#336 Byron Boots - He Turned a Polaris RZR Into a Self-Driving Military Vehicle

If you've listened to this show for a while, you know I'm a big believer in functional mushrooms and adaptogens. So mud water made sense to me right away because it gives you that morning ritual without the full caffeine hit. Their original blend is made with cacao, chai, turmeric, and functional mushrooms. It's got about one-seventh the caffeine of coffee, so you get focus and energy without feeling jittery or dealing with that crash later. And what I like about the starter kit is it actually gives you everything you need to start the routine. You get a full size 10 of your chosen blend, plus a rechargeable frothter, free gifts, and free shipping. The OG blend is where a lot of people start because it has that chai and chocolate flavor, but they also have coffee, matcha, and turmeric blends if you want something different. For me, mud water is just a cleaner way to start the morning. It feels steady. It tastes good, and it fits right into the kind of routine I already want to be building. Ready to make the switch to cleaner energy? Go to mudwater.com/srs and grab the starter kit. Use code srs and you'll get 43% off. The frothter alone is worth it. That's right. Up to 43% off with code srs at mudwtr.com. After your purchase, the last how you found them. Please show your support and let them know we sent you. Most guys will spend money on gear, training, supplements, everything else, and then still wear the same uncomfortable underwear all day. If you're moving, traveling, working out, or sitting for hours, you feel it. That's why I like sheath. What makes sheath different is the dual pouch design. It keeps everything separated, supported, and comfortable so you're not constantly adjusting or thinking about it. And I like that sheath was originally built by an active duty operator who needed something that could handle long, hot, miserable days. So it was built to solve a real problem, not just look good in packaging. They're soft, breathable, and built for daily wear, travel training, or long days on set. Feel it to believe it. Go to sheath.com/srs and use code srs for 20% off. Sheath offers a first pair of guaranteed. So if it's not for you, you get your money back. That's s-h-e-a-t-h.com/srs and use code srs for 20% off. Sheath. The underwear of legends. Thanks to Sheath for sponsoring the episode. Fire and boots. Welcome to the show. Thanks for having me. It's awesome to be here. It's awesome to have you. I can't even remember. I think I actually found you guys on LinkedIn, which I've never on. And I think I saw like a video on LinkedIn or something of Overland AI and then started found you guys on YouTube and started following you and then found out APC is an investor. So or led you around and so yeah, I wanted to get in touch and love what you guys are doing. You know, I've had lots of tech giants on here, a lot of drone stuff, Serone, I put the water stuff. I guess they I didn't even know your guy just told me I'll back that they just hit the they hit up. I ran any import. Yeah, pretty, pretty amazing. So you know, first of all, it's an honor to be on the show and in that company, I mean, that's that's incredible. But yeah, just in the last day or so, there was news of Seronex boats being used in offensive operation in the state of hormones. Wow. And so you're the ground guy. That's right. You're bringing autonomous vehicles to land warfare. So really excited to get into to dig into this. But let me let me kick it off with the introduction here. Byron Boots, you're the co founder and CEO of Overland AI, a company building autonomous ground vehicles for the US military. Before founding Overland AI, you earned your PhD in machine learning from Carnegie Mellon became a professor at the University of Washington and led the winning team in DARPA's racer offroad autonomy program. Overland AI has raised more than $140 million and become the first autonomous ground vehicle company to win a production contract with a fully integrated hardware and software platform. Congratulations. The vehicles are already being used by the military units around the world to move supplies, support operations, and reduce risk for soldiers in the field. Welcome to the show. Well, thank you so much. Got a lot to talk about here. It's been a minute since I've talked to somebody that's doing the kind of stuff that you're doing. But before we get going, got a couple things to crank out here. Everybody gets a gift. Oh, wow. Those are, uh, thank you, autonomous dummy bears. I'm excited. I love gummy bears. So right on, right on. And, um, and I got a, I got a question for you. I got a Patreon account, the subscription account. And, uh, so they're the reason I get to sit down here with you today. And so they get the opportunity to ask every single guest a question. So this is from Thomas W. We've dedicated your career to advancing robotics and AI, including technologies with defense applications. From your perspective, what responsibility do scientists and engineers have to ensure these innovations ultimately reduce human suffering rather than prolong conflict? And do you believe AI in autonomous systems could one day become tools that prevent wars through deterrence and de-escalation, or do they risk making armed conflict more frequent and easier to justify? Oh, that's a great, great question. You know, I think, uh, I think autonomous systems, they're like any other tool. The way that I think about them is, is really, um, a tool, a technology, uh, and in the context of defense. Um, it is, you know, something which allows a war fighter to, uh, you know, be safer, right? So, you know, it reduces exposure, uh, pulls them away from the point of contact, um, and then also potentially, uh, provides, uh, force multiplication on the battlefield. I'm sure we'll, we'll get into, um, some of these things. But, um, it is a tool which is used by humans and, so, you know, how you use them, uh, is really, I think, a, a human question. Um, so, you know, in, in, in the context of, uh, saving lives, I think that they, uh, will save lives for, um, our war fighters, uh, on, on the battlefield, um, it's, it's very clear how they do that, um, but, you know, they also conserve as a deterrence, like any, like any technology might, um, you know, if, if we have significantly stronger, um, robotic systems out of our disposal, then, you know, we're going to be, uh, a little bit, you know, tougher to, to defeat in the battlefield, and our adversaries will see that. And so, um, in, in that way, you know, they can certainly serve as, uh, a deterrence as well. Right on. I mean, yeah, watching some of the models you showed me outside, and then, um, you know, the, the, the videos and, and, and, and how they're going to be integrated in with the war fighters in, in combat, I mean, it's, uh, you know, is a, is a, is a former seal, who's like, seeing what, seeing the, you know, award, what war has developed into is just, I mean, it's, it's fascinating. And I mean, just, it's been over 20 years since I've been on the ground in a war. And, uh, well, I guess not, but so been over 10 years. But, um, I mean, I already have tons of questions and, and I can see so many different applications where this would be, uh, useful and just so many different scenarios. Yeah. We should, we should definitely get into it. Before we do that, though, um, I do want to give you a gift as well. Oh, I love you. So, let me, uh, let me come over here and, um, we got you a chair. Now this isn't just any chair, right on, um, this is, uh, it looks kind of like an office chair, but this is actually a seat from one of the, um, vehicles. So as I was explaining earlier, we pulled the seats out of, out of the Polaris range, and we turned them into those autonomous vehicles, um, that we, that we saw outside. Well, what are you doing with the seats once, once you've, uh, pulled them from the vehicle, we make them into chairs. So we made one, uh, for you. Yeah. Thank you. You can check it out. But it's awesome. Yeah. Yeah. Not bad. Not bad. There you go. Put this in the office. Thank you. Right. Yeah. Of course. That's awesome. All right, Byron. So before, before we get into everything overland AI, how did, yeah, let's do a little backstory on you. Where did you grow up? How did you get into this stuff? I mean, what's the backstory here? Yeah, I, I had, I had a whole career, uh, before, um, moving into, into defense tech. So, um, I grew up outside in New York in, in Connecticut. Um, uh, I was into computers and, and, uh, you know, was in the Boy Scouts and played sports and added, I think a pretty typical, um, upbringing. Um, so, you know, that's maybe where things got started. I just love the outdoors and taking apart computers and playing video games and doing all the sorts of things that kids often do. Did you watch the Terminator grown up? I sure did. Yeah. Good job. So Terminator 2 was a like unbelievable, unbelievable movie. You know, and happy to talk about that a little bit more in the context of what we're building, but obviously robotics and science fiction were, you know, something that I really enjoyed. Were you a gamer? Yeah, I used to play, so back when I was in high school, I used to play Starcraft quite a bit. Starcraft 1, as before, Starcraft 2 came out. So real-time strategy games, I did a lot of, played a lot of games like that. We talked about Warcraft before, you know, I used to play that too. So that was really what I was most drawn to, but yeah, I mean, love computer games. How, I mean, well, we'll get into it later. I was going to ask how similar, you know, is what's happening today is controlling one of those games. We'll get into that in a little bit, so where did you go to school? What did you go to school for? So I went to college at a small liberal arts school called Bowdoin College. It's in Maine, and, you know, I spent four years there. I was a computer science and philosophy double major as an undergrad. I started out really thinking about, you know, I love computer science, like I said, you know, all through high school. I also really liked history. I read a lot of history, and so when I went to college, I was thinking about maybe double majoring in computer science and history, I thought it would be cool to have a more technical degree and something which is, you know, more humanities oriented with history. But I quickly, you know, sort of figured out my first year that, well, I love history, it involved tons of reading and things like this that I really like to do, but also involved foreign languages. And you need to actually read about history through contemporary sources and, you know, in the language that folks wrote in. And so there's something I was not great at. I was not particularly good at languages, didn't really enjoy them. And so, you know, I started to, I took a couple courses in philosophy. I started out with a course called Logic Informal Systems, which was really delving into formal frameworks for understanding arguments and analyzing arguments. And I really started to fall in love with that. And so, you know, I brought that together with computer science and major in both of those areas as an undergrad. Ventures like, you're fascinated with the brain too, right? Yeah, yeah, yeah, so, I mean, in undergrad, I was, you know, I was taking computer science and, you know, took courses in artificial intelligence and started to do research in robotics. And this is back, you know, over 20 years ago. They had artificial intelligence courses 20 years ago. Yeah, it was pretty interesting. So my university is a small school. There was only four faculty in the computer science department there. And at the time, computer science was seen as like an offshoot of mathematics. And so, you know, a lot of these smaller schools had combined departments with computer science and mathematics. And you take a lot of courses in both areas. But at Bowden, two of the four professors were actually folks who studied artificial intelligence. So it was something, you know, it's been around for a long time. I mean, people were working on aspects of AI, you know, back in the 70s and 80s. But it was just starting to kind of come to the forefront and be an area that was really starting to accelerate around, you know, 2000 when I was when I was an undergrad. So I started to study AI there. And then in philosophy, I was thinking about things like philosophy of mind and philosophy of science. And you're just kind of trying to understand how the human mind worked. Well, yeah. I think that's an interesting discussion itself. Yeah. I mean, how deep did you get into that before you kind of switched gears? Yeah. So as a, again, as an undergrad, a double major in computer science and philosophy. And when I graduated, I was, you know, thinking about what I wanted to do next. I initially took a job at a robotics company as an engineer, basically working on problems related to perception and mapping and robotic systems. So these were mobile robots much smaller than the ones that we just saw outside. So robots that are about, you know, about this big, that moved around inside of buildings. And you have to determine where they are and how to get from one place to another and things like that. So I worked as an engineer working on those sorts of problems. But I was really thinking about, you know, kind of like, what do I want to do next? I knew I didn't want to just be, you know, kind of working as a software engineer. I wanted to go back to school and the question was, like, what area should I study? So really like computer science. I really liked philosophy. And one of the things that I started to think about was cognitive science, you know, just sort of how the mind works. So with AI, you know, you're trying to program a computer that can almost like think like a human that can perceive the world, that can understand it somehow. And then in philosophy, you're really thinking through language and, you know, by writing arguments, thinking about, you know, how does the human mind work? How does it contend with, you know, reality, things like this? But the piece that I was missing was actual neurobiology, right, like the, the nerve, the human nervous system, the substrate of, of the mind. And so I decided that before I went back and entered into a graduate program, I needed to learn more about neuroscience. And so I managed to get a job at Duke University in a neurobiology lab, studying human perception. So I worked as an engineer for about a year and then I went to Duke and I worked there for two years. And this was really, it wasn't a graduate program, it was just working in a neurobiology lab. And I was auditing courses on neuroscience and neurobiology, well, it was there trying to learn, you know, how does, how does the mind work? I mean, how, wow, how the human mind perceives the world. Yeah. How, I mean, did you, did that, is that helpful in what you do today? It is, it's, it's pretty interesting. So the lab that I was working in was really focused on trying to understand how humans perceive the world. And so let me just give you an example of, of why this is difficult and interesting. So the, when you look out like at an environment, like, you know, this room, there's light which bounces off of surfaces, it comes back and it, you know, moves through your eye and essentially is projected on your retina. So there's, for each of your eyes, there's a 2D projection of light from the room. And the question is, how do you go from that like 2D projection, on that 2D image to understanding what's actually out in the world, like the 3D environment, the surface reflectance, you know, properties of things like the wall or, you know, the carpet or whatever, how do you sort of solve that problem, and it's called the inverse optics problem. So it's the notion that you have a 2D image and you're trying to kind of understand this complex 3D world. And the challenge is that there's actually not an easy solution to this because you're moving from essentially, like, 3D dimensions to 2D dimensions, information is lost. And so another way to think about this is that an infinite number of different worlds could have produced the same visual image on your retina. And this manifests itself through illusions. So there are certain types of illusions, or something, for example, called an Ames room, where when you look at the room, it looks like a rectangular room. But in fact, it has, you know, this kind of crazy shape, you know, something you can look up maybe later. But the interesting thing about that is just the fact that something that appears to you to be, you know, like a normal rectangular room is actually something completely different. So that is just an example of one of these optical illusions. Now, the interesting thing about illusions is that they basically, everything you see in some ways is an illusion, right? So they're not outliers. It's not like every once in a while your mind makes a mistake and you kind of see the world incorrectly. You are always inferring some world that is not quite what is actually out there. It's the rule, not the exception. And so this is, it forms almost a philosophical problem. It's like, if you are… looking at the world, but you can't actually infer what generated, you know, the images that you see. How do you even, you know, how do you even interact with it? Like how do you, how do you continue to exist if you're not seeing things properly? And so, you know, the conclusion, one of the conclusions that we came to was that really the way that you see the world is whatever way is necessary to allow you to continue to persist. So we kind of think about this as like, you see the world in an evolutionarily sort of appropriate way in a way which informs your actions so that you kind of do the right things. You continue to exist. You can ultimately reproduce and continue on. And so, there's just one of the problems that we wrestled with. Now, what does that mean? It means that your perception of the world is really shaped through experience. You're just perceiving the world in the best possible way for you to take actions. And some of these fundamental ideas actually carried through into the work I did in graduate school and even some of the things that we do today with the systems that we build, the robotic systems that we built. - Very interesting. - Yeah. - Wow. (air whooshing) - When we first started building the Sean Ryan show store front, we didn't have everything figured out. We had products, merch, ideas, and a growing audience. But turning that into something people could actually shop from and keep them coming back, that takes the right setup. And that's why we built it on Shopify. Shopify gives you what you need to start selling without having to piece a bunch of different systems together. You can build the storefront, manage products, take payments, track orders, and keep the business moving from one place. And if you're just getting started, you don't need to know everything on day one. Shopify makes it simple to launch your store and start selling in a few steps. They also have design templates and AI site building tools to help you get the storefront looking right without starting from scratch. And once people are ready to buy, Shop Pay makes checkout fast and easy, which is huge. You don't want somebody ready to support your business and then lose them because you got a clunky checkout. I like Shopify's product so much, it's so simple. I had the CEO and founder Toby on my show. So if you've been sitting on an idea, waiting until you have it figured out, stop waiting. You can start simple, learn as you go, and build it from there. If you're ready to hear the of your first sale today, head over to Shopify.com/SRS to start your free trial today. That's right, start your free trial. It's Shopify.com/SRS, that's Shopify.com/SRS. So what do we go from here? - So I think one thing, when you think about building a robotic system, one of the ways that we kind of think about this is that when you're perceiving the environment, you're not just measuring it. You're taking the context and your prior experience and you're using that to predict what you think like the world actually is. So for example, if you look out at a set of trees, you're not just measuring that there's some obstacles out there in front of you, you're also predicting that there's free space behind them that you can potentially move through. And you get to that point by essentially seeing lots of trees like in your past, right? Like you use that prior experience and then you can understand when you see a pattern like this, that actually means that there's sort of space out there behind those trees that you can then leverage in order to make decisions faster to move more aggressively. That's really a machine learning way of thinking about things. You're using lots of data, lots of experience to understand what you're seeing in a functional way and then use that in order to move a robot more quickly or more aggressively. And those notions sort of led me from neurobiology at Duke, like really kind of thinking about data and machine learning as fundamentally interesting things towards my graduate education, which I pursued after that at Carnegie Mellon University. - Wow, but I've not heard everybody talk about that that has gotten into the machine learning AI stuff that's pretty fascinating. - Yeah, and I think like another thing which is pretty interesting here that I'll also highlight, I was originally went to Duke to try to understand the brain and how it works. And my hope was that by understanding that that would help me to maybe better understand artificial intelligence or how to build machines and things like this, what I pretty quickly realized was that neuroscience is really hard, right? People have been studying the human brain for almost 300 years and progress is slow. It's very difficult to understand how the brain works. We don't have a great grasp of it. Even now, we can describe a lot about it but not really understand it functionally. And so, one of the lessons from that was that, I almost came away with the opposite conclusion. So, instead of thinking about the brain as something that would help me to build better machines or build a better AI, I almost think that focusing on artificial intelligence and mathematics and probability and statistics and information theory and robotics helps to provide a framework that people may ultimately understand like the brain through. So, it almost works the other way that you have to really understand core principles of perception, planning, control, like these areas which are fundamental in AI and robotics to understand ultimately what the nervous system might be doing and be able to describe it. - Wow, wow. So, you worked at, what's in video too as well, didn't you? - That's right, yeah. So, before starting Overland AI, I worked for about five years at Nvidia. And this is, well, it was a professor. So, after Carnegie Mellon, I got my PhD there in machine learning. I worked on robotics problems. And then I was a professor at Georgia Tech for five years and then University of Washington for seven years after that. And one of the things which is really cool about being a faculty member or a professor who is running a research lab is that you can also work in industry. So, I had a research lab which was focused on robotics and machine learning and I had a number of PhD students who are working in that lab. But you can take 20% of your time and work in industry simultaneously. And as part of that, spent about five years working at Nvidia on machine learning and robotics. How was it work in there? - It was awesome. I mean, I think, you know, I joined around 2018. So, as before, Nvidia was really kind of like a tier one, let's say, tech company. I think like Google and Microsoft were really up there sort of defining state of the art. And but when I went to Nvidia, I think there were a lot of good people that they were hiring and people had recently seen the power of GPUs, right? Like this massive parallel processing. And I was thinking about that in the context of robotics. So, how can you parallelize tasks? How can you use it, not just these sorts of chips and this type of technology, not just for perceiving the environment but also controlling vehicles. So, an example of this, actually, it's carried through from work that I was doing at originally Georgia Tech and then Nvidia now to Overland, where when that vehicle is out in terrain, so when our uncrewed vehicles are out there and looking at terrain, they're evaluating tens of thousands of possible trajectories that they might take. Looking, ranking each one of them, determining, is this a good one or a bad one and then choosing how to drive after doing that. It's doing that about 10 times a second. And so, how do you get that to work? Well, you can use Nvidia GPUs to parallelize these tasks and evaluate many trajectories simultaneously and then decide how you're going to move based on that. - Very, very interesting. Do you still, do you miss being a professor? - Well, yeah, I'm currently, I've got a 5% appointment at the University of Washington, which means that I'm there every couple of weeks working with students. I like teaching, I like interacting with students. I think that's one of the great benefits. of being a professor is, you know, just engaging with students and people who want to learn. So that part's fantastic and I miss doing that. I haven't been teaching recently since I've been spun off the company. But I think working in industry also allows you to really scale your ideas more. So, you know, there's only so much you can do in a smaller research lab. And so, in 2018, the Army Research Laboratory spotted you at an IEEE conference, demoing machine learning. What is an IEEE conference? So IEEE, so it's association for electrical engineers. But it's one of the major types of conferences that folks publish in. So, you know, when you're a professor, one of your main goals is to publish papers, right? And those scientific papers further human knowledge. And so, in computer science, the way that you do this is you publish papers at a lot of conferences. You work with your graduate student, you develop a new technology, you then tell the world about it, right? You publish it in a paper, and you do this on a pretty fast iterative basis. So, you know, that's just one of the major conferences in robotics and we had some work there. And, you know, some folks from the Army were seeing what we were doing and had some cool ideas of how we could potentially take that fundamental research and start to apply it to Army problems. Right on. Right on. And then the DARPA racer, yeah, the crucible that bird over land AI, what was that? What's the purpose? Yeah, yeah. So, okay, so I was, you know, originally working at Georgia Tech and doing some work on ground vehicle autonomy. So, the way that this started out, we took one fifth scale vehicle. So, these kind of smaller remote control vehicles. We put computers and sensors on them and made them autonomous. And we were trying to race them as fast as possible. So, we were using data that we were collecting while we were driving these cars to learn a, what's called a policy. You think about it as like, you know, AI essentially like for the vehicle that could perceive the world and try to drive really quickly. And these vehicles are doing things like drifting around turns and things like this. So, they learn to do this, which is one of the things which is cool. Like, the vehicle's out there, it's trying to drive faster and faster and then it's learning how to do things like drift in order to drive even faster. So, that's what the Army was looking at. Now, is it actually learning or are you programming that into it? So, it's actually learning. So, you start by bootstrapping it. Like, you have a human demonstrate, you know, this is how the vehicle should drive. And then it tries to replicate what the human does. And as it does that, sometimes it makes mistakes, sometimes it does well, but it's kind of grading itself. And then it will start to experiment. Like, if I, you know, accelerate a little bit here or I break a little bit there, does this make me, you know, faster or slower. And as it does that, it learns how to drive faster and faster and it learns on its own. And so, you're programming in the ability to learn, but then it's looking at its own behavior, learning from that and driving, you know, figuring out how to drive faster. Now, is this like on a track? Yeah, exactly. Exactly. So, we started out by driving on a sand track. And we were able to achieve these, like, really fast lap times where the vehicles, you know, basically drifting through turns and things like that. And so, when I started to work with the Army, the question was, can you take these fundamental principles and apply them to larger vehicles and figure out how to drive aggressively through all sorts of different terrain? So, not just on tracks, but, you know, through forests and deserts and beaches and things like this. So, excuse me, what I was going to ask is, I mean, if this, if, okay, if we have a race car, it's going around a track over and over again. And it is learning. And then you switch up the track. Well, I mean, will it be able to take what it learned on track A and apply it to track B immediately? Or will it, do you understand? Yeah, basically what I'm saying is how do we go from a track to high speed chase in the middle of, I don't know, bandhead. So, this is kind of like a fundamental challenge in machine learning. Like, if you learn, for example, how to drive really fast on a oval track where you're always going in one direction, right, you'll learn how to drive really fast while always turning left essentially, let's say. But then if you go on another track which has right turns, will it be able to generalize and, you know, be able to perform well on a track like that? And the short answer is, you know, not without some sort of work. So, generally speaking, you want to collect data in a wide variety of environments that are inclusive of the types of places that you might want to be driving in the future. So, for example, you might have a complex dirt track with left and right turns and some wider turns and some sharper turns and things like this. If you train on a track like that, it's very easy to then race on like an oval track, right, because you've seen all of the things that you're likely to see. So, that becomes easy. But if you, for example, train on a dirt track and now you have to drive an asphalt, you may not be able to do that super well. And so, you need to collect new data as you move to that new type of environment or new type of problem, incorporate that into, you know, your learning algorithm and then it will, you know, start to do better on that new type of track. And that matters, you know, even now when we think about where we want to drive our vehicles because if you only train, for example, in the desert, you're not going to be able to necessarily drive well through a forest or vice versa. And so, you want to really train these systems, have them, collect data and learn from as wide a set of environments as possible so that they're able when they encounter a new environment to still perform well, that there'll still be aspects of that environment that they've seen before. So, so I guess what my question is, will it get to the point where it runs the route for the very first time, like a brand new route for a very first time, it could be, it could be off road, it could be, sure, in the middle of a city, but we're talking left turns, right turns, heavy braking, fast acceleration, drifting, all of that stuff, will it be able, will it eventually learn what it needs to learn to be able to, to have a complicated new route run at the first time and it will run it perfectly, it will go as fast as is the machine is capable of. Yeah, it's certainly possible, and like that's what we're always striving for. So, when we're putting, you know, ungrouped ground vehicles or Thomas vehicles in new environments, they're already performing really well because it's seen many aspects of that before, and you're trying to get it to perform, you know, optimally, right, like that is the goal, like potentially faster than a human driver, even on environments that it's never seen before. Every time our autonomous vehicles are out in the world, they are basically, you know, we've trained them so that they're, we don't like pre-map environments or anything like that. It's as if they're encountering that environment for the first time, and they're constantly learning. So, the more environments that we encounter, the more data we get, like the better and better the system gets. But then the goal is always kind of moving towards that optimal movement. How close are you to that goal? Depends on the environment. I mean, I think we can, we can already drive faster than humans in some environments. So, yeah, it's pretty cool. Wow. Very interesting. And so, how long did you spend at DARPA? So, yeah, so I started work with Army Research Lab, you know, again, around like 2018. And in 2019 or so, I started to talk with folks at DARPA. They were really interested in developing essentially rebooting ground autonomy for defense. So, like, here's, here's what the situation was. DARPA, back in 2004 and 2005, had something called the grand challenge or grand challenges. So, these were challenges where they were trying to incept ground autonomy. So, this actually goes back to the NDAA in 2001. This is basically where Congress decided in 2001. They were like, by 2015, we want a third of all military ground vehicles to be autonomous. So, I think about this, you know, over 25 years ago. The problem was, no one knew how to to do that, like there weren't, it wasn't like There were autonomous vehicles driving all over the place, and this would be easy. No one knew how to make these vehicles autonomous. Where did the idea come from? Well, I think there had been work in robotics where people were moving vehicles in simpler environments autonomously, and so the notion was like, well, what if we could do this on the battlefield? What if we could do it in these more complex environments? If you could take the warfighter out of the vehicle, you can imagine that not just provide safety, but potentially tactical things that you can do. We can maybe talk about that in a little bit. In order to make those vehicles autonomous, you needed to know how to do it, right? They didn't. So DARPA, one of the things that's great about DARPA is that it is an organization that is designed to just like create new things, right? You have some crazy challenge, DARPA is out there and can create a program, pull together some of the best minds in the U.S. to focus, like really focus on it for up to about four years and try to solve a problem. And so in 2004 and 2005, they came up with a DARPA Grand Challenge where they were trying to race vehicles from, you know, Barstow, California to Primnavada, and there's about 135 miles on dirt roads. And they just laid down a challenge that said, like, if you can do this, you get prize money. And all sorts of teams came together to attack this problem. So there were university teams from places like Carnegie Mellon and Stanford and MIT and so on. And there were industry teams, like Ashkash, you know, had a team and there were just people who were trying to put together autonomous vehicles and their garage, like just build robots. They went out there and raced and in the first, in 2004, the first challenge, no one made it beyond seven miles, like that was the, you know, so it was, it was CMU's vehicle I think made it that far. It got stuck, caught fire. It was like a whole thing. The next year though, in 2005, I think five teams competed this challenge. They made it the entire 134 miles. Those teams, like the folks from those teams, after that challenge was over, there were a few other ones, there was something called the Urban Challenge and some other DARPA programs which followed up on this. But many of those people then moved into industry and started the self-driving car projects and companies that then turned into things, you know, companies like Waymo or Aurora, innovation and so on. So these autonomous driving companies, commercial companies. So by, you know, 2012, let's say a lot of work was being done in the commercial sector. Bootstrapped off of this DARPA work, right? So the military, you know, started this whole thing because they wanted autonomous vehicles. All started to build autonomous vehicles because of this DARPA program. But then basically just went off into industry and were working on like robot taxis and autonomous trucks and things like this. So by like 2019, coming back to my story, DARPA was in a position where they were like, well cool, we have autonomous taxis, you know, there's been a lot of progress in this area. But you know, where's our autonomous tanks, right? Like where are autonomous military vehicles? The whole point of this was initially to support the military and so DARPA racer was a program that got stood up. It started in 2021 to reboot autonomy for defense. So specifically to take a lot of the learnings that had been produced over the previous, you know, 20 years or so for the on-road autonomous driving industry and, you know, work which had been done in robotic perception and, you know, robotic vehicle control, bring that back together and focus on defense problems. And so that meant trying to drive much faster, you know, larger and faster vehicles offroad, what we call complex natural terrain. So, you know, no roads at all, right? Like through deserts, through forests, through snow, things like this. And contested terrain. So thinking about, you know, how do you move when you not only do not have infrastructure which is there to help you, like roads or road networks or signs or things like this, but infrastructure which might be in the environment which is there to defeat you, to stop you. And so that's what the DARPA racer program was. So I had already been doing work at, you know, at Georgia Tech and then University of Washington working with the Army on developing off-road ground vehicle autonomy. And I then put together a team, it's called a performer team, to attack these problems for the military through DARPA starting in 2021. So they recruit you from the IEEE conference. So the way that that worked was like through the research I was doing and the publications that was, you know, putting out to the world, Army saw that the technology that we were, Army research labs saw that the technology we were developing might be really helpful for the types of problems they wanted to solve. I then started working with Army. So the way that that works is when you're running a university research lab, you have a bunch of PhD students, you know, they're doing research, you're publishing it, you're making it publicly available for other scientists to see. Okay. But you need funding to run that lab. And so people pay you to essentially do research, they pay your lab to do research. So Army, U.S. Army was one of the organizations that started to fund my research. When you fund research, you can say, okay, here are the problems we want you to solve, we'll give you, you know, this much money to solve them with your PhD students. And then you provide those solutions back to the Army. So that's what I was doing when I was working at Georgia Tech and University of Washington, my lab was partially funded by the U.S. Army, then I worked on problems that were interesting to them. We provided those solutions back to the Army. And then we started to work with DARPA, which is at the time Department of Defense, but Department of War, like level organization. They saw the work we were doing with the U.S. Army, and then they decided to fund my lab at like a much higher level to attack these problems of how do you drive vehicles off-road and all sorts of different types of terrain at high speeds. Wow. So you've really been at the cutting edge of this thing the entire time. Yeah. So we've been working on these problems for more than 10 years, more than a decade. And my lab was doing a lot of that work in academia before we spun out overland. Wow. Well, before we get into overland, let's take a quick break. There's been a lot of attention lately on how important sleep is. Not just for recovery, but for your brain, hormones, immune system, and for your performance. And the more research comes out, the clearer it gets. If your sleep is off, everything else gets harder. That's why I use Helix. I've had my Helix mattress for a while now, and it's been a real upgrade from what I was sleeping on before. Helix has over 20 mattress models, so you can find one that actually fits how you sleep instead of guessing your way through it. And if you sleep hot, they've got cooling upgrades that help you stay comfortable through the night, especially during the summer. For me, Helix feels high quality, durable, and comfortable. It's not just a mattress that shows up in a box. It feels like something that's built to last. Helix shifts straight to your door for free in the U.S., and you get a hundred and twenty nights sleep trial, a limited lifetime warranty, and they're happy with Helix guarantee. So returns in exchanges are simple if it's not the right fit. Helix is also the most awarded mattress brand. Tested and reviewed by experts like Forbes and Wired. Go to HelixSleep.com/SRS for up to 30% off. Do you ever wonder what it takes to make an episode of the Sean Ryan show? In this exclusive studio tour, I'm taking you behind the scenes for an in-depth look at every part of the operation from the editing room in the main studio to the spaces where we film range day, content, and more. You'll see how the show comes together. Meet some of the people behind it and get a closer look at the work that happens off-camera. When you become a paid member of the SRS Patreon community, you get more than just the podcast. Watch new episodes early alongside other members. Join monthly live shows with guest Q&As and submit questions just like you see on the show for upcoming guests on the protector tier. You'll also unlock exclusive range day videos behind the scenes content in premium ambience videos that you're not going to find anywhere else. Join the Patreon community today and get access to the full experience. Let's talk about the Textron M5. Sure. Yeah. So, as part of the DARPA racer program, so this is DARPA program where we're trying to develop ground vehicle autonomy for the military. We started out working with Polaris Razor side by sides. So these are very capable vehicles that can go very fast. We've got four wheels, about 3,000 pounds, about the size of a small SUV, something like this. And in the first part of the DARPA racer program, there were multiple teams and we were essentially like racing these Polaris Razors, so trying to record the best possible times through a wide variety of different terrains and test scenarios. Once we progressed through the program, the DARPA program manager didn't want us to just be using those small vehicles, so we moved up to the Textron M5. The Textron M5 is basically a robotic combat vehicle prototype. It's about 12 tons, so it is much larger. You went from a side by side to a 12 ton vehicle? Yeah, it's a 12 ton tracked vehicle. So it's about the size of a 113, so it looks like a tank, like a sort of like a small tank, I guess, but pretty heavy vehicle, tracked, it's electric, so really aggressive. It has the ability to essentially go from zero to 60, like think about like a Tesla, but like a tank version of that. It's a really, really awesome vehicle to work with, and we're putting autonomy on that vehicle and then driving that through off-road terrain. Okay. How did that? And this was a competition, correct? Yeah, so DARPA racer, it started out with three teams, so it was our team, and there were two other teams that we were competing against. By the first summer, about a year into the program, we had by far the best technology, and pretty quickly, the other teams essentially dropped out, and we were the only ones left. We were focused on not only driving on the side by sides, but also the much larger vehicle, so we're the only team that actually worked with these. And there aren't that many of them. There's something like four or five, tech strut M5s ever produced, but we had all of them, and we were just smashing through terrain in these huge tracked vehicles, pretty awesome. So what is the kind of the point of the competition? Do they buy the technology from you, or do they fund the technology? That's a great question, so how does it work? Yeah, this is a really good question, actually, because the way that DARPA normally works, they have these programs where they're trying to develop new technology, in this case, develop ground vehicle autonomy for the U.S. military, so for our soldiers and Marines. But often, in a DARPA program, even if the program is very successful, like ours was, how do you actually take that technology and then get it into the hands of the warfighter? How do you actually do that? It's a good question, and very few people succeed. So this known as transition, transitioning like a new capability that's developed from DARPA to the actual warfighter is something like 4% of technologies ever actually get transitioned. So just because you succeed in developing a new capability doesn't mean that the warfighter gets that capability. This is actually why we created Overland AI. We were in this program, and we had this new capability, and we were just like, this is incredible. We can drive extremely fast off-road, we can do it in a variety of different types of vehicles. We have the thing that DARPA wanted us to produce, like we actually did it. Now how can the warfighter benefit? Promising way of making that transition was to spin off a company that would then take the capability that we developed as part of this DARPA program, commercialize it, so turn it into a robust, reliable, commercial autonomy stack, and then sell that back to the Army and Marine Corps so that they could use it on their programs, so on their vehicles, and so that warfighting units could start to work with this technology and start to integrate it into tactics. So that's why we created Overland AI. We thought, this is just going to disappear if we don't do this. Why would they fund the research? So DARPA funded all of this to develop the capability, but just because DARPA. Then what has to happen is the Army and the Marine Corps, or whoever, one of the services, has to then put up money to be able to then buy and transition that technology because they're on DARPA or from. Yeah, so the way that DARPA works is that DARPA, like these DARPA programs, are generally four years long. They have a time limit, and that's part of the point, it gives you a very focused window in time to just do whatever it takes to try to develop this new technology, but at the end of those four years, that program is over, and you have to find someone who is then going to support the continuation of that technology to move it into war-fighting units. You've created it, now someone has to continue to transfer it, to get it onto military vehicles to push it into units, and that's where a lot of these programs die because you create the technology, but no one picks it up when it's over. I would think there would be. What would you call this, like an incubator? You get picked to get funding to be an incubator, and then DARPA doesn't even, they don't own the actual. Well, they do, they own the IP, so also on the IP. Yeah, so the way that this works is when DARPA pays for the development of this core technology. In this case, I'm running a lab at the University of Washington, so they pay the University of Washington to develop the IP. When they do that, that IP belongs to both DARPA and the University of Washington, right? Because DARPA pays for the University to develop it, both parties essentially own that IP. Once you have that IP, in this case, it is raw technology, right? It is research grade code that can demonstrate this capability, that will allow a vehicle to drive fast, but it's not going to be reliable in the same way that a Waymo is, where you have a whole team of engineers that is making production quality code. So, they own that, they own the basic capability, and they own that IP, but that then has to be taken and turned into essentially a commercial production-ready piece of software. It's almost like a rough draft, even if I don't think there would be. But then someone has to kind of pay for that transition, and that's often, again, where these programs end up having problems, where you develop a new technology, but then who adopts it, who actually transitions it to the warfighter? Are these classifying? I mean, I would think there would be venture capital firms galore just surrounding these projects. Well, yeah, yeah. So, sometimes, yeah. That's right. Only 4% make it to the warfighter, that's 96% of the projects, like, oh, yeah. So, but in some sense, that's actually what happened here, right? Because we said, okay, like, we're going to take this IP, we license it, right? So, that IP is licensed to Overland AI. We created this company to essentially take that IP, and then turn it into a commercial autonomy stack that is reliable, and we keep building on it, we keep improving it, and then we can transition that back to the warfighter. So, Overland AI got its start by focusing on the software portion of this, right? Take these good ideas that were developed during DARPA. We license that IP at Overland. We build on top of it. We turn it into a commercial autonomy stack, and now we have this software that you can put onto a vehicle to make that vehicle autonomous. And we had worked with different types of vehicles. We discussed the side by side or the, you know, a big Textron M5 that we can make autonomous. You know, we really want to develop software which would work on any vehicle that the military had. And so, Overland was stood up to do that. And to your point about VCs, you know, we're a VC backed company, right? So, they're essentially making that bet that you're saying. It's like, okay, we can step in, provide additional funding to turn this into a mature product that can then be transitioned over to the war fighter. But now you're in like defense tech territory, right? Where you now have to fight all of these battles to go from, you know, this good core idea, good core IP to actually getting it procured, right? And that takes, you know, time and effort. And so, like, our story as a defense tech company is really similar to a lot of other stories except that we got this start by, you know, doing all of this, you know, sort of initial research to accept a new technology that the military really wanted. - Very interesting, very. And so, so the DARPA competition ends, you win. - Yeah. - So, all the information with you and develop Overland AI, right? So, we started Overland AI and, you know, we, Overland actually was formed before the DARPA competition completed. So, we became part of the competition as well. We then, as the competition was going on, you know, we were developing new technology from moving like really quickly through a lot of different biomes so different types of environments. And then we were really trying to think and this is the Overland AI started in December 2022. So, three and a half years ago. And we were initially focused on, okay, we have this software, we're, you know, we're finding it, we're turning into a commercial piece of software. How do you then get traction with like the Army and Marine Corps? I mean, not just DARPA, but like, you know, the people who ultimately need to use this technology. And the problem was that even though we had this like, great piece of software, it has to go on a vehicle. So, it's like, what vehicle are we gonna put it on? And we started to talk to companies and units that had vehicles which could be made autonomous, but there just weren't that many of them out there. We had something which could be very effective, but we didn't have vehicles to put it on. So, we started out, you know, again, with a software stack, but we quickly realized that we had to actually also build the hardware so that we could get this idea of autonomous vehicles into the hands of warfighters faster. So, we just started to build the vehicles too. And that's what resulted in like this ultra-vehicle that, you know, you've seen outside where, you know, this is an autonomous vehicle that we built and we put our software on so that we could start to put many different types of payloads, you know, on the vehicles and start to work with warfighting units to really integrate autonomous systems into their concepts of operation. So, our path really is initially research and development starting with DARPA. That then we worked with a defense innovation unit and got a prototype contract with them. We then started to work with Army Applications Lab and, you know, winning, you know, these severs, these small contracts with the Marine Corps and the Army. We're doing really well with those. And, you know, we then, last year, we revealed like the ultra-vehicle, fully autonomous, you know, vertically integrated vehicle that you could push into hands of warfighters. We started to push this technology into units who were constantly testing them and then this led to a production contract. So, we just recently won the first production contract for autonomous ground vehicles in the US military and that's with the Marine Corps. So, it's really this whole process of, and it didn't take that long, right? It's about three and a half years where we went from like, University Lab, R&D, through prototype, fielding with warfighters, production contract. - Wow. - So, we felt this could have here. - Yeah. - Yeah. - Wow. - Wow. - I mean, so, who won the contract or who are you contracted to? - Yeah, yeah, with the Marine Corps. So, the Marine Corps is buying a whole set of the autonomous vehicles that we produce. It's part of their ground-based air defense program. And so, they'll initially start using those vehicles for autonomous resupply of basically air defense systems. So, trying to make sure that they get enough ammunition as they're shooting down drones. - Can I ask how many, you're gonna manufacture for them? - Yeah, so we're manufacturing the first tranche about 15 of these in the next year or so. So, that's what we're starting out with. And one of the things which is pretty interesting about these vehicles, they're modular. So, a lot of units want to start using them for resupply where they're just kind of putting supplies on them and moving those supplies back and forth. But we've been doing a lot of work with other units, like 82nd Airborne, 173rd Airborne, other warfighting units to put other types of payloads on board the platforms as well. And so, one of the things that we've been talking to the Marines about is putting sensors and, you know, affectors. So, basically, Connecticut County UAS payloads on the vehicles and thinking about how to disaggregate them. So, one of the things, one of the ways that you can use autonomous vehicles is by, you know, essentially putting your sensors and putting your, your counter UAS, for example, payloads on vehicles and moving them away from the places that you're trying to defend, right? So, you're disaggregating, you're dispersing, you're essentially making it much harder for an adversary to be able to take everything else, like all of the pieces out when you're defending an area. - Interesting. So, they're going to be using this for logistics. You're going to be using this for defense. And I would imagine there's going to be an offensive commode. - Yeah, yeah. So, - And already. - The way that we're thinking about this is, so, logistics, so you can think about resupply and casualty evacuation. A lot of people, when they think about autonomous vehicles, this is like the first thing which comes to mind because they think about a vehicle and they're like, "Oh, okay, I can put stuff in it and I can move it," right? Like, that's like what a vehicle does. But I like to think about these vehicles more like robotic systems. They can sense, they can track, they can move on their own, on the battlefield. And I think like the real product market fit here is that you want to move these vehicles out in front of the warfighters, right? So, you want to be using them for things like intelligence, surveillance, and reconnaissance, being able to move them through, you know, bad weather, they can persistently, you know, stay in a location for a long period of time because they're just on the ground. So, they might be more effective than drones for some of these things. You can use them for, you know, defense or strike capability, so you can put kinetic payloads on them, you can put drones on them, launch them off of the vehicles, so you can, you know, move them into areas that might be too risky for a human, but they can potentially hold ground or try to take terrain in those areas. Breaching is a major thing that we're working on, so we're working with a number of different combat engineering units on removing people from breaching operations, which are just exceptionally dangerous. And then air defense, as I was discussing, is essentially putting sensors and shooters in different locations and moving them autonomously, reconfiguring what that defensive position might look like. I mean, I could think of a whole slew of things, but let's, let's run out back and take a look at this thing. Yeah, sounds great. You guys know my schedule. I'm in the studio all day. I'm on the road, and I've got kids. The last thing I have time for is standing around a gym, wondering what to do next. That's why I'm on the ladder app. I've got a real coach in my ear on every set. What's next? What way? Why? It's a new plan every week that builds on the last, and it goes where I go. The garage with dumbbells or a hotel room on the road, 30 to 45 minutes, and I'm done for less than $1 a day. I've trained my whole life. In the teams, nothing we did was random. Somebody who knew more than you wrote the program, and you got results. That's the difference between working out and training. Ladder puts that kind of real programming in your ear. Remove the guesswork with ladder. And get a real coach in your ear telling you exactly what to do for every workout. No thinking. Everything planned for you. If you're a guy who. been meaning to get back to it. This is the way back. If you have an iPhone, head to ladder.fit/srs and take a quick quiz to get matched with your coach and the right plan for you. Use my link and get a free seven day trial with no credit card and $10 off your first month if you join. Hi, we're out here with Byron Boot CEO and founder of Overland AI. You're ready to take a look at this beast here. What are we looking at? So this is an ultra one of our autonomous vehicles. So let me just kind of show you what it is. You can tell it's an off-road vehicle. It's got long travel suspension and big wheels, but it's fully autonomous. So it has sensors up in the front of the vehicle. There's stereo cameras. There's actually three of them, like one here, one there, one on the other side, LiDAR. So this allows it to see in this whole area in front of the vehicle. Not what degree. Actually, you can see 360 degrees around it. So you have these sensors in the front and there's also sensors in the back. So it can just kind of see this whole whole area. And then if you come along this way, this is a payload deck so you can put a wide variety of payloads on it. You can see there's all these attachment points. So it makes it easy to integrate new payloads onto the vehicle. Below this deck, you have compute. So that's where your computer systems are, your power batteries, the alternator, like all this is below the deck. And then back here, you've got comms. So satellite comms, we can integrate tactical mesh comms as well on the vehicle walking around. Back here, you can see there's more sensors. So stereo cameras and LiDAR, again, what is the LiDAR do? So LiDAR basically allows you to see depth in the area around the robot. So think about it as like a depth sensor. It tells you the distance to surfaces in the terrain. Okay. Yeah. What what what what engine are we running in here? I think it's it's you have to ask Chad, it's about about I think 115 horsepower engine. So this is based, this whole vehicle is actually based on a Polaris Amrase or commercial side-by-side vehicle. Okay. Okay. So the engine drivetrain chassis that all comes from Polaris. We take out the seats from the vehicle, take off the roll cage and everything, and then transform it into this autonomous platform. Right on. And so what what kind of stuff would you be mounting of here? So you can put yeah, you can put all sorts of different payloads on here. This particular vehicle takes about a thousand pounds of of payloads and pounds. So we've done everything from, you know, fairly straightforward things like comms making a communication node to electronic warfare payloads to remote weapon stations. So you can mount, you know, a machine gun on on the vehicle. Nice. So if you mount a weapon system on here, is it all run through autonomous autonomous stack? Is that what you call it? So the autonomous stack controls the vehicle. It allows an operator to tell, essentially tell the vehicle where they want the vehicle to go. So it will say something like, you know, go 10 kilometers to this location, orient in this way. And then they will access the payload through the comms network. So the human is still in the loop whenever you're using the payload, but the movement of these types of vehicles is autonomous. And what that allows you to do is instead of having to just like remote control the vehicle, essentially, you know, drive it through its sensors, the human can just say I want one vehicle to go to this location and want two vehicles to go to this location and so on. So it really allows for force multiplication where a single operator can control many different platforms and then access the payloads on those platforms. It will show you how to do that. It's like the old computer games like Warcraft. It's exactly like that. So that's actually how we go here, how we think about it. Kill this thing, you know, I think like our vision of this is a single operator could control potentially hundreds of different assets on the battlefield. And you're going to do that through an interface, which is a little bit like a real-time strategy game, like Warcraft or Starcraft, where, you know, it allows you to, you know, select the units, tell them where to go execute the payloads and so on. Wow. So I just, I mean, with the changing landscape of the battlefield now, there's the first land, like autonomous land, vehicle I've seen. So you could mount like an apparatus, direct energy weapon on this thing. Yep. So let's say, I mean, Enduro's got their stuff coming out, Shield AI's got that new X-Bat they came out with. You could mount it. So what I'm asking is, I don't know if we're there yet. Maybe we're already there, and this is old news, but with all these companies, you know, like yours, the submarines, the Seronic, you know, Overland AI. When we do go to a full-scale war, are you going to be like, is the operator or the battle or the ground force commander or just the commander of the entire operation? Are they going to be controlling Shield AI's X-Bats, Overland AI's ground vehicles, Seronics, Surface Warfare vehicles, submarines, drones, all of it, all of it on the same system. So military is working on this right now, finding ways to integrate all of these different pieces into the same sort of system so that you have a unified view of the battlefield. Now the way that that may play out, I think they're you're likely to see there are single systems where you can see all of the different pieces, and then there will likely be systems which will allow the warfighters to actually be controlling some subset of them on the battlefield, but it all have to be linked together to provide that overall awareness of what you're pushing out there. And so if you have 100, 200 of these things right here, what is the name of this? This is an ultra. The ultra. Yeah. 100, 200, 300 ultra is out here and they've got surface-to-air missiles for air defense. They've got, I don't know, 50 calibers for other ground vehicles and anti-personnel and rocket launchers and grenade launchers and drones. Are they all going to read off each other or somebody going to have to? So what the way that this is going, we're taking, so the the basic idea is that you start with just being able to move like one vehicle at a time, right? So you can say I want this vehicle to go to this location, you just let it go, you know, execute a payload there. We're already can do that really well. So now we're starting to build up coordination where you can move multiple vehicles at once and they coordinate in order to achieve a task and then we'll keep building on that. So the basic idea is that, you know, over like as we build out the technology and as we field more and more of it, you're going to have a situation where a single operator will be able to move multiple vehicles into formations, have them, you know, send them to achieve a particular task and they're going to go out there and just do it. And that's part of what's called mission autonomy or orchestration. So platform autonomy is basically the autonomy that lives on the vehicle that allows it to see the terrain, see where, you know, vehicles and people are out in that terrain and move through it. And then you tie that together where vehicles are coordinating with each other and that's mission autonomy. They're coordinating autonomously to actually conduct like a full mission. And then all of that, like all of these different autonomous systems and uncrewed systems which are in the battlefield will be pulled into command and control systems which allow people to see everything which is out there. Wow. I got a I got a million questions for you, but it's it's hot as shit out here in humans. Yeah, yeah. Let's see. Let's see. We'll do that inside. That that sounds great. But let's see what this thing can do. Do you want to drive it? Yep. Yes, I want to drive it. Okay. So the upper part of this interface here, this is where the vehicle is on a satellite map. You can see like the name of the vehicle. And then down here gives you a view of what the vehicle sees. So over here, this is the front camera on the vehicle. And then this is actually like the AI view of the vehicle. You don't have to necessarily spend too much attention like looking at this, but the magenta areas are lethal. So those are areas that you know, the vehicle you don't want to move the vehicle into with as you're as you're tele operating it, you can actually move it anywhere that you want. So this is that big pond that we're digging down there, bro. This is this is going to be about 50 meters around. So it's only seeing up here on this little area, what we can do is give you the controller and the way that this works is when you pull, like you have to pull this down so you're lower left and you hit A and that moves you into autonomy and then you can move this joystick moves the vehicle forward, holy shit. So that's basically moving it forward and then you can turn using this so we can turn it to the right. And when the vehicle is here in front of you, it's tempting to look at the vehicle but you should actually look at this which is what the vehicle can see and then you can control it. That's right, yeah and you can control it beyond line of sight so you can do this from like 5,000 miles away. So left down, and then forward. It is, it's, we're not looking at the vehicle. Yeah, I want to look over there, so. Holy shit. Can I go, how do you go backwards? Okay, so to go backwards, hold down this, so both that and this and then that and then move that backwards. There you go. Oh, there we are. It will automatically stop but won't run us over. Let's not test it. Alright, yes. Yeah, there it is. Slowing for a person. Yeah, yeah. Well, that's not going to work anymore, Byron. You can, you can turn that off. Dude, this is crazy. So it's somebody be looking at this or like a VR headset or does, I guess you could probably do no matter what. You could do whatever. Like generally you're looking at this. You can look through the other sensors on board the vehicle as well. So you can look at the rear sensors or off to the sides. You can access the payloads on the vehicle through the interface up here. Wow. Alright, and then if you want a. Oh, shit. There's a tree yet. Maybe back up. Okay, so you can't tell I'm not a gamer. I'm not great at that either, but we like to control it actually through this interface and I'll show you that in a moment. I think one of the challenges for tele-operation is that you don't have like a vehicle sense, right? Like you can't feel like how the vehicle is moving like you would when you're driving. And now you're trying to interpret what the vehicle can see like through its own sensors. You should probably stop. Right, yeah. It's actually much safer to put it into autonomy and just tell it where you want it to go and it will find a way to get there without you know hitting anything. We do that while staying safe. Yeah, so I can do that. That's awesome. Alright, so. Oh, shit. I did almost hit that damn tree. Okay, let's send it to go get a pizza. That sounds great. I'm not sure though. Let me do this on the road. Can we go over there? Can we go over here and watch it? So before you guys run the route, how is that thing determining what is a human being and what's a tree and what's a vehicle and what's a rock? Yeah, it's looking at the whole environment around it and then it's determining where it can drive and where it can drive. So that's the first part. It's notion of traversability. Like what part of the terrain is traversable? And then on top of that, there's a semantic understanding of the terrain, which means you can understand that, you know, it can see a person or a car, you know, determine what that is, where they are relative to it. And right now, it's running a safety system, which essentially says like, don't, you know, stop if you get too close to a person or a vehicle. Right on. I mean, how does it just pick up like body temperature? It looks and so it's just using a camera and a silhouette. Yeah, and it's saying like, this looks like a person, this looks like a vehicle and it's able to pick them out. Cool. Yeah. And this is full autonomy while we're good really to do here. Yep. Yep. Holy shit. Oh, we're, we're in the way. Let's move back away from it. So yeah, as it's moving through the terrain, it's also tracking where all the people are, right, where the vehicles are, where they are relative to, to it. And then deciding how to, how to drive it picked them up. Yeah, so we can actually pretty wild that it picked them up in the middle of all those weeds and trees. Yep. You can see one of the benefits of this is, you know, you can just tell it, I want you to go to again, like this location and you can let it go and we'll, we'll do its thing. So I was in the run, basically, came up the hill, you know, went around the field and then stopped there. So very short run for what it normally does, but you can get a taste of kind of like, how it moves. That is sick. All right, well thanks for showing us what the ultra can do and let's go wrap up the interview. Yeah, thanks for letting us come out here and show us. Thanks for bringing that bad ass thing. There's nothing more frustrating than putting in the work and still feeling like your body is not responding the way it used to. You're training. You're trying to eat right. You're trying to sleep, but energy still drops off. Recovery takes longer and you don't have that same drive that you used to have. That's why I started taking Marsmen. What I like about Marsmen is that it's not forcing hormones into your body. It's designed to help your body use more of what it's already producing. As men get older, SHBG can bind up testosterone. So even if your levels look fine on paper, a lot of it may not be available for your body to actually use. Marsmen uses clinically dose ingredients to help free up bound testosterone plus vitamin D and zinc to support healthy T levels energy, stamina and recovery. And since taking it, I've noticed steadier energy and better recovery. It's not like a caffeine spike. It's more consistent drive throughout the day. Marsmen is made in the USA. Third party tested and backed by a 90 day money back guarantee. For a limited time, our listeners get 50% off for life plus free shipping and three free gifts at mengotomars.com. That's mengotomars.com for 50% off and three free gifts when you check out after you purchase. The last you where you heard about them. Please support our show and tell them our show sent you. Okay, so after taking a look at this thing, I, this, this, this just has so many different capabilities and could be used for so many different things. I mean, even just keep, I mean, you, it sounds like the first thing the military is interested in as is logistics. Which I mean, just keeping major supply routes open. Yeah, I mean, constant detection because these things detect, I mean, when we were up there, it detected, I mean, I know you guys have that safety feature on it, but it detected one of my camera guys who was, I mean, really wasn't even that close to it. It looked like he was made of 25 yards, 30 yards from it in the middle of a bunch of trees and a picked them up like that. Could this thing pick up IEDs? Is there a way that you could, that it could detect IEDs? I mean, yeah, so you can put a variety of different types of sensors on the vehicle. So you can certainly detect like people and vehicles and then you can also, you know, we've been putting tethered drones on these. So, you know, the drone can go up in the air, be able to look down on the ground. So, if you have, you know, ground penetrating radar or other types of sensors, you can certainly detect things like IEDs or obstacles in the environment that then can be ideal. identified and reduced. So that's certainly something they can do. Into your point, I think you can, of course, move things around. But just the ability to push these forward and sense the environment is, I think, will be game-changing. You don't have to put a person out there to do it. Do you see these integrating with human forces that do you see these as a forward observations platform for maybe a slew of tanks or emrades or whatever we're using, and then kind of push that information back to the larger force, or as it go in all autonomous? I think the-- so there's a number of statistics out there which are pretty interesting. So you've got like Ukrainian ground commanders basically saying that they're going to replace something like 80% of infantry with uncrewed ground vehicles in the near term. And those statistics are interesting. I think the reality will be a little bit more subtle. So these are machines that can integrate with human formations. You should think about them as giving humans more capability on the battlefield. So I think the way that they'll actually work with US military is that we'll have formations with humans in them. They will be pushing these vehicles out in front as you suggested in order to be able to get a better understanding of the terrain that they're going to be moving into. But you can use them for all sorts of things, right? So it's not just sensing what's out there. You can use them to create like diversions, right? Like you can use them for strike capability. You can use them to create that kind of counter UAS bubble to protect humans. I think about them as something that will be a force multiplier. I mean, you're saying that Ukrainian military saying it will replace up to 80% of infantry 80% of the infantry. And you're saying the US is saying, well, maybe not the US, but you see a subtle integration. Why won't we do a subtle integration? I mean, I know on one hand, I'm going to push the war fighter up because I love fighting war. So on the other hand, I'm thinking about my toddlers at home, with all the wars that we're involved in and re-involved in. And I don't want my kids going over there having to do it. And so why do you see a subtle integration? I wouldn't we go full scale and say, hey, like we don't need private Joe driving this M, you know, driving this M-Rap into hostile territory. I don't want to say a wasted life, but it would be a wasted life if he got killed when we have the capability of going full autonomous right now. I think we'll eventually get there. So one of the value propositions for ground autonomy is that you're really focused on one of the hardest domains in the military. So just kind of stepping back, there are a lot of folks who are focused on air power, naval power, or missiles. But at the end of the day, wars are ultimately one in the ground. That's where people live, right? We don't live in the sea. We don't live in the air. We live on the ground. And that's where we have to fight. And I think when you look at the US and our casualty rates, something like more than 80% of combat-related casualties since World War II are infantry. So of course, I think focusing on the ground makes a lot of sense, because we want to bring technology to the warfighter that will save those lives. And so yes, we want to take humans out of harm's way. And I think in Ukraine, we're seeing technology increasingly being pushed out in front of the warfighter. And I think they're very optimistic about how they want to use them. And ultimately, we do want to do that. We want to pull, we want to essentially reduce risk on the battlefield. Pull warfighters away from these points of contact, have more people who are maybe operating teams of robots that are far away from location, where they could get hit. But in the meantime, I think that the way they're going to be integrated is in a way that helps warfighters on the ground to stay safer, to potentially give them more options, more tools, without totally pulling them out. So to your point, yes, we want to ensure that we ultimately want to pull as many people out of harm's way as possible. But these are still machines that aren't quite as creative and adaptable as humans. And so I think what we're really trying to do is find the places where we can maximize the ability for them to be able to take risk while also allowing our humans to do what they do best and do maybe more specialized roles on the battlefield. So I think it'll be a process. I mean, just thinking about my time in Iraq, Afghanistan alone, I mean, I have right. IED thread was huge. And then here is a key statistic here. 44% of US killed in action from 2006 to 2021 were killed by IEDs. 44%. Now, I mean, if these things did have IED detection capabilities, it's got to be better than humans than-- I mean, there's 43. It's almost 50%. Yeah. People that were killed would still be here today. Yeah. Just with that application. I think there's a lot you can do there. I mean, there's sensors you can put on the vehicle. Human sensors, like I was just saying, with caught my guy out in the weeds trying to-- In a bunch of trees, I mean, rolling down Iraq, I mean, there were sniper problems. There was the guy at the Trigger Men of IEDs. I mean, I would think that it would pick them up immediately. I mean, they can't hide from the machine. Yeah. How is it sensing them? Yeah, so the machine has cameras on it, right? And we, depending on the loadout for it, you can have sort of normal RGB cameras. You can have thermal cameras, right? Like there's a lot that you can do to detect people and vehicles in the environment. So I was just using cameras out there. I was able, as you said, to just be able to pick someone up, even in the woods and in the weeds. But you should think about-- there's a lot of different ways that these can be used out-- again, out in front of the warfighter, you can have convoys of autonomous vehicles that don't have people on them that can move very quickly and move supplies back and forth. Even if they're targeted, at least no human is being killed. You can have them autonomous vehicles out in front of human convoys as well, right? So really making sure that there is no one out there that moving first through the terrain. So generally speaking, there's a lot of ideas in Army and Marine Corps about how to use these vehicles out in front, again, like leading convoys. Or it's like a protective onion almost, right? You can think about a whole set of vehicles that are surrounding higher value assets, whether they're people or tanks or proudly fighting vehicles or XM30, which is a replacement for that. Having autonomous vehicles that can provide sensing and protection for those formations. So we think about them as adding to some of the things that we're developing. How many of these scenarios of you kind of wargamed or run scenarios on out at wherever? Quite a few. I mean, I think people are-- once they start to see the vehicle and they start to think about it, they are coming up with lots of ways of potentially using them. And as we-- one of the things we've been trying to do in the last year is work with a lot of different units. So we've worked with over 20 different units integrating this technology into the unit. For example, the 82nd Airborne, we worked with 382 for almost six months, just embedded with them up through one of their training rotations down at GRTC in Louisiana. And we started out just doing resupply. They were like, this is kind of the obvious use case for this. Let's use this to support our logistics to allow us to move faster. And we were able to do that. And so then they were like, OK, can we get cameras and drones and things like that on it? The answer is yes. And so we started to add different types of payloads to it. They started to use them in more creative ways. And by the end of this, they were using these autonomous vehicles in this four-some-fourths exercise. Not just for resupply, but also for intelligence surveillance and reconnaissance. There were snipers who were using the vehicles as decoys. There's just a lot of things you can start to do. And the first step is really just getting it in the hands of the warfighter, right? Let them think about the problems they're trying to solve, experience the technology, and then start to iterate on potentially new tactics and things like this that you can do with the vehicles. Now, are these truly autonomous? Or is there an operator in the rear that's going to have to be behind each one of these things with a remote control? Yeah, this is a great question. So this is sometimes glossed over. A lot of people say that they can do autonomous things. And people think about uncrewed ground vehicles. People talk about UGVs all the time in Ukraine versus autonomous vehicles. So what are the differences between these things? So remote control basically means you have a controller like in your hand, and you're looking at the thing, and you're driving it. So think about a toy car, right? That's a remote control vehicle. Now, that's fine, except you have to actually look at the terrain yourself, look at the vehicle and determine where it's going to go. That obviously doesn't work beyond line of sight or in areas where you can't see the terrain very well. So the next evolution of this is teleoperation. And people talk about teleoperation a lot with ground vehicles. This is basically where you take a remote control and you look through the vehicle's own sensors, whether it's like camera or thermal sensors. And then you drive the vehicle based on that sensory feedback that you're getting over a network. So you can go beyond line of sight because you just see what the vehicle sees. You don't have to see the vehicle itself. So you can tele-operate something from across the world if you have a really good connection to that vehicle. If you have a good satellite connection or radio connection to the vehicle. And a lot of the work which is being done in Ukraine is tele-operated, uncrewed ground vehicles where you'll actually have a set of people, up to five people, who are controlling each vehicle, and essentially driving it through the terrain using a handheld controller. Or they're looking through potentially either the vehicle's camera's sort of cameras from a drone overhead. And they're just trying to drive this thing around. The thing which is hard about that is that it occupies at least one person's attention at all times because the person's making all of those decisions. But it also is something where if your comms gets disrupted, the thing's just dead, right? Like it requires a human to drive it and the human can no longer connect to it. So one of the strategies is you cut the comms to the vehicle, you use EW or whatever to cut the comms, and then you strike it because it's just a sitting duck. So autonomy is extremely powerful because it allows the vehicle to sense the environment, represent it, plan through the environment, all on board, like at the edge, makes its own decisions. And it doesn't require a human to essentially drive it. So when you have an autonomous vehicle, you can tell it where you want it to go, and it will just go there. And that means that you can focus your attention on something else. That might be another autonomous vehicle. It might be another task. So for example, you can send a vehicle back to resupply you, but you don't have to be focused anymore. You can do your job. The second thing about autonomy, which is important to understand, is because the vehicle's making decisions based on its own sensing and own on-board compute, if your comms get cut, the thing's going to keep going and continue the mission. And so one of the things we're starting to see out of the Russian Ukraine conflict is that the Russians are starting to add autonomy to ground vehicles to deal with the contested comm situation, because it's so hard for them to maintain comms. Autonomy allows the vehicles to continue to move without human oversight. So that's another piece of this. So we think about this as kind of force multiplication and resilience to a contested comms environment. So another thing with autonomy, too, correct me if I'm wrong, is that let's say you have 100 days. You say you have 500 days. And then you have all these other-- we were talked about at the beginning-- all these other companies that are doing shield AI with their X-Bat. We got Seronic, with the autonomous boats. We got Enduro-making stuff. We have Mach-making stuff. All these companies are jumping all out-- lots of autonomous stuff. But just overland AI alone. I mean, let's say that there's 500 of these vehicles. We're talking, I don't know, the invasion of Fallujah. These will all be able to communicate with each other in accomplish the mission without each one being micromanaged. They'll all read off each other, communicate with each other, know what everything is doing, correct? Yeah, so think about that. So this is our vision for the company. So you should think about it as like-- imagine that there's five people that are controlling something like 500 vehicles. And each of those five people, maybe each one's controlling 100 vehicles. And those vehicles are coordinating with each other to complete that mission. So we think about the autonomy, which is onboard the vehicle. This is what we call platform autonomy. It's how each individual vehicle analyzes terrain, makes decisions, decides how to drive. Then there's a notion of mission autonomy where multiple vehicles can coordinate with each other. And the real idea here is to make it really easy for one person to have immense effect on the battlefield, to control many, many different vehicles and the payloads on them. And so going back to the earlier part of the conversation, we view this as something like a gamer who is playing a real-time strategy game, something like StarCraft, where you're controlling like 200 different units. Can a person do that in the real world with real platforms like our ultra-platform? And that's what we're building up towards. So that whole idea is single operator, massive force through potentially hundreds of different vehicles and payloads. Wow. And so we'll move to the point where I just brought up all these other companies. Will it move to the point where these Overland AI vehicles are communicating with shield AI's X-Bat with Seronix boats? I mean, will it get to the point where the entire battle space is coordinated under one brain? So the short answer is that, yes, the ground vehicles will be communicating with aerial vehicles and the different payloads. And they'll be coordinating in order to complete a task. I think one of the things which is really important to keep in mind is, again, these are all just tools. So you really want to enable an operator to achieve their objective as effectively as possible. And that means not just handing everything over to an AI brain, but having essentially AI assistants who are allowing a human to choose courses of action to coordinate ground vehicles or payloads or aerial vehicles more effectively and so on. Let's talk about African Lion. What happened there? OK, so African Lion is an exercise. It's one of the largest exercises in Africa. So it's a place where US war fighters work with partner nations and things like that to essentially train and demonstrate capabilities. We're part of African Lion working with 173rd Airborne. So 173rd Airborne had two of our ultra vehicles with different types of payloads on them. One vehicle had a crows remote weapon station with an M240 machine gun on it. Another vehicle had a rocket propelled breaching system on it. And they were using these vehicles for a breaching operation. So essentially what happened was they sent one vehicle forward with machine gun on it, providing security. And then a second vehicle was quickly following it, moved to a breached point, essentially applied the payload. So, there. It's an explosive line charge, something like a mcklick if people know what that is. But essentially an explosive rope that gets launched out in front of the vehicle and produces a, it blows up and produces like a safe cord or that can then be proved with another vehicle. So they're using this to reduce obstacles out in front of the force as part of an assault. And the thing which is really important about this is that it's extremely dangerous operation. Like when you're doing things like breaching, every area of the defensive obstacle belts are being watched by an adversary and as you move forward to try to reduce those obstacles, the combat engineers, everyone's targeting them. So even in a successful breaching operation, you're expecting something like 50% casualty race. So if you're able to do that with autonomous vehicles, like the way that like 173rd airborne was was demonstrating, that's taking, you know, in their estimate up to almost two platoons about 40 people out of that extremely dangerous situation, you're just sending the machines forward to do that, to work that problem to create the breach in the obstacle belt and then they're able to move through. So they're very excited about the technology, but it's just an example of like warfighters using using our vehicles, putting payloads on them, coordinating as part of their unit's movement maneuver. So pretty exciting to see. It's fascinating stuff. Yeah. Wow. The game has changed a lot. Yeah. 40 years. Yeah. Holy shit. Here's the pitch. It takes two minutes and you'll never think about it again. Download it, hit one button, close the app. That's the entire job. No manual, no settings, no tech background. Nothing for you to get wrong. And then it just runs. Blocking the trackers, the fishing sites, and the surveillance ads across every app on the device all day, quietly without anybody having to do anything. It's first week out. It hit number six on Apple's top downloaded productivity chart, built by former US intelligence professionals. I'm a co-founder, glacier, downloaded on the app store. And remember, privacy isn't paranoia, it's protection. Most of us picked a bank at 16 and never looked back. Sound familiar? You're older now. You handle your money differently, but odds are you're still paying fees to a bank that hasn't done a thing for you in 20 years. That's where chime comes in. Chime is changing the way people bank. They're not like traditional old banks who charge fees and gatekeep perks and rewards. Some offers the most rewarding, fee-free banking, no overdraft fees, no monthly fees, no minimum balance fees, and the benefits stack. You get 5% cash back on your chime card in a category of your choice, like gas or groceries, all well-building credit through regular everyday spending with no credit check. You can grow your money faster with a savings rate that's nine times the national average. And if you're ever in a pinch, spot me, let you overdraft up to $200 fee-free. Join the millions who are already banking fee-free with America's number one choice for banking. Head to chime.com/SRS, that's chime.com/SRS. Sign up now, it only takes a few minutes. Chime is a fintech, not a bank, banking services and chime card provided by Chime's bank partners. If you spend any time off-road like I do, you already know how easy it is to end up second-guessing where you're at, whether you're still on a legal trail or if the route ahead is even worth taking. That's why I've been using ON-X Off-Road. It's an off-road navigation app that shows trails, public and private land boundaries, places to camp and detailed trail info all in one place. And what makes it really useful is the amount of actual trail data. You can check difficulty ratings, terrain details, trail photos, even recent reports from other riders before you head out there. The other big thing is you can download maps ahead of time, because once you lose service out there, your phone is pretty much useless. But with ON-X Off-Road, everything still works offline. And if you're riding with a group, their location sharing feature lets everybody stay on the same map so nobody gets lost or separated. What I like most about it is it gives me a lot more confidence when I'm exploring somewhere new. I'm not wasting time backtracking, accidentally ending up on private land or trying to figure things out once I'm already deep into the trail system. I can plan ahead, know what I'm getting into, and spend more time actually enjoying the ride instead of worrying about navigation the whole time. Which ON-X Off-Road in the App Store or Google Play? Again, that's ON-X Off-Road in the App Store or Google Play. All right, so we're back from the break. And one thing I'm kind of wondering is, what other, I mean, you mentioned Russia was doing some of this, correct? What about China? They should be pretty on the ball with just about everything. Yeah, we don't have, at least I don't have detailed knowledge. Of everything that they're doing. But we do read academic papers that come out of Chinese institutions, and they're certainly working on ground robots, and I think there's several different types. They've been working on robotic dogs and integrating those types of robots into infantry formations. And they're also working on off-road ground vehicle autonomy, you know, similar to some of the stuff that we're doing. We think we're pretty far ahead right now, but, you know, they're working hard to catch up. So what kind of, is there any, what other kind of weapon systems do you think will be putting on to these, these ultras? I think for any of, you know, you think about uncrewed ground vehicles and autonomous ground vehicles, you can put, it depends on the size of vehicle, you can put, like, almost anything on them, right? So any type of vehicle that you have today, which can carry anything from, you know, a very small vehicle, might be a couple hundred pounds, up to tens of thousands of pounds, you can put on a potentially autonomous vehicle. So that's, you think about smaller things, might be smaller, remote weapon stations, might just be sensors, right, radar, optics, things like this, but much larger vehicles, large missile systems, right? So ship interdiction missiles, anti-air missiles, all sorts of things. You could put pretty much anything on those. Pretty much anything. Are you familiar with Epirus? Yes. So you could put Leonidas, mounted on one of these damn things and then you have, yeah, drone defense for an entire fucking battalion. Yeah. Usually when we're thinking about drone defense, you know, it's going to be a layered defense. So you're going to want maybe something like, you know, Epirus's system, but you might also want kinetic air defense. That might be drones, like, you know, drone interceptors, it might be machine guns, but you can imagine a whole set of these, you know, many different vehicles with different defensive systems, creating that layered defense, and we think it'll probably be something like this. So that, you know, even if drones are getting through one type of defense, they're getting hit by another. I mean, even if you lose some of your vehicles or systems, they're more to take their place. Let's talk about the technical mode, the team, and the peer threat, sure. So the way that we think about the technology that we're developing, we really are thinking about it as, you know, developing an autonomous version of core battlefield functions, right? So like, either the ISR or breaching for as examples, where you have multiple vehicles with payloads that are performing a task. Now in order to do any of that, in order to have a set of autonomous vehicles with payloads that's, you know, doing something complicated, think about like first principles, like what do you have to do first? Well, you have to be able to carry stuff and move it, you know, like move it from one place to another in the environment. And when you look at our history as a company, we started with what we think is the hardest problem there first, which is being able to understand terrain and move through that terrain. Even that basis, you can then build off of that. So, you know, one of our strongest technical modes is the fact that we're the best in the world at being able to see and understand the terrain. stand terrain and move vehicles through it. We can put payloads like pretty much anywhere that a vehicle can drive, right? And so once you have that, as you now combine this to say, like instead of just moving one vehicle or one type of payload, now I'm moving multiple vehicles, multiple payloads, that foundation allows you to build up those capabilities. So we think that that's one of our biggest technical motes is we have just this incredible team coming from deep tech places like Waymo and Crews and self-driving car companies, some of the top artificial intelligence labs. They're working for us focused on these problems and creating this AI essentially that allows you to move payloads in the environment. So that's technical mode. Can't quite remember the other questions, but that's really a foundation of everything. - I mean, one thing that I didn't ask, I think we started talking about it out there and then I said, we'll come back in and discuss it 'cause it was so damn hot out there. But what does it look like when we're talking about controlling all these autonomous vehicles, especially when it comes to hundreds? I mean, it's not remote control. So what does it, I mean, we'd kinda discuss a little bit out there but what does it look like for the team or the person or whatever is controlling hundreds of these at once? - Yeah, that's a great question. So you're not gonna be just like looking through the vehicle sensor and remote controlling it because you have hundreds of them. How can you do that? So the way that we have designed the software is that you essentially have overhead maps, something like a satellite map, and you're able to see all of the different vehicles on that map, all the different autonomous assets on the map. So that gives you, you know, and you can zoom out and you can look at, so you can look at terrain features, you can look at where all the vehicles are, you can see what payloads are on the different vehicles. And then the question is, how do you now coordinate and control hundreds of those assets? And you need to be able to do that by selecting vehicles, grouping vehicles, telling groups of vehicles to go to like one place or another, telling them to execute a certain payload, you know, at a particular location, so that you can kind of quickly move between, you know, sets of vehicles and essentially tell them where to go. Again, a lot like a real time strategy game. We're also working on tools. - So this looks like, this is what we were talking about again, this looks like, sounds weird, looks like world of warcraft. - Yeah, yeah, you're sending a group of things, giving it a task. - It's not under the next thing. - Yeah, and I think, you know, one of the reasons why we've looked to some of these types of games is because it's one of the few places where humans are actually controlling something like hundreds of different assets. Starcraft is a great example of that in Warcraft, but, you know, the tools that people use for grouping, for moving, for coordinating assets in those types of games are things we can take those ideas and apply them here as well so that a single operator can control many, many different assets. Another thing that we're developing is things like AI assistance, which can help to suggest, you know, particular tasks, so you can just say, like, hey, I need to move 10 vehicles, you know, into this area, provide a reconnaissance here, and it will tell me how those vehicles are going to move, suggest to me solutions to these problems so that I can make decisions faster. So this is a part of, like, essentially using AI to increase your decision advantage. Now we're not saying we hand over, you know, the actual decision making to the AI, but the AI can suggest solutions that can help you to handle, you know, more assets at once. - Wow, wow. I got a hot question for you. All right, ready? In 1941, 353 Japanese planes came off six carriers and hit Pearl Harbor in under two hours. They sank or crippled all aid battleships and killed over 2400 Americans. And here's the part people forget, we had warnings. We'd broken their codes. Our own ambassador flagged a possible attack a year earlier and Congress later said the real failure wasn't intelligence. It was imagination. Nobody could picture it until the harbor was on fire. Today, China's arming robot dogs, they build 90% of the world's drones and they just flew a mothership not long ago that launched a hundred kamikaze drones in a single swarm. The warnings are everywhere again. Are we sleepwalking into a robotic Pearl Harbor? - I hope not, but I think it's correct there are warnings everywhere. We're seeing robots used in the battlefield again in Ukraine that there's both by the Ukrainians and Russians. So we and I think other defense tech companies are certainly paying attention to this. We are trying to understand what capabilities our adversaries have and also ensure that the US military has similar better capabilities. And so we think about these sorts of things all the time. It's the new technology, which is going to be defining like these next conflicts. And I think it's all of our jobs to ensure that the decision makers understand that this technology exists and what it's capable of. So I think we are certainly aware of it. I think we're pushing the military to adopt new technology and defenses against it, like as quickly as possible. But, you know, there may be work to be done. I sometimes these things aren't really real to people until unfortunately they actually experience them. So our best bet is to really be watching and taking seriously what is happening in these conflicts and what our adversaries are developing. I mean, do you feel that the Department of War has taken this seriously? Do they understand the capabilities that you guys have? I mean, when you looking at that thing out there, what it's capable of, it's very surprising and almost alarming to me that they, you know, 15, 15, you have a contract for 15. Yeah. Why don't you have a contract for 10,000? Yeah, I mean, it should be, you know, I think the Department of Defense is trying to, our Department of War is in the services are trying to move faster. But the procurement system was really designed for a different era and a different type of war. And that's something that I think, you know, all of the defense tech founders who you've had on here will probably agree with, right? And something that we're all fighting hard against. Now, we've seen the Department of War speed things up. And, you know, some of the, like I was saying, but a little bit earlier, we have actually gone from, you know, university research to production contract, even if it's, you know, a relatively small initial one, in like three years, which is about right, maybe a little slow for like, you know, the tech sector. But that's lightning fast for Department of War. And we've made use of defense innovation unit and a lot of the new tools, the app fit contracting process and things like that that have come online recently. So, you know, I do want to credit the Department of War for moving faster, but there's still so much work to be done in order to move at the speed that we need to move at. So, you know, it's something I think like we're all working on and trying to. - Are other countries better customers than our own country? - Or is there a problem for like. - Not necessarily like their own stuff or stuff that America, not overland AI for things that America, Americans, like yourself are developing here in this country. Are other countries, maybe Ukraine. I mean, I think, you know, I've heard of these issues happening and here we have the best that the fucking world has to offer right here, developing groundbreaking new technology that's going to change how wars fought. It's, I mean, it's, it's, and I mean, it's such an upgrade. - Especially with a company like, all these companies, man, like which everybody, you know, everybody that's sat across with me. in the defense tech sector. What you guys are developing is it's fucking incredible. Yeah, you know, but I see it being utilized in Ukraine. You know what I mean? And I see it bothers me when when when we're sending all this stuff. Maybe we aren't, but see things similar being utilized in Ukraine and it just seems like our country is not taking advantage of the talent that we have here. It's fucking scares me. Yeah, I mean, we there's there's kind of a like we have all this we have all this talent like there's things to be concerned about and I think there's there's things to be hopeful about. So like on the one hand, you know, you're I think you're right, I think countries like Ukraine. I mean, they're adopting tech as fast as they possibly can. There's just a statistic out saying something like there's been two million casualties in in that conflict, right? And about a million and a half Russian and about half a million Ukrainian. And one of the ways that Ukraine has been able to continue to fight and hold off the Russians is through, you know, all of this technological innovation. It's existential for them. They they have to do this or they will lose, right? Like and so they will do whatever it takes to to win here. It's not existential for the US yet. And I think that is means that adoption is just slow like people aren't feeling the the pressure to to do it. And I think that, you know, it's unfortunate, of course, because we have a little bit of luxury right now like we are not in a conflict like the Ukrainians. So we have the space to be able to potentially build this technology and transform our forces. But we don't have the urgency, right? And so I think I think that's part of the problem. On the on the positive side, I do think that we have the best minds in the world here. I think we can do it. But it's again, it's a matter of arguing it. We are doing it. We are doing government that's not bumping up with you guys. But it could be it could be done faster, right? Like we could accelerate this. And you know, that's a fight that we just have to continue to to make right to continue to fight to get this tech into the government faster. And I think going back to thinking about some of the DARPA stuff earlier, like we're creating incredible new technology. But then you really got to fight to like actually get the Army and Marine Corps and services to actually, you know, try it, iterate, you know, ultimately, ultimately use it and incorporate it. If we were attacked tomorrow, do you think we could survive a sea or ground invasion? I mean, I think I think we would absolutely survive a sea or ground invasion. I think we would mobilize very quickly. I think those the urgency would be there. And you know, we would do whatever it took to win. So I'm, you know, I think very positive and optimistic on that. Like I think when we're pressed, we can do a lot. But I also think that, you know, we could be better prepared. So we too. Last thing. Yeah. And you factoring. Are you guys manufacturing these yourselves? Yeah. So scale up. So we're trying to scale as quickly as possible. We are manufacturing ultra vehicles. I think I mentioned before that they're based on Polaris, you know, engines and drive trains and things like that. But we're upgrading the suspension. We're, you know, putting in the payload deck, adding the compute, the sensors and everything. We've got factory running in Seattle that's doing this right now. But and we, I think we've like five X manufacturing over the last last six months or so. And we just, we need to do a lot more. So we're building up that capability as fast as possible. And again, the more support we get from, from the government, the more we actually work with war fighters, the more demand there is. And there's, there's a ton of demand right now. So we're scaling as quickly as we can. Right on. Well, Byron, I really appreciate you coming. It was an honor to interview you and a love, love everything overland AI is doing and that thing out there is super impressive. Thank you. Yeah. Thank you so much for having me. I really appreciate you taking the time to learn a little bit more about what we're doing and inviting us out and being able to show you some of the things that we're building. So thank you. Thank you. No matter where you're watching the Sean Ryan show from, if you get anything out of this at all, anything, please like, comment and subscribe. And most importantly, share this everywhere you possibly can. And if you're feeling extra generous, head to Apple podcasts and Spotify and leave us a review.