Mark Zuckerberg NEW STATEMENT Changes EVERYTHING!

So there was a crazy announcement a couple of hours ago and I honestly should have covered this earlier, but there was a lot of information that I wanted to include into this video because Mark Zuckerberg just made an announcement about AGI that honestly left me quite shocked because I didn't see this coming from him, considering his company was just focused on the Metaverse not too long ago. So take a look at this announcement because we got to do a deep dive about this and I'm pretty sure it's going to show you guys too.

Hey everyone. Today I'm bringing Meta's two AI research efforts closer together to support our long term goals. Building general intelligence, open, sourcing it responsibly, and making it available and useful to everyone in all of our daily lives. It's become clearer that the next generation of services requires building full general intelligence, building the best AI assistants, AIs for creators, AIs for businesses, and more. That needs advances in every area of AI, from reasoning to planning to coding to memory and other cognitive abilities. This technology is so important and the opportunities are so great that we should open source and make it as widely available as we responsibly can so everyone can benefit. We're building an absolutely massive amount of infrastructure to support this. By the end of this year we're going to have around 350,000 Nvidia H1 hundreds, or around 600,000 H100 equivalents of compute, if you include other GPUs. We're currently training Llama 3 and we've got an exciting roadmap of of future models that we're going to keep training responsibly and safely to. People are also going to need new devices for AI, and this brings together AI and the Metaverse because over time I think a lot of us are going to talk to AIs frequently throughout the day, and I think a lot of us are going to do that using glasses, because glasses are the ideal form factor for letting an AI see what you see and hear what you hear. So it's always available to help out. Ray Ban Meta glasses with Meta AI are already off to a very strong start and overall across all this stuff, we are just getting started.

So that announcement is incredible because there is a lot to dive into if you know anything about AI. So he also did post this on Instagram. This is basically the text version and I think this is a crazy announcement because not only does he talk about what he's doing with AGI and mainly he basically said AGI, he wants open Source AGI, which is absolutely incredible because. Because before he was just doing standard LLMs and now they're doing open source artificial general intelligence. And I find it very interesting and fascinating that he's also made the same shift that OpenAI did, because if you remember, it was only around three or four months ago that OpenAI made the shift from saying that we're going to be building XYZ to our main focus is AGI. But Mark Zuckerberg is talking about open source AGI so that everyone can benefit. So that is a really good thing and that is truly, truly fascinating. Now, something else that he did say as well that was pretty crazy was that the cluster announcement. So their GPU cluster is absolutely incredible, guys. 350,000 H1 hundreds and almost 600,000 H1 hundreds equivalent of compute by the end of the year is absolutely incredible, guys. Like that is so much. There was even an image that I'll show you guys later on in the video that just goes to show how crazy of an announcement this is. And later on in the video we also will talk about AI devices because I think that their Ray Ban Meta glasses are absolutely incredible. But one of the things I do want to talk about is the fact that people may have missed the mark here that he says he's building open source AGI. So that's really actually going to change the world. I think OpenAI will change the world, but they're a company, they're developing a product. But open source AGI is truly going to change the world because that means that everyone can access it responsibly, which means essentially it's going to be free. So that is incredible. Okay, now he also tweeted this on threads, but it's pretty much the same thing. And yeah, so let's get into some of the other things because there was someone at Meta that said something that I did also find rather interesting. So Meta's AI chief doesn't think that AI superintelligence is coming anytime soon and is skeptical on quantum computing. Now, if you don't know who this person is, this is Yann Lecun. And this guy, pretty sure he's at Meta, so it says. Facebook parent Meta held a media event this week in San Francisco highlighting the 10 year anniversary of its fundamental AI research team. And society is more likely to get cat level or dog level AI years before human level AI, Meta chief scientist Yan Lecant said. And I think I don't want to say that he might live to regret that statement, but I do think that if certain companies are working on what we think they're working on, then I do think that it is coming faster than many people do think. And I do think that the current, you know, area and the current speed of how fast we're going, I think it's currently being, I guess you could say underplayed, and I think rightly so, because if it was overblown, I definitely think there would be some more regulations trying to stop this. But it makes sense to downplay, honestly. So he said, unlike Google, Microsoft and other tech giants, Meta is not making a big bet on quantum computing. And quantum computing, there's a recent video that actually talks about how quantum computing is essentially could be going away in the sense that AI is taking the spotlight because AI allows you to get much more out of traditional computers. Whereas quantum computers, you know, since they are just so technical and so hard to get compute out of in terms of actually making them work and error free, AI is making that, you know, somewhat redundant, I guess you could say. So. His article was fascinating because of course, you know, he's a chief scientist of Meta and Mark Zuckerberg just comes out and says, you know, we're focusing on open sourcing AGI. We're, you know, getting 600,000 H1 hundreds the equivalent, which is insane. Which means they're going to be training loads of models and larger models. And, you know, later on the video, I'll show you guys some interesting stuff because we literally talked about llama 3 yesterday. So I find it fascinating that we talk about llama 3 and then Mark's talk about it today. He's like, yeah, Llama three is the thing. So essentially you can see Meta's scientist and deep learning pioneer, Yann Lecun, said he believes that current AI system are decades away from reaching some semblance of sentience equipped with common sense that can push their abilities beyond merely summarizing mountains of text and creative ways. I disagree, okay? And I don't think I'm smarter than the chief AI scientist at Meta. So that's not what I'm saying. I'm just saying I disagree, okay? He says his point of view stands in contrast to that of Nvidia CEO Jensen Huang, who recently said that AI will be fairly competitive with humans in less than five years, besting people at a multitude of mentally intensive tasks. And the thing we have to think about is that, remember, every single time we've made AI predictions before, we said AI is never going to be able to do creativity. It did. Creativity. AI is Always going to replace the artists last. It replaced the artists first. AI is gonna never really replace mentally challenging work first. It's always gonna be the easiest jobs that get replaced. And it looks like AI is automating the most mentally challenging tasks, including data analysis and things like that with code interpreter. So I just find AI predictions very, very hard because humans do have a poor, poor time at predicting the future in terms of technology. I mean, if we just go back 30 or so years and if we look at the technology we have now, it's absolutely incredible. So he said. I know. Jensen Lecun said at a recent event high the Facebook parent company's 10 year anniversary of the fundamental AI research team. And he said the Nvidia CEO has much to gain from the AI craze. There is an AI war and he's supplying weapons. That is a very, very decent point. And he said if you think AGI is in the more GPUs you have to buy. Le Kun said about technologists attempting to develop artificial general intelligence, the kind of AI on par with human level intelligence. As long as researchers at firms such as OpenAI continue their pursuit of AGI, they will need more of Nvidia's computer chips. And that is pretty true because they're pretty much the only company that has. Of course there's companies like AMD. Nvidia owns like I think 85 or 95% of the market share in terms of what GPUs people are training on. And even companies like Google that make their own GPUs, they just aren't fast enough. And companies really are trying to make their own. But I think it's, I don't, I wouldn't say it's pointless because of course you don't want to depend on a company like Nvidia because if you're dependent on one company, if anything ever happens to that company, yours kind of slows down. But you know, Nvidia are just absolutely incredible at what they do. And I find that trying to compete with them and what they do is, is, it's admirable, but I don't know if it really makes any sense from a business standpoint. You can see right here that text is a very poor source of information. Lacan said, explaining that would likely take 20,000 years for a human to read the amount of text that has been used to train modern language models, train a system on the equivalent of 20,000 years of reading material. They still don't understand that if A is the same as B, then B is the same as a there's a lot of really basic things about the world that they just don't get through this kind of training, LeCun said. And I think that while his statements are fascinating about the fact that AGI isn't that quick and it's not going to come anytime soon, I think that if certain companies like OpenAI and Google and other independent labs like Mistral allowed to keep progressing at the rate that they're progressing, and companies like DeepMind are continuing to build on systems like Gemini, I really do think that we could definitely get AGI within the next five years. And I don't think that it's decades away. Like really, really don't think it's decades away. Because, you know, recent videos circulating on the Internet from like AI explained showcasing that smart GPT and other systems have so much more potential just based on changing how we prompt and GPT4 has been out for pretty much just over a year now and we've seen the abilities of GPT4 just based on prompting like it increased what this model has been able to do. So I think OpenAI in that time have definitely developed smarter and smarter systems and maybe they're going to release them, maybe they won't, maybe they'll lobotomize them, maybe they won't. I mean we really have no idea. And I do think even if OpenAI invented invented a super smart system, I don't think they would release it straight away because I think OpenAI wants to gradually introduce us into these models because that makes sense. And I think that if the model was just absolutely incredible, I think it would definitely put fear into the markets. And of course that is not something you want to do, especially when OpenAI doesn't really need to do that, especially considering right now they are the market leader. So this image right here shows us the Nvidia H100 GPU shipments by customer and these were estimates based on some data, but it goes to show just how crazy Meta's investment is. And this was last year and I'm guessing that next year it's going to be even crazier. So we don't know how much Microsoft is getting. I mean there are some estimates rolling around and I think there was an estimate that there were around 60,000 GPUs used to train GPT4. So either way, GPUs are going to be the future to train these models. And I do think that in the future we're going to see bigger and bigger training runs. But and additionally I Don't think the Moore's Law for training runs is going to be there because I'm pretty sure that these systems are going to get more and more efficient with within the same amount of space. So one thing that nobody seems to be talking about, and everyone's talking about, you know, llama 3, and I'm going to talk a little about llama 3 in a moment. But one thing that nobody seems to be talking about, okay? And if you haven't, guys, if you are someone who is in the AI space, I would urge you, I would urge you, I would urge you all to take a look at this. Okay? So the glasses, the glasses, the glasses, the glasses, okay? This was something that people just glows, glazed over because people were like, llama three. Llama three. Okay, Llama three is amazing. But the reason this is crazy, okay, because he basically made an announcement about the glasses that people just missed, okay? He said that the two major parts of our vision, AI and the metaverse, are connected by the end of the decade. I think lots of people will talk to AIs frequently throughout the day using smart glasses, like what we're building with the Ray Ban meta. So Mark Zuckerberg basically just said in his video that in the future, in terms of like, you know, end of the decade, which is six years away, okay, so then six years, we're going to have glasses that are going to be having an AI system on board. We know that it's completely possible on board. We know that it's possible with real time latency, and we know that with increasingly better vision models, this is going to be absolutely incredible in terms of what AIs can see. So I think that that is going to be absolutely incredible. Now, you might be someone that hasn't paid attention to the Ray Ban smart glasses, but I have, because I like looking at technology, I like staying up to date with these things, and these things are actually good. Like, if you remember Google Glasses from a couple of years ago, you'll know that Google Glasses were strange. They were weird things to put on your face. Google Glasses just doesn't compare to this because these glasses are really, really good. Now, I haven't tried them, but I've seen the quality, I've seen the reviews, and I've ordered a pair, and I really can't wait to try it. So essentially what these glasses do, they have a little camera in them, like right here. So I'm going to play you guys a review in a second. But essentially they have a camera right here that you can record with. And they don't look out like, they don't look like some big, huge glasses. They look like completely normal. They have a little camera there. They have music that can play. You can talk to them. You can be on a phone call with them, you can listen to music. You can do a lot with them. And they actually do act as glasses. So, for example, if you do have glasses, you're someone that wears prescription glasses. You can actually change these out to work with your. Whatever it is that you have, spec savers or whatever. And if it's someone that just wants them to be stylish, they actually do function as sunglasses as well. Now, you might be thinking that this is just all talk, but I really want to tell you guys why the future is going to be crazy. Because as cool as these glasses are, imagine having an AI in them that can see what you see, can hear what you hear, and can help you out in your daily life and is really, really effective. So there's a clip from. And I want you guys to see this because it's really cool and it just explains how good this really is.

Chin lenses, which is. I'm just starting playing music with the R. Transition.

Yeah, I thought they were just clear.

No, check this out. I thought they were just clear too. I went apple picking this weekend and I put them on and I thought my friends were going to be like, david, you don't wear glasses. What are you doing? Are you being lame again? And then what actually happened was they looked like this. And for aud listeners, they turn into regular sunglasses. Oh, they did. Wow. Yeah, they transitioned to regular sunglasses. And that's. So let me explain what everyone loves Transition. Oh, yeah, we don't. We even explain what these actually. Okay, so we're wearing some glasses. So Meta made smart glasses in collaboration with Ray Bans. If you've seen Ray Bans. Yeah, if you've seen Ray Ban sunglasses, they've been popular and expensive and well known for a while. Hundreds of years. Then Ray Ban stories came along. That was the first gen. This is the second gen from Meta. This is a pair of sunglasses that inside of them has a camera, some microphones, some speakers, a computer, a touchpad on the side, and some batteries to keep it powered for a couple hours. And a shutter button on top for your photos and videos. And it looks kind of just like a normal pair of sunglasses, almost exactly like a regular pair of sunglasses. Like, this is one of their styles and this is how. This is how it would look out, out the box if you didn't see the sort of a semi clear thing on the side. You wouldn't even know. Or the camera, there's some tech in there. And the other cameras, so this is what they would look like. And you could walk around, take first person videos and do a whole bunch of computing. Yeah, on these, the photo and video quality is really good, by the way. Yeah, 12 megapixel cameras. So I'll say I agree. Yeah.

And the clip is really, really cool. So you guys can basically see just how they're talking about it. And yeah, here's where you can see the actual quality of this. It doesn't just look like some, you know, really, really bad stuff. It looks really, really cool. So I mean, you know, some people are not going to really care, but I think that the trend of AI wearables are going to be really, really fascinating. We've seen a lot with Rabbit, we've seen a lot with many other different companies like Humane, the Humane Pin. But I think this is, this is the one that probably is going to prevail the most because this one is one that isn't a new device that people have to get into. It's not something that people have to learn. It's not a new trend that people have to adapt. And we know that humans are, you know, we're creatures of like, comfort. And nobody's gonna start wearing something until, you know, everybody starts wearing it. Which is why trends are really hard to spot and they're really hard to take off. But I think this one makes sense because people wear glasses anyways. Some people need to wear contact lenses. These are not intrusive. A lot of people, you know, wear these on a daily basis and they just work. They actually do work. Like, but imagine having these with an AI system that's able to see what you see, can talk to you in your ears. You can have a conversation with it. I mean, it really does seem like Meta could be poised to take over. But that also brings me to the question, is another company going to beat Meta to the punch? So that is going to be something that I will have to see, but I kind of doubt it because, you know, doing all that research, design, engineering, in order to beat a company that has billions and billions of dollars, you know, it's definitely going to be hard. But I do think that with any kind of technology, there's always going to be smaller companies to fill the market. Now, something that Mark Zuckerberg did talk about was absolutely crazy. And this is what I'm saying, we talked about this literally yesterday and that is of course llama3 and yester, I don't know if you guys watched it, but we talked about llama 3 and we talked about Mistral and essentially these are open source AI projects that are basically going to be on the level of GPT4. So essentially there was a leak that we talked about that was saying that llama 3 was essentially planned, or essentially is planned to be on the performance level of GPT4. So open source level GPT4 will change the game because that means there's going to be GPT4 level AI systems that are going to be readily available for everyone to use as much as we want. That would be absolutely incredible. And that also does mean that OpenAI is now forced to release a better model, which we know is largely going to have a big enough gap so that they stay and, you know, stay in the lead and maintain the lead, you know, above open source and above Google, which is means once again we're going to be getting more and powerful, more and more models. So, yeah, so they basically said that if llama 3 is as good as GPT 4, will you guys still open it? Yeah, we will. Sorry, alignment people. So, and also in that video, we did also talk about how GPT4 has a more sophisticated architecture than your standard Llama. And if, you know, there are some crazy, crazy things. Like the other day there was this thing called, I'm not sure exactly what it was called, but it was a coding system that beats AlphaCode 2, which is absolutely incredible. And that's something that I don't want to cover. But the point is, is that we are moving pretty, pretty quickly. Okay, this announcement, I think the fact that they're getting so many GPUs is going to be insane. Like, I think all these GPUs, like, you know, the kind of system that they're going to be training is going to be crazy because I think AGI is not that far away. Once they're able to do it, that's going to be crazy. I also think that this right here is going to be crazy. An AI device with glasses. Although Rabbit's device is good, although the humane pin is good, I think this might just be the thing that takes the cake. And I'm willing to bet that Meta manages to do this. And the fact that they're saying, you know, you're going to be talking to an AI device, you know, a lot, that doesn't make sense. And since they've already got the Meta quest with the Metaverse. I can definitely imagine there's going to be a bunch of AI agents in the metaverse. We're going to be talking to maybe NPCs. I mean, of course, maybe you don't want that, maybe you don't want these glasses, maybe you don't really care. But I do find that it's going to be fascinating, you know, smart glasses and yeah, it's going to be a really, really interesting time to live through, to see this next stage of evolution. Now, my question to you is, if these smart glasses do come out, are you going to be buying them? Do you really care about AI systems like that? Are you just going to wait for people to adopt it? I personally will be testing them. We'll be wanting to see how they perform. And I know that is, of course, speculation, but I mean, it just makes sense. So I'm willing to bet that they're probably going to do this sometime in the Future. And with Llama 3, I'm not willing to bet that it's going to be on the level of GPT4, but I do think that Llama 3 will be close. And I do think this announcement is essentially really, really crazy because he also did say lots more to come soon. So with all that being said, with a huge announcement being made by Mark Zuckerberg really quickly, what do you guys think? Are you guys on Yann Le Kun side? Do you think that AGI is far away? Do you think that we have many years before we're able to get comprehensive AI systems on the level of humans? Or do you think that this crazy techno future is on the rise? I think it's a bit of both. I don't think we're gonna get AGI this year, maybe not next year, but I think definitely in the next five years we will most likely have it, provided that everything remains stable. So with that being said, let me know your thoughts and I will see you all in the next video.