Keynote · 2:00 PM · Thursday, May 14, 2026
Our next speaker is somebody who I've really enjoyed getting to know recently, Cathy Hackl. She's an author, entrepreneur, futurist, and one of the most influential voices shaping what comes after the AI hype cycle. And she's frequently cited as one of the most accurate predictors of future tech and consumer behavior, and she's earned the title the godmother of the metaverse. So I'm not going to go on into too much more detail because we want to hear you speak. So please welcome Cathy Hackl who will challenge us to think beyond LLMs.
Thank you so much. I am thrilled to be here. I took the red from Seattle yesterday. So, but I got, you know, I took a nap, so I'm good. I'm really excited to be here, and I think out of the talks that you've heard, I think some people have mentioned physical AI. Mohamad mentioned a little bit what they're doing with robots. I'm going to take you guys a little bit further into a lot of the things that are about to take off when it comes to AI beyond LLMs. So let's go there.
So I think a lot of us can agree that the next AI revolution is not necessarily going to continue to happen in the chat windows. We will all continue to use the chat windows, but the next phase of where AI is going is the physical world. Like I mentioned, I just got back from Seattle. I did a very similar keynote for a couple of executives at T-Mobile. I think a lot of people are starting to figure out and think about what comes after LLMs, right? LLMs are all having a huge impact in our lives, but what comes after, right?
I love starting my talks with this slide from Jensen Huang. He's shown this slide several times at several of his keynotes of the trajectory of where AI is going, right? I'm not going to focus on perception AI because that's kind of been there and done that. Generative AI, which obviously, as Mohamad mentioned, you know, once chat GPT hit, everything changed. We're in this agentic AI moment. And from agentic, we go into physical AI. At least this is how NVIDIA looks at where the curve, the AI curve is going.
So how many of you are here actively deploying agents? Just for me to kind of understand. So, perfect. We're using Claude, Cowork, probably OpenAI, all these different things. And a lot of that is eventually gonna bleed into physical AI. That is a term that you're gonna continue to hear across the technology space. Physical AI, physical AI will be one of those terms. Because in this era that we're leaving, you know, LLM can predict the next word, but eventually we need physical AI to predict the next action. Intelligence, we need intelligence that can perceive and understand the physical world.
I think it's really interesting that Mohamad was mentioning the robots, right? They had, I can't remember the number of robots they have, but he mentioned a human in the loop. And I'm gonna get into that, because right now most of these robots are not fully, fully autonomous. There's still a human in the loop, and this is kind of where this paradigm shift starts to happen. So obviously we've created a lot of amazing things with generative AI. Our agents right now are making a lot of decisions for us, some smarter than others. There's been obviously a lot of media coverage on some agents that have gone rogue or have deleted everyone's emails or things that shouldn't be happening. But AI is deciding to do things for us currently. Right now, that shift from deciding is into existing in the physical world.
And this is what I call the physical AI stack. It's just kind of an idea of some of the things that are coming down the line when it comes to AI. So in order for these agents that we're all using to eventually do things for us in the physical world, they're gonna need a body, right? And that body could be a human or robot, it could be a hologram. I think a lot of people don't really think about holograms too often. Everyone's really obsessed with the human or robots, and I understand why. But also those AI agents need to be embodied and they might be in holograms. So if you're a Star Trek fan, Star Wars fan, you probably are excited and waiting for the holograms to come.
We're also gonna be talking about the new interfaces. Everyone is using their mobile phones, but everyone in Silicon Valley and across China and different tech companies are all running to replace your mobile phone. And a lot of people will think smart glasses, but there's a lot of different options. We're gonna talk about that. Also the importance of connectivity and how the conversation is eventually gonna move from compute to connectivity. I am gonna challenge you guys a little bit. The way I view agentic AI, I see it as kind of the nervous system for this next era that we're going into, right? We're gonna need a body and interface and a nervous system for this to work together. And I think that's where agentic starts to come in, more so than just an agent that can, you know, write emails for you or recruit people. We're actually gonna need that nervous system to connect all this intelligence.
So, wanna talk a little bit about world models. How many of you here have heard the term world models? So very few. Anyone working on world models? No, not quite? So world models, and I'll just start by saying this, world models is where a lot of the attention is starting to flow when it comes to AI. So Dr. Fei-Fei Li, if you know Dr. Fei-Fei Li, she is known as the godmother of AI. I'm the godmother of the metaverse. I know the metaverse isn't cool to talk about anymore, but I feel some sort of sisterhood with her. But the godmother, yeah. But the godmother of AI is working on a company called World Labs, which she's working on world models, right? She, and I will get into what world models are. I'm just kind of painting the picture of where the funding and the attention is going.
You also have Yann LeCun, who used to be the chief AI scientist at Meta, leaving Meta, a lot of potential drama there, lots of stuff written about that, but he left to create a company called AMI working on world models. And if you have not been following the news, he just raised, get this, his seed round, 1.1 billion seed round. 'Kay, that doesn't even sound like a seed round, I'm sorry, but $1.1 billion seed round for world models. Right? Why do we need world models? Why is all this funding going into world models?
And I will also mention some other companies that are getting into world models. Waymo, Waymo is referencing what they're doing as world models. Roblox is getting into world models. Runway, which does video, is now talking about world models. Even Goldman Sachs has a report on the importance of world models that just came out a couple weeks ago. It's actually pretty good. But this is kind of where a lot of the attention and the funding is starting to go.
But why? Because unlike LLMs, world models do not predict words, they actually predict the next frame of reality. Right, so in order for those robots to not have the human in the loop that Mohamad was mentioning, they need to be able to understand the physical world. So if I tell you right now, I think a lot of us usually think of robots in the manufacturing space, but if I told you you're gonna have a robot at home taking care of your children or your grandbabies, would you trust that robot right now? No, right? Would you trust the robot in the kitchen with a knife? Probably not, there's too many movies, right? So I think it's important to start to think about why world models are important.
And what world models pretty much do, it's AI that is able to simulate the physical world, right? They are able to train these agents in virtual spaces to eventually understand the physical world. So it gives them a way to understand physics. Dr. Fei-Fei Li again did a wonderful TED Talk about I think two years ago. I recommend if anyone's really interested in what's happening with spatial intelligence and world models go look at it, it's fantastic. And she talks about spatial intelligence. We as humans when we're born, it takes some time for us to learn how to walk, takes some time for us to go up and down the stairs and learn the physics of the space, right? Or learn, a little kid learns that eventually if he puts a glass of milk on the table and he swats it, doesn't put it right here, it'll fall. That is something that we eventually learned. There's a lot of kids spilling sippy cups and all these sorts of things. That's why we have sippy cups, right? But basically, what we're training through world models, we're training the AI, so the agents, eventually the robots that will embody those agents, to understand the physical world and eventually be able to do all these sorts of things.
So world models in practice, it is a foundational shift. These are very different models. I am of the mindset, and I think a lot of people in the AI space are of the mindset that we cannot reach AGI, if that is what we want, or eventually superintelligence, that we cannot reach that with just large language models. You need, yeah, and I see people nodding. We need new models and new architectures. That's why LLMs are not enough. This is super important for anyone here. LLMs still hallucinate. That has not been solved. You cannot have a robot or billion, let's say you have millions of robots everywhere, you cannot have robots or you cannot have agents or holograms that are hallucinating on a constant basis. Once again, it goes back to trust. So these world models are gonna allow the AI to actually understand physics over language. That is critical. For example, what Dr. Fei-Fei Li is doing is using a lot of these virtual worlds to train the AIs. So the agents are being trained inside virtual worlds to do things, to do tasks and all that. That's eventually gonna translate into the physical world. So think about world models as a way to simulate the physical world. That is how you start to teach AI to understand the physical world in a much stronger way.
This is something that most people out there, and it baffles me that no one's really talking about this, is geospatial models, right? So you've got world models, that's where a lot of the money's going, and then you got geospatial models. This is the layer that's gonna connect the AI to the actual world. It is gonna enable AI to reason about place, terrain, movement, and the built environment. So it's AI that understands language, you know, AI that understands language is powerful. I think it's changed a lot of our lives or our workflows or what we do, but AI that understands place, location, and destination is transformative. There are several companies in this space working on geospatial AI models that are starting to get the attention they deserve because sometimes you might deploy some of these robots, you know, drones or whatever, into places that might not have access to GPS or there might be certain issues there.
When it comes to geospatial models, and by the way, I'm just gonna kind of, I don't wanna get ahead of myself, you know, I might go up one slide, a couple slides. Let me go here. Just for frame of reference for people to understand when we're talking about geospatial models. So you've got the large language models, which are words, predictive patterns, all the usual suspects that everyone's using. World Foundation models, which is simulation predictive behaviors, training a robot or an agent to do something a million times in a virtual space before it actually hits the road or goes to the factory floor or eventually comes to our house or to our hospitals. That's World Labs, NVIDIA Cosmos, Google DeepMind. And then you have large geospatial models, which is predictive spatial mapping. That is one of the missing pieces that people are not talking about, and that's where companies like Niantic Spatial, Maxar, Google's also getting into this.
If you think about that, for example, Google as a company, many of you use Google Maps. Google Maps knows where you're going. They also have your Google Calendar. They know where you're going, where you're navigating, but they also know where you're gonna go next. Right, so that's a lot of really interesting data they're gonna have access to. Waze or Apple Maps, that's another reason. And we are gonna talk about hardware, but I think it's very relevant to mention in the spatial frame of what we're talking about that Apple is gonna be doing AirPods with cameras. So it's not just about spatial audio, it's cameras that are gonna be mapping the physical world on a constant basis. That's gonna give them even more data for that geospatial, for those large geospatial models. So you're gonna see the race, you know, let's say the hype and the money going from just LLMs, which they're still gonna raise tons of money, into world models and large geospatial models.
One of the reasons that this is happening is because one of the data moats is in physical world data. That type of data, any company that is doing anything related to the physical world, whether it's in the factory, in the manufacturing space, physical world, whatever, that is a treasure trove of data. That is where certain countries that have been tracking all their citizens and doing a lot of things are gonna have a lot more data than other countries are gonna have.
I will mention this, and I'm gonna go back a couple of slides, to the defining the spatial stacks. So I am considered one of the world's top experts when it comes to spatial computing. I've worked in hardware for a very long time, both VR hardware, spatial computing hardware, large-scale gaming simulation. I wrote a book on spatial computing probably three and a half years ago. It did not do well. And you know what's happening right now? Everyone's buying the book. It's crazy, it's a bestseller in China. I'm like baffled by this. I was too early, apparently. I'm a futurist, so I got there too early.
But for people to understand spatial computing, it is not a headset, 'cause I think people are so obsessed with thinking it's VR. No, spatial computing is an evolving 3D-centric form of computer that at its core uses AI, computer vision, extended reality, and other technologies to seamlessly blend virtual experiences in someone's experience of the physical world. It combines software, hardware, data information, and connectivity. It's not just one thing, it's computing. It's the evolution of mobile computing into spatial computing. Then you've got spatial intelligence where a lot of these world models are working, and then spatial AI where the world models are also, where the geospatial models are heading.
One of the reasons that we need not just large language models and world models is, and one of the reasons we need geospatial models is because there's a ground truth advantage. This is a word that Niantic uses, and I think it's very, very favorable, because real-world data is irreplaceable. So synthetic data, you can scale that indefinitely. So that's a lot of what world models are doing, is doing a lot of the simulation, right? But as AI moves from digital reasoning to physical action, you need that real world data. So there is an economic case here for real world data to take us to that physical AI moment.
All right, wearables, smart glasses, new hardware. This is once again where I've spent a lot of my time. Anyone here using the meta Ray-Ban glasses? Okay, I'm in person. I usually find like a couple of people whenever I go to the talks that are using them. So there are people starting to adopt this, right? There are millions of people across the globe. There are actually a lot of people in China also starting to use a lot of these devices. So I think we, a lot of people are obsessed with thinking it's going to be smart glasses. I personally think it's going to be a combination of form factors. Those AirPods with cameras, you know, do you guys remember when AirPods came out? People were like, I'm not going to wear that. They look like earrings on men. Like now everyone wears AirPods, right? So the idea of wearing AirPods with cameras, people are like, I don't know if I want to do that. That seems like it's going to invade my privacy. I think it's gonna, you know, I think that's already happened, but I think there's gonna be a lot of different form factors.
OpenAI, for example, is working on hardware, right? Everyone is starting to look at hardware. I get asked, like, why is Google going back into the game? Google Glass failed. What is, you know, why is Apple gonna do Glasses? This is why. Because the hardware is the new moat. Hardware is hard. As someone that has worked in that space, it's really hard. But the moat is moving from the models, right, especially the general purpose models, into hardware because whoever has access to that hardware, whether it's, you know, devices, pendants, AirPods, whatever it is, that's going to be scanning the physical world on a constant basis, right? Those cameras are going to be telling, gathering all the type of data of what's happening in the physical space, and that is where there's going to be a lot of interest in that data.
So, you know, why are we going into new hardware? Because we are going to move AI and these AI into devices that we can actually interact with. Right now we're mostly doing our phones and we're doing mostly our computers, but eventually you might want to talk to your agent through your glasses or through a pendant, or you want your agent to tell you where to go, all these sorts of things. So it's going to move, the interface is going to move from just our phones and heads down, which I think any parent of a teenager, right, I've got two teenagers at home, you don't want them looking down. If they, you know, eventually they'll be looking up and there'll be data in front of you, but we're moving more into the heads up and on you. Anyone here wearing an aura ring as well? Maybe a couple of you? That's another reason that wearables are so important is because of the biometric data that they collect on you. So there's gonna be tons of data, I mean once again it's data collection, right?
So we're gonna move from the cloud to the edge to on you. That is why you're gonna start to see more and more companies like Nvidia, T-Mobile, Nokia, and all these sorts of companies work on what's called AI radio access networks, AI RAN, to bring the compute, the connectivity, not so much the compute, the AI compute, but so much more the connectivity closer to the user and on the user. So it's called on-device intelligence, and that'll be one of the reasons that this is gonna happen.
Game engines. I think most people think of gaming and they think of maybe a teenager playing Minecraft or they think of a 50-year-old man in a basement. One in three people in the world is playing video games. Okay, that's a huge number of people. People don't consider themselves gamers. I usually don't even ask this anymore in a room, like who's a gamer because it's mostly men that are gonna say they're gamers. A lot of women qualify themselves as gamers, but a lot of these women are playing Candy Crush, or playing different games, they are gamers. They just don't give themselves that label. But one in three people in the world is playing games. And if you look at the younger demographics in the Middle East, for example, MBS, the real crown prince in Saudi, is actually a huge gamer. So there's a big push for a lot of gaming.
Gaming is not just entertainment, it is infrastructure. So when you look at world models and where these world models, the agents are being trained, this is all driven by game engines. So game engines were built to create worlds and sometimes simulate reality for entertainment. If you're playing one of those zombie games in the desert, you want the zombie to pop up in the right way. But now they're actually being trained, they're being used to train AI to operate in reality. That's where Unreal Engine, Unity, NVIDIA Omniverse, all of these come in to allow that training to happen. And it's becoming important and critical infrastructure.
This is from Roblox. If any of you guys have kids in your lives, yeah, uh-huh, we all work for Roblox, Minecraft, or Fortnite. That's where all our money is going 'cause the kids are not coming like, mom, I want $20 to go to Target. No, I want Robux. Yeah, and this is the funny part, and I wasn't gonna mention this, but with my children, I remember for one Christmas, one of my sons was like, what is Auntie so-and-so gonna send me? And I said, $40, whatever. And he immediately did the conversion to Robux. He can't do that with Euros. He can't do that with South African Rand, but he will tell me exactly how much Roblox that is and he will spend it immediately. But this is a really interesting thing, especially when it comes to the gaming component, that a lot of these younger kids are gaming native. This is their social network.
So you've got Roblox, for example, just last week announcing what's called Roblox Reality, which is gonna combine, what their game engine is really, really good at doing is creating games, you know, and creating all these sorts of games, but it doesn't really understand the physical world. They're starting to do video world models to combine these two to allow developers to create better and better worlds. This is not available just yet, it's coming, but once again, one of those things, you're starting to see Roblox get into world models.
So there is an infrastructure shift that's happening. We're gonna, you know, physics is gonna simulate, it's gonna simulate at scale. We still need synthetic data generation. There's gonna be even more advancements in digital twins and world models. Obviously a lot of healthcare work is being done right now in using virtual surrogates to test treatments and see how things might pan out for someone.
Humanoid robots and holograms. This in my perspective is definitely the most visible but most misunderstood expression of physical AI. I got into actually a really interesting debate with a friend the other day on why do robots have to be humanoid, right? He was like, well, 'cause the physical world is created for humans. I'm like, are we doing it to make ourselves feel more comfortable? Because I don't know if that's the most efficient way for a robot to perform, right? Yeah, if there's stairs, they have to go up the stairs and they need two legs. But I think there's a reason we're creating these robots to look like us. Jensen Huang said this not too long ago. He said humanoid robots are going to be the next consumer market. That's a pretty interesting, bless you, pretty interesting thing. They're not gonna be pre-programmed machines, right?
At the end of the day, if you're gonna have millions if not billions of robots or vehicles, you know, or delivery robots or drones or devices or holograms, all these sorts of things, you can't continue to have a human in the loop with every single one of them, right? So that's why training them to understand the physical world will be important. Then also holograms. You're gonna start to see a lot more focus on, especially once you have the AirPods with the cameras, being able to scan the physical world, and then obviously using the phone as your camera as well. You're gonna start, like, the idea that you're gonna hologram with your grandbabies in the future, right, my kids are little, so I think I always think about that. I'm eventually gonna do hologram calls when they're married and have kids. So that idea, I know it sounds really sci-fi, but it's a lot of people have been working on this.
Intelligence is entering the physical world, and the digital presence that our agents have needs to be embodied. They need a body, and whether it is the hardware or the interface, that is where this is gonna come in. Oh, I want to make a note here because I also think that when we talk about the labor shortages, whether it's in manufacturing, whether it's in nurses, there's also an imperative to get some of these devices to actually work in the physical world if we cannot find enough humans to do the jobs.
So when you're building the post-LLM enterprise, here are some of the things you guys can consider as you're leading your companies. I'm not saying you need to do all of these right now, but just some ideas and some thoughts on questions you should think about. If you have any type of real world data, geospatial data, how are you utilizing it? You're obviously piloting a lot of digital pilots with agents, all these sorts of things. How do you start to think about, depending on your line of business and your company, pilots in the physical space, potential and infrastructure audit, also talent. You know, back in, like what, a year ago, two years ago, people were hiring prompt engineers like crazy. You should also be thinking about who in your company knows how to use game engines, who knows about geospatial data, spatial computing, all these sorts of things, and connectivity.
The connectivity strategy is gonna be incredibly important. This actually hit me, recently, what was it, a couple months ago before the war, I was in Riyadh. I do a lot of work in the Middle East. And I was having a conversation in Riyadh with one of the men that helped create the AI strategy for Saudi. And he said this to me, I did not prompt this. He said, I believe that as compute starts to get cheaper, obviously there's certain constraints right now, but as compute starts to get cheaper and some of these models become more efficient, that the conversation will move to connectivity, right? Because when you're talking about millions, if not billions, I always go back to this, of devices, robots, glasses, I mean, we don't currently have the connectivity needed to be able to power this physical AI future. That's why Nvidia starts to invest in a Nokia. That's why all these, you know, all this connectivity isn't gonna be incredibly important. 5G's not gonna cut it, I will tell you that. Because at the end of the day, once again, if you're gonna trust this robot in the kitchen, it needs to know what to do with a very, very small amount of failure.
So as a leader, here's some of the things you can think about. This is a great slide to take a picture of, right? We're not gonna go through all of them, right? But it's about starting to think about what moat might be moving for your company? What type of data do you have access to? If your customers are eventually gonna adopt smart glasses or new devices, how does that change the services that you might be providing? So just some of the questions to ask yourself as a leader.
How will this work in the physical world? I want you to kind of think about what I just shared with you. I'm not gonna paint the picture for you, but I want you to think about how this could work in the physical world, right? If there's gonna be robots, delivery robots, drones, holograms, all sorts of new devices in this future, let's say five to 10 years into the future, how would that work in your company? What would that mean for your employees who comes to work? What are they coming to work wearing? Because they're gonna wear a lot of these devices. And how does that change the services and the services and products that you offer? So the infrastructure is being built right now for a lot of these world models, geospatial AI and hardware, the interface is slowly arriving. Don't kid yourself. There is a mad rush in Silicon Valley to replace the mobile phone. And the moat is slowly moving. Thank you. That was a lot. Sorry, guys.
That was terrific. That was terrific. It was just really good to hear some refreshing talk around because we've just been talking about how hard it is to achieve AI outcomes and all that stuff for the last 36 hours. CEOs of all the top AI services and firms here. So maybe it would be good to sort of throw it out to the audience before I ask you a question. On maybe Michael over there. You wanted to ask Cathy something?
Yeah. I was going to say, if you could just raise the point that AI has made, so that the geospatial program, right, and doing enhanced reality, was expensive beyond belief two years ago. And now it's gotten, and I was looking for all kinds of developers, and we were all at the Naval Yard over here and all, and they were each working on their own very parochial kind of stack. And charge a ton and you could never really complete anything. That's somewhat gone right now. That's a huge benefit of AI. So we can start to think of more innovative solutions. So you hit a bunch of them. We were looking at, you know, blockchain for housing, tied to the energy grid, tied to your robots and everything else, and how do you design that, which now people understand, oh, we can do that. So thank you for that, by the way, because I read your stuff all the time.
Thank you.
But the fact that you've made this a reality, people need to understand, this is not really five, 10 years, and kids are gonna get this right away?
Yes. Yeah, they absolutely do. I feel demographically, one of the reasons I do so much work in the Middle East and like in emerging Asia and then South Africa is because demographics are a lot younger, right? And there's a lot of gamers in that space. So they intuitively understand this. And you mentioned blockchain for property. I mean, there's so many interesting deployments that are happening right now with tokenizing, this is a different type of token. Not that token maxing, tokenizing property. But I think we're gonna start to see, one of the other things I focus on is convergence, the convergence of all these technologies kind of coming together and how that pushes us forward. I also think with a lot of big tech laying off amazing talent, that also gives you an amazing opportunity to tap into people that might know, you know, spatial, that might know gaming, that might know all these sorts of things and bring them into your company. So, yeah, I think there was a question over here too. Oh, there's more. Yeah, I'm like, if no one had questions, I'd be concerned. So, yes.
Hello, yeah, hi. My question is not so much about the technology. I think this is fascinating. It's borderline sci-fi, if not actually sci-fi. And it's very exciting to hear about all of this. But side by side with this, we need to also have considerations of responsibility and ethics. You haven't mentioned any of that. In fact, I haven't heard any of that in the last couple of days at all. Where does that, what does that look like? What do we have? What should we have? And my concern is not so much about whether the machines are going to take over the world. I'm not so worried about that. I'm worried about the humans who control the machines. I'm worried about the ethics of humans, not of the machines. So if you could just tell us your perspective on, you know, what that whole area, is it developing alongside the technology? Is it a step function behind? Is it five step functions behind? Does it even exist at all? What's going on there?
Here, I'll tell you two things that I think are very relevant is when people ask me, am I worried about AI? I'm like, I'm not so much worried about AI. I'm worried about the companies controlling the AI and the incentives behind it, you know, the financial incentives. That's one of my biggest worries. With that being said, when it comes to physical AI, new hardware, all these sorts of things, what keeps me up at night, and it does not, you and I, I think, have talked about this in your podcast, it keeps me up at night, does not keep almost anyone else in the world up at night, is the concept of virtual air rights, who owns the air around you, because when you're talking about these new devices, right, if they're in front of your eyes or they're in your ears, what, you know, yes, the world becomes the canvas, which is the positive part, but it also becomes real estate. So who owns what is within eyesight and earshot of you? Who controls that right now? Because if I go to Times Square, you know, I can, like, look away from the billboard, but if someone's controlling what my eyeballs are seeing or what I'm hearing, all these sorts of things. That is another level of issue with privacy, right? Is anyone working on this? I think a few people are trying to work on this. I am personally having conversations with governments across the world about this, and I think some of those governments that are more future forward are thinking about do we need to create, and this is not in the immigration sense, sanctuary areas where there's no content being put overlaid over your reality, right? This is more on the sanctuary, what I mean by sanctuary. So I think there are efforts and there are conversations and there is interest. There are several companies that are starting to try to work on this, some of them in blockchain, for example, but I think that is gonna be a very, very important conversation that we're gonna need to have. And sometimes I do mention it in my talks, but yeah. Thanks for asking though. Yeah.
People used to try and control us through traditional media, newspapers and TV, and now they're trying to control us through many different other things. I've even heard about GPS apps now giving you routes home, and you don't even realize they're taking you past your favorite fast food restaurants and things.
Sending you on that route because Chick-fil-A is right there.
Exactly. They know exactly what you want. Sorry, next.
Yeah, just to add to what Hemant said in the front there, you know, I'm probably worried more about not humans impact on these gizmos or whatever that will control the world. You know, I'm worried about the opposite, actually, right? Because we'll reach a point, I think it's a tipping point, where we will take for granted what we learn from machines, right? I mean, we'll just assume it is true, which is probably not a bad assumption because humans themselves are not very good at being consistent over a period of time, right? So what's your, you know, and what are your thoughts on that? Also, the other thing I want to say related is that we as human beings have been very bad at predicting the impact of anything on our lives, okay? So should we even worry about it or should we just let it evolve? Okay, that's my question.
I mean, you know, move fast and break things. I don't know if that's served us really well at all. I do believe in having these difficult conversations. See, I'm glad these questions are being asked now instead of after we deploy all these gizmos and devices. When people ask me a little bit about this, I say I think the future is less Ready Player One where you're putting on a device to escape because the physical world is so horrible. And to me, it should look like or I want it to look like where I think certain people are working for it to look more like Jarvis in the Marvel movies, right, where you're actually collaborating with the AI, whether it's through a device or some type of, you know, you know, spatial interface in front of you where you're actually collaborating with the AI. I do think, and I will say this, this is what I'm seeing with the younger generation, like my kids, they're Gen Z, Gen Alpha, is that they like using AI for certain things, but they are revolting against AI and other things. So I think that also from a generational perspective, I think we're going to face some challenges there where some of the younger generation is going to be like, this is not okay. This is not okay. So I'm not sure. I'm not sure. But the conversations are being asked and some people are asking the questions.
Someone over here.
Cathy, very interesting conversation. How do you see physical AI showing up in the corporate world? Are there any specific samples that you see in the future happening?
Yeah, I think especially you're already seeing it, I think, more in the manufacturing side. That's always been. Here's something I'm going to say when it comes to physical AI on the hologram and device side is that the consumer side of the metaverse, okay, no one needs, you know, legless avatars. I think everyone agreed on that, okay? So not the Zuckerberg version of the metaverse, but the industrial metaverse work that was done inside companies, that's how Omniverse started, right? NVIDIA's Omniverse. A lot of the work that has been done in the industrial metaverse is now being translated into physical AI has been very powerful. So I think you're already starting to see, you know, there's case studies, there's many case studies about how some of the work done in the industrial metaverse translates. When it comes to corporate, I mean, we'll have to see how it all rolls out because right now we still need, we still need a lot of things to be solved for true physical AI to be deployed on a more corporate level, right? Hallucination is number one. Connectivity, so there's different things. So happy to connect with you, maybe paint a picture of what I envision it to be. It's kind of a longer picture. But yeah, I think a lot of things need to be solved. We're not there with physical AI, right? Not there yet, so. But I think it's where a lot of these companies, especially after Nvidia showed that graph, everyone's like, okay, physical AI, that's all I wanna talk about, so.
Okay, thank you, Cathy. That was a fascinating stitching together of the future. I have a question on your timeline, and how two things fit into this. AGI and quantum. Is this scenario you're painting pre or post?
This is all pre-quantum. Once quantum hits, oh goodness gracious. And I am not a quantum expert, I will tell you that. I am definitely not a quantum expert. I do think quantum's gonna obviously change a lot of things from an encryption standpoint, from many different perspectives. I do think with the quantum conversation, I think people are really focused, I don't know why we do this as humans, but quantum computing, I was like quantum computing, but within the quantum space, there's quantum sensing, there's quantum communications, there's other things that are not necessarily computing that I think are gonna have impacts. So most of what I am talking about here is pre-true quantum breakthroughs, right?
AGI?
Oh, AGI?
Is this pre or post AGI?
Oh, I thought you were talking about quantum breakthrough, both, oh. I think, okay, here's my big beef with the AGI term is that I think even people in the AI space don't have, I think we all agree on a definition, but we don't agree on a definition, right? So I think that remains to be seen with truly achieve AGI, right? And then you throw in super intelligence. So I prefer the term singularity, to be honest. So this leads us, once we get into true physical AI, truly believe we have singularity, you know? So I prefer the term singularity.
Do we have time for one more? We can do one very quick one. One very quick one. I'll ask the final question. I'm a big fan of quantum computing. I wrote a paper on it with Joel. I don't know where Joel's gone. Where's Joel? A couple of years ago. Here he is. He described it like a shoal of fish, and the energy was like this, and one fish falls out of the shoal, and the whole thing falls over. I thought it was very clever. When are we actually going to really start to see quantum? Come on.
I mean, we've been talking about it for a long time. Oh, gosh. I don't know. Like, it depends on who you talk to. Let me answer that question when I get back from China. Because I think they're doing some really interesting things there. At least I hear they're doing interesting things about some quantum breakthroughs. I'm not sure. I'm not sure. You know.
I'm not excited about the new AI deal they'll be signing out there today. Can you just imagine? Anyway.
I will say, I think I'm starting to see some really interesting advancements in new materials like graphene. I know people like haven't heard that term in forever. Like if anyone's heard the term graphene, they're like, what? And it's a 3D material, et cetera. But I think we're starting to see kind of new advancements in graphene that will probably push quantum forward. So yeah, I don't think we're getting the graphene space elevator anytime soon if anyone's into sci-fi. But yeah. Okay.
Yeah. Well, this has been wonderful, Cathy. So glad we got you here.
Thank you. It was a pleasure being here. Thank you, Phil.
If you don't have an account, Register here |
With the exception of our Horizons reports, most of our research is available for free on our website. Sign up for a free account and start realizing the power of insights now.
Our premium subscription gives enterprise clients access to our complete library of proprietary research, direct access to our industry analysts, and other benefits.
Contact us at [email protected] for more information on premium access.
If you are looking for help getting in touch with someone from HFS, please click the chat button to the bottom right of your screen to start a conversation with a member of our team.
