EP 236: NVIDIA GTC Recap - 3 ways NVIDIA is going to change the AI world

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

Nvidia's announcement from their GP GTC conference last week will change our technological future. But it's not the headlines that you read that will make that big of an impact. It's actually connecting all of the dots that matters. So I was lucky enough to partner with Nvidia for the GTC conference. So today we're not just going to give you a recap of the GTC conference, but we're going to tell you the three ways that Nvidia is going to change the AI world. World. All right, I'm excited to talk about that to more today and more on Everyday AI. What's going on y'? All? Thanks for joining us. My name is Jordan Wilson and I'm the host of Everyday AI. We're a daily live stream, podcast and free daily newsletter helping everyday people learn and leverage generative AI. So we are live, unscripted, unedited, the realest thing in artificial intelligence. So thank you for joining us. All right, so we are going to get going here in a second, but I have to remind you, if you're listening on the podcast, this is one of those ones you got to go to your everyday AI dot com. Today's newsletter is going to contain so much important information. Not just that you need to know to understand the future of generative AI, but a whole lot more. So make sure if you haven't already, to go check out our website. I tell people it is a free generative AI university with more than 230 now backlogs of episodes, live streams, etc, so make sure you go check that out. But before we get into the three, the three ways that Nvidia is going to change the AI world, let's first do as we do every single day and go over the AI news. All right, so first, MIT researchers have found an AI image breakthrough. So MIT researchers have discovered a new method for generating high quality images with a single step, reducing the time and computational resources needed. So previous diffusion models required multiple iterations to generate high quality images. But the new method, which is being called Distribution Matching Distillation, only needs one step. That's a mouthful, right? But intas DMD or Distribution Matching Distillation, produce images comparable in quality to those generated by more complex original models and achieved similar state of the art performance in text to image generation. So actually some, some pretty big news there coming out of mit. All right, our second piece of news. Pretty relevant to today's show about Nvidia's dominance. But is Nvidia too dominant? Some big companies are banding together because they think so. So Intel, Qualcomm, Google and others have created a group to combat Nvidia's dominance. So they've created the Unified Acceleration Foundation, a consortium of tech companies aimed to challenge Nvidia's dominance in the AI market through an open source software suite. So the group also. So here's the complete lineup. So it consists of Intel, Google, arm, Qualcomm and Samsung. And it aims to deliver software and tools that can support various AI accelerator chips and liberate developers from the consumer constraints from what they say are the constraints of Nvidia's exclusive technology. So the project is expected to be fully up and running this year, aims to eliminate obstacles and enhance adaptability for AI developers across different platforms. So despite Nvidia's market dominance, though, here's the important part. The consortium plans to eventually support Nvidia's hardware and code while also seeking collaborations with other chip makers, chip makers and cloud computing companies. Wow, that's got to say something when you never thought that companies like Google and Intel and Qualcomm and Samsung would band together, right? You say that five or 10 years ago and you say, hey, it's to go up against Nvidia's dominance. That wouldn't make sense, but that's just the day and age we live in. All right, our last piece of news for the day. OpenAI has released a lot of new examples of some of the best and most versatile examples of its AI text to video software, Sora. So the examples are from artists and creatives who were granted early access to SORA to bring their ideas to life. And this is actually the first at least official kind of announcement that We've heard from OpenAI about Sora since the announcement now a couple of months ago. So, so these, this kind of behind the scenes look that's sharing some of the best examples from creatives, shares how SORA has helped them break free from traditional constraints and visualize new and surreal concepts. So make sure to check out today's newsletter to see those examples. So go to your everydayai.com and sign up for that. And hey, also Sora, if you didn't already know, SORA is about much more than just the AI video, but going to be talking about that today's show as well. All right, hey, thanks for joining us to our live audience. Where are you Joining us from. It's, it's been a while since we've had a, a full week of live shows, so I'm excited. So thank you all for joining. You know, Mike joining us and Dr. Harvey Castro, Rolando, Brian, Juan, thank you all. Carolyn, Brian and Chris, appreciate you all tuning in. You know, I'd really like to know from you what would your biggest takeaway of Nvidia gtc. So I'm going to go ahead and. Yes, very. We were very lucky here@everyday AI to partner with Nvidia. We got a lot of behind the scenes access. We got, you know, kind of some exclusive information and behind the scenes, uh, intel on exactly what's going on at Nvidia. Uh, so before we get into, uh, kind of the three ways that Nvidia is going to change the AI world, uh, I first just wanted to give you a quick recap of the conference as well. All right. Actually, no, I'm going to skip to the end. Let's skip to the end, shall we? So I'm going to give you the, the answers right now. Right? So if you have a super busy day, you can get on with your day, but here are the three ways that Nvidia is going to change the AI world. Ready? So for the first time ever, we now have a compute sweet spot. All right, number two, the world will now be a simulation. All right, and number three, Nvidia is creating humanoids and AGI. Oh, interesting. But they're not. But they are. All right, so I'm gonna. There's, there's the end. But today is hot. Take Tuesday, y'. All. So, so let me know if you are joining us live. I can take it nice and easy. You know, give me, give me one, two or three flame emojis. You know, how, how hot should today's hot take Tuesday be? Let me know. I always listen to you. If, if, if you all want it kind of middle of the course, we'll, we'll keep it there. So before we get into those three very important ways that I believe Nvidia is going to change the AI world, first, let me give everyone a very brief recap of the Nvidia GTC conference. So the conference was last week, so now we are full week away. So it was essentially Monday through Thursday of last week. So about four days of hundreds of speakers, workshops, tens of thousands of attendees. But here's just the biggest takeaways if you missed everything. All right, so what grabbed a lot of headlines was the Blackwell Platform. All right, so that is Nvidia's new GPU chip and reportedly it is four times more, more powerful than current GPUs, which you have to keep in mind that's crazy, y', all, because Nvidia's Hopper GPU chip was already the world leader. It was already the most powerful gpu. And if you're not quite a dork like me, let me explain why this is even important. Well, all generative AI needs these GPUs, right? So when you go in and you put a prompt into, you know, copilot, or you run, you know, an AI video in Runway or in Sora, etc, right? Eventually, somewhere along the line that is being computed, right? To talk in extremely simple terms, your, you know, a lot of people are, are astounded the first time that they use generative AI and they say, oh, how can I put in a, a ten word prompt and get something unbelievable on the, on the other end? Well, it's compute, right? It is these GPU chips, right? So these GPU chips, it's, it's, it powers everything. You know, technically these GPU chips power our economy, right? People looked at me like I was crazy when, you know, now about nine months ago, I said Nvidia is the most important company to the, to the U.S. economy. People didn't understand that. This is well before, you know, I'd even partnered with, you know, Nvidia on, on anything, right? This is before I knew anyone at Nvidia. I said Nvidia is probably the most important company to the U.S. economy. People didn't understand that nine months ago, a year ago when I started talking about it. But this is why, you know, the, depending on reports right now, Nvidia has anywhere from, you know, know 70, you know, mid 70 to mid 80% market share on GPU because their GPUs are so much more powerful than everyone else's. And literally now all of the work that we do because generative AI is being integrated into everything that we do, whether you've realized it yet or not, right? Like if you're working on a new, you know, Windows machine, you might have Copilot running your entire operation, right? That is all being run by GPUs. And for the most part it is Nvidia's GPUs in their, you know, their current, you know, at least as of last week, their current most powerful GPU hopper, this new Blackwell. So the Blackwell system, or the Blackwell chip, however you want to refer to it as, is four times more powerful than the current. Like I'm laughing because that's never happened before. Like it doesn't happen where you have a leader in a certain type of technology and a company updates said technology to make it four times more powerful, reportedly 12 times more energy efficient. Which energy is extremely important here when we're talking about the power of generative AI in the future of compute. All right, so the biggest recap, you know, at least what was grabbing headlines is the new Black Wealth Platform 2 was NIMS. So this is, and again this is just our recap, not our, our three ways Nvidia is going to change the AI world, but nims, which is essentially a new way to distribute software. So that is the Nvidia interface models and more or less it's a way to easily deploy custom AI models within Nvidia's kind of AI software suite. Nemo and the Nvidia AI foundry. This is essentially an end to end cloud framework in a powerful toolkit. And then last, last but definitely not least number four, Omniverse and Isaac Robotics updates. So that's essentially, think of it like this, it's a robot gym, right? It's extremely impressive and we're going to be sharing about it in the newsletter. So again go to your everyday AI.com and sign up for that free daily newsletter if you haven't already. But updates to the Omniverse and the Isaac Robotics, you have the Groot kind of humanoid robotics framework that we talked about with the director of Nvidia's robotics last week as well. So a lot of huge announcements from Nvidia at the GTZ conference. I mean essentially they're, they're changing the way the future is being built, right? And it's also important to talk about this. Nvidia makes chips, but they are so much more than a chip maker, right? Like that was one of the biggest messages that I took away is Nvidia trying to, you know, shift away from this thought that they are a chip making company, right? So even Nvidia's CEO Jensen Wong said that, right? So Jensen said, like we are not chip makers. We are creating the, the future of technology, which I would obviously agree with. All right, so let's see. All right. Carolyn said hot. She said bring it hot. Brian said two flame emoji. Cecilia said three hey Juan said burn it down. All right, Juan, I'll burn it down. I'm feeling a little spicy today. Today. All right, so now let's get in, let's get back to the three ways that Nvidia is going to change the AI world and I'm going to go into depth on these. And hey, because y' all wanted it, I'm going to bring a little heat. All right, so again, as a reminder, the three things we now have a compute sweet spot. That's number one. Number two, the world will now be a simulation. And number three, Nvidia is creating humanoids and. Hei. Yeah, I saved that one for last. All right, let's get into it. Let's talk about a. Let's, let's talk about number one. We now have a compute sweet spot. All right, again, everyday AI is for the everyday people. It's for everyday people. So if you are a huge GPT and compute nerd, excuse me, because I'm going to, I'm going to go ahead and try to simplify things and you know, I might get things only 95% accurate when I'm trying to simplify it to make sure it makes sense for everyone. All right, so what does hitting a compute sweet spot mean? Well, I would say for the, you know, whatever, whatever timeline you want to put onto it, but let's just say generative AI, because even generative AI, the definition of what it is and what it means has changed exponentially, you know, over the course of the last couple of years. But let's say we've been living in a, the generative AI world, let's just say since ChatGPT, right? Since Chat GPT debuted. So let's say we've been living in a generative AI world for, for two years, give or take. Right? We haven't had enough compute collectively. Right. Does the, the, the Blackwell announcement mean that we have enough compute? Absolutely not. Right. It's one thing Jensen was, was talking about during his keynote and you know, in some closed door, you know, Q and A's, he said we need to build more chips, bigger chips, more powerful chips. Agree. So at least for right now though, we have a compute sweet spot. But don't get me wrong, compute is still currency, right? That's kind of the hot, the hot thing all the smart tech CEOs are saying. I've been saying it on the show here for months, but now it's cool because all the, all the billionaires are saying it, that, that compute is still a currency and it is. Right? It doesn't matter for the, the hottest AI startup out there or you know, the, the enterprise company that's, you know, deploying generative AI across their workforce. It doesn't matter how good you are, how great your ideas are, if you don't have the compute, right? Think of it like power outlets, all right? You could have, you know, 10 computers and 10 phones and all these things, but if you're sitting at the airport and there's only one available power outlet, you can't do all that work. All right, again, oversimplifying here, but that's what compute is, right? So as the collective business world, especially here in the US we've talked about 2024 is the year of generative AI implementation. You need enough outlets on the wall to plug, oh, here's, here's our new, you know, AI enabled CRM. Here's our new generative AI powered customer service, you know, solution. You need those outlets, you need the compute, right? And it has been and it still will continue to be a problem having enough compute. All right, but the fact that we went again without getting too technical here, Nvidia's Hopper GPU chip was by far the most powerful in the world. And then we get this Blackwell chip that was just unveiled last week that is reportedly four times more powerful, 12 times more energy efficient depending on, you know, how you're, you know, looking at the different tasks, different benchmarks, there's different, you know, different multitudes or multiples, but essentially this, it is way more powerful, way more energy efficient, but it's still not enough. So here's why we have a sweet spot. Foreign. Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on gen AI. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because, because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too, just like some of the biggest companies in the world do. Go to your everydayai.com partner to get in contact with our team or your you can just click on the partner section of our website. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to roi. On gen. Models are actually getting smaller, right? We've talked about it here on the show. Not just with edge computing or generative AI on actual devices. But when you look at what Mistral is doing, right, the French company that is creating large language models, but you know, we, we still do have models getting much larger. Okay, so your, your trillion parameter models like GPT4, presumably GPT5 will be trillions. Gemini, Claude. Claude's new, you know, their new models. So we still have large language models getting bigger, but we have now so many models getting smaller, which reduces the need for compute. Right, the same thing. I mean we even talked about a very small example. But you know, the MIT researchers with their distribution, matching, distillation, the DM. Right, the DMD. So so many researchers, startups, etc. Have been spending the last year and a half because compute has been a problem. They've been spending so much time, energy and resources on how can we make generative AI faster, how can we make models smaller, how can we, you know, kind of reduce the number of generations needed, right? So let's just say in the cloud something's computing 20 times over. They're saying how can we get it once, okay? So that's why we have at least right now a compute sweet spot. And I don't think a lot of people fully understand that. You know, this is something I was talking to exhibitors on the GTC floor about. You know, hey, what do you think of Blackwell? And a lot of companies were excited and rightfully so. You know, you have to think of it like computer software, right? Or maybe your, your smartphone software, right? When a new update comes out until all of the other apps or you know, everything else starts demanding more power, you almost have this sweet spot where everything just runs better than expected, right? That's where we're going to be at in the coming months. And maybe, I don't know how long this is going to last, maybe a year or so. So this is actually a great time, a great, great moment for generative AI, because you are going to see now all of these companies we've been Talking about this 2024 is the year of implementation and companies for the most part will be able to implement generative AI solutions top to bottom, company wide, enterprise wide, in a somewhat cost controlled way, right? Where I think, let's just say if we weren't getting new GPUs from Nvidia or Qualcomm or AMD, etc, if we weren't getting all these new chips, we would be having a problem as a country getting enough power outlets, right? Because now all of a sudden every single business wants to take their Generative AI, plug, empower it and plug it into a socket, right? But two years ago there weren't really enough, right? Like researchers and AI companies were looking around and saying hey, we're running out of outlets, right? And if, and if businesses want to power, you know, their business with generative AI, we either need to reduce the power needed or create more outlets. So we kind of have both of those things happening in tandem with the focus on on edge AI or you know, models small enough to live on a device. Which really negates the need for so much of these cloud resources, right? If you can run models locally on your phone or locally on a machine, it obviously reduces the, the, the load that you eventually need, the compute power that you actually need because it all is living locally, right? But every time you go on to, to chat GPT or to, to Runway or you know, maybe once we get all, get access to, to Sora, that that just eats away computes. And you know, unfortunately right now compute is still expensive and it is not terribly energy efficient. New updates like Blackwell help to solve that, but it's still, I won't say problematic, but it's still something that you have to take into consideration. So we have that compute sweet spot. And I also think, right, all the companies, the OpenAI's the, you know, I keep saying Runway mid journey, anthropic, right? All, all of these companies are using Nvidia's chips, right? So presumably another way that this changes and kind of working in this compute sweet spot is I'm sure over the last year or so that you or your company has been using different generative AI services. You get errors all the time, right? And, and, and a lot of, you know, consumers that don't fully understand kind of this compute paradigm look at this generative AI technology and they're like, oh, it kind of stinks. Time's out all the time. We get errors, you know, there's, there's outages. Well, it's because hundreds of millions of people are all of a sudden rushing to use all of these services that didn't exist before, right? It's, you know, Jensen said something in a Q A session, he said, you know, we're creating demand from nowhere. We're not borrowing all of this demand from somewhere else. So we do have a sweet spot in compute. And I also think as an example, right? So we talk about Sam Altman, you know, reportedly raising $7 trillion for compute alone. Well, I would say my hot take here, since you wanted the fire, is I think he Saw, right. I, I think at some point all these big partners who are spending tens of billions of dollars, I'm sure they get a look at what Nvidia has in the pipeline and I'm sure he saw the just huge jump right between Hopper and, and Blackwell and saw, wow, we should be investing in compute ourselves. We should be creating chips. Right. That's why you now see, you know, Amazon, Microsoft, etc, creating now chips in house because everyone is still going to need compute, but at least for right now, kind of the quote unquote software in the hardware. I think in the coming months, maybe year plus are going to be working in harmony now you have more companies that didn't even exist a year ago popping up that are allowing you to share kind of compute resources in the cloud versus, you know, spending, you know, a couple a billion dollars or hundreds of millions of dollars to get, you know, a thousand of these, you know, GPUs running at your, you know, at your company. All right, so that's number one. Yeah, yeah. Hey, we're getting there, Ben. Yes. The world is a simulation. Yeah. Worry. We're getting there. All right, so that's number one. Number two, the world will now be a simulation. Yeah, it will. Right. I'm not going to get too meta on this. Obviously we're still living in a physical world, but the way things are going and this is, I would say at least things that I learned the most at the Nvidia GTC conference is just the power of simulation. Right. Obviously I knew how simulation worked and some of the use cases that AI powered simulations and digital twins. Right. We've talked about it, we have experts, we've had experts on this show before talking about the power of digital twins. But a lot of these new announcements from Nvidia changes what is even possible with digital twins and simulations. So if you don't know what that is, let me try to do my best to explain it to you. Think let's say your company is investing, you know, $50 million in building a new manufacturing plant. All right, so maybe the way that you would traditionally, you know, build this plan is you look at your current plant, you model it after that you hire some expensive consultants to come in, you know, run some, maybe you run some simulations, but for the most part you're probably just looking on what you think has worked and you're putting forward your best guesses. That's incredibly inefficient. Right. And it's not leveraging technology. Right. So with simulations, essentially you can simulate an entire New factory that doesn't exist. Right. Based on millions and millions of other simulations that have been run. So that's an example of something that you can do, you know, not just in Nvidia's platform, but many other, you know, generative AI and AI powered cloud platforms is you can run literally millions of simulations of your warehouse that doesn't exist yet based on your projected data, your products, you know, how much space do they need? How many people are going to be, you know, operating in this facility. And you'll see by, you know, these platforms running literally millions of simulations for your manufacturing plant that doesn't exist yet, going to see problems. Oh, these, these shelves are, are too close together. You know, they're not going to be able to, you know, operate and move things from this point, you know, point A to point B. There's going to be a lot of collisions here. So we, you know, you need to rethink how you set up this part of your manufacturing or this part of your production line. Right. So literally that's just an easy example hopefully to understand on how simulations are run and how they're already, and they've been being used for years. So now we have some updates to the Omniverse. Right. And so the other thing also going back to point one about the sweet spot of COMPUTE is it also changes who can use all of these platforms because at least for the foreseeable future, it's, it's lowering the barrier of entry to play here, right? To play in some of these very powerful, normally enterprise only environments. Right. It's, it's, it's kind of like with chat GPT and we've talked about this on the show, you know, so many, so many times before. This is the first time, I believe, right? I've, I've been, you know, kind of in Martech and in communications now for full time for 20 years. And this is the first time in my 20 years on earth that the average company can go out and spend, you know, 20 bucks, 50 bucks, a hundred dollars a month and get similar level of technology that enterprise companies can get. That's another part of, you know, kind of being in this COMPUTE sweet spot. But also point number two on how the world will now be a simulation because now more and more companies can start to use, just in this example, Nvidia's Omniverse or you know, other kind of world simulators powered by AI to simulate everything. Right. Hey, you guys wanted hot takes? Here's, here's what I think. I think simulations are going to get weird. I think they're going to get weird. Right? Right. Now, the example I gave of, all right, your company's spending $50 million building a new production plant. That makes sense, right? That makes sense. Yeah. You want that to be simulated with millions of simulations using, you know, Nvidia's data from, you know, all of their systems, uploading all your data. But I think a lot of things are going to be simulated, right? I could even see if, if you want to get super weird. I could see a time in the, in the future, I don't know if it's months or years away, where I could simulate this exact podcast, this exact live stream, you know, and I tell a system, hey, here's, here's what my topic is. Here's some of my bullet points, and it's going to go ahead, it's going to run millions of simulations. Oh, what if you go down this route? What if you go down this route? Here's how your audience is going to react to all these different things based on all the data that we have on you, your audience, how they respond to certain content, etc. When, when, when you tap into these powerful platforms like the Omniverse, they're not the only ones, but if you bring your data and you are able to tap into literally a world of knowledge, simulations are, I don't know if that's weird or extremely powerful or both, but that's, that's where I see things heading, right? Where every single, I think in the future, every single product, service that you interact with has already been simulated, right? As, as a digital twin in the Omniverse, in the meta, in the Metaverse, etc, where almost every single thing that we interact with, the cars we drive, the devices that we use, you know, the software we send emails on, I think everything will have already been simulated millions or billions of times already. Simple day to day interactions with products that we use all the time. Hopefully what that means is more intuitive products and services, better experiences in the real world. That's the point of simulations. That's the point of, as an example, the Omniverse. But I do believe the world will be a simulation. I think almost every, you know, single, you know, the, the clothes you wear, the shoes, you know, the, the new pair of Nikes that you go buy, they will have been run through a, like a digital simulation millions or billions of times right across different textures. You know, what happens if, you know, this person has a bigger left foot than right foot? What happens if, you know, they constantly go from wet to dry surfaces. Right. Like, all of these products and services that we use every single day will have already been simulated millions or billions of times. Hopefully. Like I said, hopefully. That doesn't make it weird on our end, but it just makes for better product services and experiences in the real world. All right. Yeah. Hey, Earth 2. Hey, great, great point, Juan. Thanks for joining us live. Yeah, Earth 2, that's, that's a great point. If, if you read some of my kind of insider takeaways from GTC Earth 2, that's something that Nvidia announced as well. A literal Earth simulator. Right. So you can't look at me too weird and be like, all right, Jordan, this is weird. If you say the world will be a simulation. Well, Jensen Wong talked about that at, even at his keynote. Some, some new updates to the earth 2 kind of their world simulator, and they're trying to, you know, reduce climate change and to better predict extreme weather sooner. I think literally every aspect of our daily lives will very soon have already been simulated. Hey, I love this comment from Josh. And hey, shout out to Josh. Josh gave me, like, probably the funniest reply to a comment ever on LinkedIn. But Josh just said, this is like when Dr. Strange ran through millions of simulations to identify the reality that would defeat Thanos. Yes, exactly. Yeah, if you're Avengers fan. Fan, exactly. That. You know, running through literally millions or billions of different scenarios to give all of us, hopefully, ideal business and personal outcomes. That is what AI powered world simulators mean to all of us. All right. Hey, Tanya, I don't, I don't think I can get into this one. She said, what about running simulations for potential marriages? I'm sure it's already happening. Right. I'm very happy that, that I lucked out with mine, so. Yeah. But I'm sure, I'm sure that's going to be a thing where, you know, people. Yeah. Are simulating everything. I'm sure that already exists. You know, simulating things in, in dating apps or matchmaking. I'm sure that already exists, but yes, everything is going to be simulated. All right, last but not least, I know some of you listen, you know, when you're walking your dog or on an exercise machine and, you know, you might only have 35 minutes on that machine. You can't listen to me forever. So number three, Nvidia is creating humanoids and AGI. So here's the thing. They're definitely not, but they are. Okay, let me, let me tell you what that means. Humanoids. All right, again, we talked with Amit Goel, the Director Nvidia of Robotics. Nvidia is not building physical robots, but their Groot system. Love the name. I think it's General stands for general robotics. 003 I, I, I think is technically what it stands for. So the Groot system allows every other robot manufacturer to use Nvidia system to. Again, we talked about the Isaac, right? Isaac Robotics, essentially a gym for robots, right? Seeing, seeing some of those simulations where you just see millions of these robots walking around in a kind of digital twin real world Omniverse. Isaac Robotics, it's their gym. They're just existing and they're getting smarter, right? So all of these companies that are building physical products are more than likely already using either Nvidia's chips or their software solutions like Isaac Robotics, like the Omniverse. So Nvidia is not creating any physical robots, right? But companies as an example, like figure. So we shared about figure on the show about what, nine days ago, right? And we showed you that video where there's a robot that is not only interacting with a person, so it is doing general robotics where it's, it's reacting to things in real time, it's talking to a person, it's completing tasks based on a person, it's doing one task and answering questions on another. Right? That's the example of the figure01. It's using Chat GPT to listen and to process and to speak in natural language. You know, that's figure 01 is presumably running on Nvidia GPU chips. Nvidia is, is one of the main investors, financial investors. They're one of the biggest partners with, with figure, you know, and I do think figure, at least in my opinion, is one of the companies on the robotics or the humanoid side that's kind of running away with things. But guess what, it's no surprise that they are one of their biggest investors is Nvidia, right? So all of these companies that are creating these humanoid robots are either using Nvidia chips, they are using Nvidia's Isaac Robotics platforms. They're going to be using the new Groot kind of foundational system. Nvidia is intertwined into everything, literally everything. If you want to talk Meta, Nvidia, if you want to talk OpenAI, Nvidia, whatever your favorite generative AI system is Nvidia, right? So they're creating humanoids and in, in fact AGI, right, we were able to ask Nvidia CEO Jensen Wong questions, you know, kind of in a closed door session and he was asked about AGI, right? And how far, how far out are we? And you know, what he said is, you know, hey, depends on what you even consider to be artificial general intelligence, Right. Human intelligence is obviously changing. You know, what technology and AI is capable of is, is changing all the time. You know, his kind of blanket statement was probably five years. I think we're probably a lot closer than that, right? There's always an argument and people say, oh, what is AGI? Right. But that's essentially when. Let's just use the figure robot as an example because I think that's pretty, pretty easy to visualize when you have a physical form humanoid robot, right? But you power it with a large language model like CHAT GPT, right? So at what point can that, you know, figure slash CHAT GPT perform tasks, both mental and physical, general tasks at the same level better than the average human? I think we're a lot closer than people think, right? Especially when, you know, probably still the strongest model out there. Sorry, Claude. 3. Sorry. Gemini 1.5. I still think GPT4, which is an old model, is still the best model out there across the board. When you look at outside functionalities and features, when you think of right now that a leading model is two years old and when you think, hey, what happens when the next GPT, whether it's 4.5 or 5, what happens when that comes out? Yeah, I think we're getting close, closer and closer to AGI. Right? But a couple of things, some of my takeaways, and I opened the show by talking about connecting the dots, right? Between what was said and what was not said, but what was implied or maybe things that a lot of people miss, right? So during, during a closed door Q and A, I was literally about 10ft from, from Nvidia CEO Jensen Wong. So I was able to get a lot of insights, even just looking at his, his facial expression, how he reacted to certain questions. And he was asked a lot about AGI, about compute, about humanoids, right? And he said, hey, we're going to have our CHAT GPT moment for robotics soon, right? And I think robotics and humanoids and AGI are, are all a lot closer tied to each other than most people realize. But one thing that I took away from both closed door and open door sessions at the Nvidia GTC conference is the only thing in questioning that the CEO of one of the largest, I believe they're either the third, I didn't look this morning, but either the third or fourth largest company in the United States by market cap and The CEO only brought up in a closed door session one aspect of AI more than once without being prompted. All right, so what that means is, you know, hey, Jensen, A or B. And he says B, hey Jensen, 1, 2 or 3. And he says 3. But bringing up things unprompted, I think tells about not just the direction of where maybe his mindset is, but also the direction of the industry. And the only thing in that closed door session, and I didn't see anyone else talk about this or write about it, the only thing that Jensen brought up more than once unprompted was Sora. All right, and here's the other thing. Brought it up unprompted. It's not even his product, right? Technically, I mean, Obviously Nvidia powers OpenAI, but you could think of them as, in theory, a competitor because Sam Altman is reportedly trying to raise $7 trillion to create chips, compute, fusion, power, etc. Right? But he brought up Sora on multiple occasions and talked about its implications, not as an AI video tool. That's the thing that most people are forgetting or they're overlooking. He was bringing it up because of AGI. He was. You know, there's a transcript somewhere. We'll try to, you know, try to include that in our daily newsletter today. So sign up@youreverydayai.com. however, he brought up Sora on multiple occasions, not for its video editing prowess. He brought it up because he said he was extremely impressed with its ability to understand the relationship between the physical world, objects in the physical world, and what that means for not just AI, but what that means for humanoids, what that means for artificial general intelligence. So had to connect some of the dots between what was said and what was unsaid at the Nvidia GTC conference. So as a reminder as we wrap up the three, the three ways that I believe that AI or Nvidia's announcements at the GTC conference, that their conference that their announcements are going to change the AI world. All right, we're going to recap. Number one, we'll hit a compute sweet spot, maybe for the first time, it might be short lived. Number two, the world will be a simulation. What these announcements mean for digital twins can't overstate it enough. And number three, yes, even though they're not technically working toward humanoids and AGI, Nvidia is actually helping to create Humanoids and AGI. All right, speaking of Nvidia gtc. Hey, I found the date. I think you still have about 10 more days, so make sure to check out the link in the description. But you can still, even though the conference is already wrapped up, you can still sign up for free. Literally. Some of the leading experts in the world are running workshops that you would normally pay thousands of dollars to attend these workshops. You can go sign up, find the link, you can sign up and still watch all of these replays and sessions for free for I think about another 10 days. So make sure that you go do that and by signing up. If you sign up with our link, you will enter into our giveaway for a free Nvidia GeForce GPU as well as DLI learning credits from Nvidia. All right, that's a wrap for today's show. Make sure to join us Thursday. This is going to be an amazing one. So learning from the Giants WWT's Gen AI blueprint, we are going to be talking with the founder or the co founder and CEO of Worldwide Technology, Jim Kavanaugh, one of the largest y'. All. Let me, let me restate this sitting down one on one with one of the largest companies in the world with their CEO. This was actually, we just recorded this last week at gtc so I can tell you how this conversation goes and it's amazing. You are not going to want to miss that and you're also not going to want to miss today's newsletter. Go to your Everyday AI dot com. Sign up for the free daily newsletter. And hey, if this show was helpful, go ahead please. We spend hours putting these shows together. It takes you about 10 seconds to go repost this, re it, share it. Here's what I'll do. If you're on the live stream, you see this. If not, you're on the podcast I'm sharing. I, I have a picture here with Nvidia CEO Jensen Wong, but there's actually a pretty embarrassing story that happened before this. This photo was taken. So if you repost this, if you share this on social media, let me know and I'll tell you in a private message. I can't, I can't put myself on blast there. It's extremely embarrassing. I'll tell you a funny story. So thank you for joining us. We hope to see you back tomorrow and every day for more Everyday AI. Thanks y'. All.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit youreverydayai.com and sign up to our daily newsletter. So you don't get left behind. Go break some barriers and we'll see you next time.