I really think that achieving a balance between the different things you do helps you do them all better. I think I'm a pretty intense person. Yes. I try to have compassion with the people I work with and not concentrate too much energy on a single project because I want to make sure they have the flexibility and space to do their thing. So yes, I think those aspects are part of being an interesting and complete person and learning is similar to wanting to have different experiences. I'm very happy to be here in Beijing. I know you've always been fascinated by China and I also know that you learned the Mandarin language. What do you know about where China is in terms of the IAEA and the IAG? Personally, I don't know much. There is a question that arises, and it is how the United States should approach the competition of the IAEA against China. There is a current of thought that maintains that we must close everything. I think that is completely wrong, since the United States is really thriving with an innovation that is not only open, but also decentralized. This is how our economy works. This is how we create incredible things. But I think the leading companies should collaborate with the government of the United States and they should do so with the aim of ensuring that our national and similar defense maintain a continuous advantage as pioneers in the most advanced technology worldwide. Is this how we win the IAEA wars? I think the real question is what you expect to achieve. There is an aspect that I do not consider a reasonable objective, and that is if the question is whether the United States should be five or ten years more advanced than China. I'm not sure that can be maintained, but I do believe that there is a reasonable objective and it is to maintain a constant advantage of between six to eight months. And this is achieved through a guarantee, by ensuring that American companies and Americans who work in this continue to develop the best IAEA systems. And I think that if the United States manages to maintain that advantage over time, this is already a giant advantage. There are many participants. What do you think of Sam Altman's leadership? I think Sam really deserves a wide and well-deserved recognition, because of the way that organization has developed. I can give my opinion after having received a lot of public attention myself, and that is that when it happens to you for the first time, you can't handle it as well as you would like, but Sam is handling it with a lot of grace. I think he's handling it better than I could, and it's a little ironic to have an organization called OpenAI, but that somehow is a leader in building closed-code IAEA models, and it's not necessarily a bad thing, but it's a little funny. Facebook and you have been blamed for many things. Do you agree or not? Should we trust you? I think that's a trendy question. We have been blamed for many things. I take our performance very seriously, and I think we try to handle this in the best possible way. I'm not sure it was fair, but I like to think that we are a relevant company, so scrutiny is generally good. One of the defining things about open code is that anyone can examine the work, and that's why I think it simply puts a lot of pressure to make sure that the quality of the work you're doing improves quickly. Let's talk about the presidential elections. Yes. Facebook has been a critical point in many elections around the world, and you have also been pointed out, recently by former President Trump. They are important elections. What's at stake? Well, clearly they are elections that are very important, and without a doubt they will be historic elections. What I hear the most from people is that they want to see less political content in our services. And it makes sense, because they use our services to connect with other people. So that's exactly what we'll do. We'll give people control over this. But in general, we're trying to recommend less political content. I think our services will have less impact in these elections, unlike in the past. At the time, I did things that I don't plan to repeat this time. And when I say that, I don't mean to support any of the candidates. A lot of crazy things are happening in the world. And by that I mean that we've seen Donald Trump get up after being shot in the face, and then we've seen him raise his fist with the American flag. Both things are the most impressive things I've seen in my entire life, without similar background. And I think that at a certain level, as an American resident, it's quite difficult not to get excited. It's difficult not to get excited about that spirit and that fight. And I think that's the reason why a lot of people like this guy who's running for president. But I think our role here is to make it easier for everyone to have the ability to express their opinions and points of view on these issues. And I think that's the best role we can play. We're facing a mental health crisis, especially in adolescents. The Surgeon General demands a warning on social media Is he right? There's clearly a problem with mental health in the country. Mental health is a very important issue, and it's especially important for children and adolescents. And I think the approach they want to take is the right one. I have three little daughters, and right now I think that being a father is really difficult. Everything matters a lot to you, and my only wish for them is to make sure that they can live a good life. So from that perspective, I'm looking for us to develop our services in a way that is aligned with the parents. And in this way, we provide them with the controls they need to basically supervise how the services work for their children. What is shown through the data today differs a lot from what the basic meme that circulates through the networks reflects. I think a lot of people act as if there was a proven connection between the aspects. And the reality is that I don't think science supports that today. It's also likely that phones represent a problem. For example, if you get a notification or the phone vibrates, and somehow this prevents you from sleeping, that's not related to social media. But there are clearly problems in society, and we want to make sure we're part of the solution. You do your best to make us more virtual, but you love the real world. Yes. How do you fight those two ambitions? I don't really know if they clash. I think several people in the technology community think they should separate their consciousness from their intelligence and put them in the cloud. I think it's ridiculous. I think it's a little more than that. And I think what makes us people is that we are active and that we have energy. We are not just minds. Minds that think and reflect. I believe that aspects such as energy, love, and other of our emotions are in a way fundamental to define what makes us the people we really are. I'll learn wake surfing from you. I've heard you surf and you're from Hawaii. No, I'm from Hawaii and I've surfed. Great, but I'm not a surfer. Good, you surf. Yes, Priscila is pretty good. I'm glad Priscila is here because I've heard she puts you in your place. Yes, sometimes she does. Yes, she's very good. What's your best dad joke? When Mark makes a mistake with the girls, he just says, it wasn't my best job, it wasn't my best job. It's just something he says. When I braid Max's hair. You know? Yes, I braid Max's hair. And I usually say, I think I did it very well today. Or sometimes I say, this wasn't my best job. Now the girls braid his hair and tell him it wasn't his best job. How has it been to be in a relationship with all this, all this time? I think being in nature and doing sports is super important, especially for Mark. And this is because Mark has too much energy. He needs to spend a lot of energy every day. And how do you keep up with him? I don't follow him. Okay, do your thing. Mark is supposed to teach me. Let's do it. Yes, my success really depends on how good a teacher I am. Surf waves come from behind. These waves will come from the side. The force pushes you in that direction, and that way you can lean on the wave much more than you would lean on a normal surf wave. Don't worry. Yes, you'll be fine. Your body will know how to move. That's our family sports motto. Your body will know how to move. I'm ready to go. Let's rock it. Ready. Aloha. Let the board turn. Get up. It's so close. You'll make it. This is the part where they ask if this is really fun. You're almost there. Almost. Let's go for it. I'll hear you. Have you driven differently for her than for us? No. Your body knows how to move, Mark. Yes, my body knows how to move. Let's go. It seems so easy. There you have it. Now point your nose this way. You've been here for 20 years. How long do you want to keep doing all this? Surfing is fun. Fighting is fun. Building is fun. At this point, I think about what I want to build. I think about the future and what will happen. I think the AI will take 10 or 15 years to fully develop. And it will take this same time to operate on the future computer platforms that we are developing. And then I suppose new things will emerge over time. I have an opinion about technology in general. And it seems that we are the proof of a great experiment. How does all this end? I'm not sure that there is a final point really. But I do believe that there is a considerably wider arc. I know that there are many concerned people who reflect and think that the AI will prevent us from our creativity or similar things. I think it will be the opposite. When I was starting with my company, I needed to know everything about programming to be able to build the idea I had. So I had to learn it. And in the future, maybe it won't be necessary. Maybe it's possible to express what a person wants to try to create. And from there, let an AI system generate the coding. Obviously, we will face many challenges along the way. Things change and that leads to new challenges. But I remain quite optimistic about this type of great art that we have present. I have an idea to protect us. I brought you some sunscreen. This is more or less the amount of sunscreen I need. So I thank you. Thank you.