Latest / NO FILTER / Byron Boots - 5 People Could Control an Army of 500 Autonomous Vehicles _ SRS _336

Transcript

Byron Boots - 5 People Could Control an Army of 500 Autonomous Vehicles _ SRS _336 · NO FILTER

• 0:05
  Speaker 1: Firing boots.
• 0:06
  Speaker 2: Welcome to the show. Thanks for having me. It's, it's
• 0:09
  Speaker 2: awesome to be here.
• 0:10
  Speaker 3: It's awesome to have you.
• 0:11
  Speaker 1: So I, I can't even remember.
• 0:12
  Speaker 3: I think I actually found you guys on LinkedIn, which
• 0:15
  Speaker 3: I'm never on. And, uh, and, uh, I think I
• 0:18
  Speaker 3: saw like a video on LinkedIn or something of, of
• 0:21
  Speaker 3: Overland AI and then started, found you guys on YouTube
• 0:26
  Speaker 3: and started following you and then found out APC is
• 0:29
  Speaker 3: an investor. And, um, so, or led, led you around and, um, So, yeah,
• 0:36
  Speaker 3: I wanted to get in touch and love what you
• 0:38
  Speaker 3: guys are doing. Looks, you know, I've had lots of
• 0:41
  Speaker 3: tech giants on here, a lot of drone stuff, Saronic
• 0:46
  Speaker 3: with the water stuff. I guess they, I didn't even know.
• 0:48
  Speaker 3: Your guy just told me out back that they just
• 0:50
  Speaker 3: hit the, they hit an Iranian port.
• 0:53
  Speaker 4: Yeah, pretty, pretty amazing. So, you know, first of all,
• 0:56
  Speaker 4: it's an honor to be on the show and in
• 0:58
  Speaker 4: that company. I mean, that's incredible. But, yeah, just in
• 1:02
  Speaker 4: the last day or so, there was news of, uh,
• 1:05
  Speaker 4: Saronix boats being used, um, in an offensive operation in
• 1:09
  Speaker 4: the Strait of Hormuz.
• 1:10
  Speaker 3: Wow.
• 1:11
  Speaker 5: Yeah.
• 1:11
  Speaker 1: And so you're the ground guy.
• 1:13
  Speaker 2: That's right.
• 1:13
  Speaker 3: You're bringing up autonomous wheels, vehicles to, uh, to land warfare. So, uh,
• 1:20
  Speaker 3: really excited to get into, to dig into this, but, uh,
• 1:23
  Speaker 3: let me, let me kick it off with a introduction here.
• 1:26
  Speaker 1: Byron boots.
• 1:27
  Speaker 3: You're the co-founder and CEO of Overland AI, a company
• 1:31
  Speaker 3: building autonomous ground vehicles for the U S military. Before
• 1:35
  Speaker 3: founding Overland AI, you earned your PhD in machine learning
• 1:39
  Speaker 3: from Carnegie Mellon, became a professor at the University of Washington,
• 1:43
  Speaker 3: and led the winning team in DARPA's Racer Off-Road Autonomy Program.
• 1:49
  Speaker 3: Overland AI has raised more than $ 140 million and become
• 1:53
  Speaker 3: the first autonomous ground vehicle company to win a production
• 1:57
  Speaker 3: contract with a fully integrated hardware and software platform. Congratulations.
• 2:03
  Speaker 3: Your vehicles are already being used by military units around
• 2:07
  Speaker 3: the world to move supplies, support operations, and reduce risk
• 2:10
  Speaker 3: for soldiers in the field. Welcome to the show.
• 2:15
  Speaker 2: Well, thank you so much.
• 2:17
  Speaker 1: We've got a lot to talk about here.
• 2:19
  Speaker 3: It's been a minute since I've talked to somebody that's
• 2:22
  Speaker 3: doing the kind of stuff that you're doing. But before
• 2:26
  Speaker 3: we get going, I've got a couple things to crank
• 2:29
  Speaker 3: out here. Everybody gets a gift.
• 2:31
  Speaker 6: Oh, wow.
• 2:33
  Speaker 1: Those are autonomous gummy bears.
• 2:37
  Speaker 2: I'm excited. I love gummy bears.
• 2:40
  Speaker 5: Right on.
• 2:41
  Speaker 1: And I got a question for you.
• 2:45
  Speaker 3: I got a Patreon account. It's a subscription account. And
• 2:49
  Speaker 3: so they're the reason I get to sit down here
• 2:52
  Speaker 3: with you today. And so they get the opportunity to
• 2:55
  Speaker 3: ask every single guest a question. So this is from
• 2:58
  Speaker 3: Thomas W. You've dedicated your career to advancing robotics and AI,
• 3:03
  Speaker 3: including technologies with defense applications. From your perspective, what responsibility
• 3:09
  Speaker 3: do scientists and engineers have to ensure these innovations ultimately
• 3:13
  Speaker 3: reduce human suffering rather than prolong conflict? And do you
• 3:19
  Speaker 3: believe AI and autonomous systems could one day become tools
• 3:23
  Speaker 3: that prevent wars through deterrence and de-escalation? Or do they
• 3:29
  Speaker 3: risk making armed conflict more frequent and easier to justify?
• 3:34
  Speaker 4: That's a great, great question. You know, I think autonomous systems,
• 3:41
  Speaker 4: they're like any other tool. The way that I think
• 3:43
  Speaker 4: about them is really a tool, a technology, and in
• 3:48
  Speaker 4: the context of defense, it is, you know, something which
• 3:52
  Speaker 4: allows a warfighter.
• 3:53
  Speaker 2: To, you know, be safer, right?
• 3:56
  Speaker 4: So, you know, it reduces exposure to pulls them away
• 3:59
  Speaker 4: from the point of contact, and then also potentially provides
• 4:04
  Speaker 4: force multiplication on the battlefield. And I'm sure we'll get
• 4:07
  Speaker 4: into some of these things. But it is a tool
• 4:11
  Speaker 4: which is used by humans. And so, you know, how
• 4:14
  Speaker 4: you use them is really, I think, a human question. So,
• 4:22
  Speaker 4: you know, in the context of saving lives, I think
• 4:25
  Speaker 4: that they will save lives for Our warfighters on the battlefield,
• 4:31
  Speaker 4: it's very clear how they do that. But, you know,
• 4:35
  Speaker 4: they also can serve as a deterrence like any technology might.
• 4:39
  Speaker 4: You know, if we have significantly stronger robotic systems at
• 4:44
  Speaker 4: our disposal, then, you know, we're going to be a
• 4:47
  Speaker 4: little bit, you know, tougher to defeat in the battlefield
• 4:51
  Speaker 4: and our adversaries will see that. And so in that way,
• 4:55
  Speaker 4: you know, they can certainly serve as a deterrence as well.
• 4:58
  Speaker 3: Right on. I mean, yeah, watching some of the models
• 5:00
  Speaker 3: you showed me outside and then, you know, the videos
• 5:05
  Speaker 3: and how they're going to be integrated in with the
• 5:08
  Speaker 3: warfighters in combat. I mean, it's, you know, as a
• 5:13
  Speaker 3: former SEAL, seeing what, seeing the, you know, war, what
• 5:21
  Speaker 3: war has developed into is just, I mean, it's fascinating.
• 5:24
  Speaker 5: Yeah.
• 5:26
  Speaker 3: It's been over 20 years since I've been on the
• 5:28
  Speaker 3: ground in a war. Well, I guess not. It's still
• 5:32
  Speaker 3: been over 10 years. But, I mean, I already have
• 5:37
  Speaker 3: tons of questions, and I can see so many different
• 5:40
  Speaker 3: applications where this would be useful in just so many
• 5:44
  Speaker 3: different scenarios.
• 5:45
  Speaker 2: Yeah, we should definitely get into it.
• 5:47
  Speaker 4: Before we do that, though, I do want to give
• 5:49
  Speaker 4: you a gift as well, if you don't mind.
• 5:52
  Speaker 3: So let me....
• 5:55
  Speaker 2: Come over here and we got you a chair. Now
• 6:00
  Speaker 2: this isn't just any chair.
• 6:01
  Speaker 3: Right on.
• 6:03
  Speaker 7: This is, it looks kind of like an office chair,
• 6:07
  Speaker 7: but this is actually a seat from one of the vehicles.
• 6:11
  Speaker 7: So as I was explaining earlier, we pull the seats
• 6:14
  Speaker 7: out of the Polaris Rages and we turn them into
• 6:17
  Speaker 7: those autonomous vehicles that we saw outside. Well, what do
• 6:21
  Speaker 7: you do with the seats once you've pulled them from
• 6:24
  Speaker 7: the vehicle?
• 6:25
  Speaker 5: We make them like the chairs. So we made one
• 6:27
  Speaker 5: for you.
• 6:28
  Speaker 6: Oh, man.
• 6:29
  Speaker 3: Thank you.
• 6:29
  Speaker 2: You can check it out.
• 6:30
  Speaker 3: This is awesome.
• 6:31
  Speaker 5: Yeah, yeah.
• 6:37
  Speaker 3: Not bad, not bad.
• 6:39
  Speaker 5: There you go. Put this in the office.
• 6:42
  Speaker 3: Thank you.
• 6:42
  Speaker 2: Yeah, of course.
• 6:43
  Speaker 3: That's awesome. All right, Byron. So before we get into
• 6:47
  Speaker 3: everything Overland AI, let's do a little backstory on you.
• 6:51
  Speaker 3: Where did you grow up? How did you get into
• 6:53
  Speaker 3: this stuff? I mean, what's the backstory here?
• 6:56
  Speaker 4: Yeah, I, I had, I had a whole career, uh, before, um,
• 6:59
  Speaker 4: moving into, into defense tech. So, um, I grew up
• 7:02
  Speaker 4: outside of New York in, in Connecticut. Um, uh, I
• 7:07
  Speaker 4: was into computers and, and, uh, you know, was in
• 7:11
  Speaker 4: the boy Scouts and played sports and had, I think
• 7:14
  Speaker 4: a pretty typical, um, upbringing. Um, so, you know, that's
• 7:19
  Speaker 4: maybe where things got started. Just love of the outdoors and, uh,
• 7:24
  Speaker 4: taking apart computers and playing video games and doing all
• 7:27
  Speaker 4: the sorts of things that kids often do.
• 7:29
  Speaker 3: Did you watch The Terminator growing up? I sure did, yeah.
• 7:32
  Speaker 3: Did you?
• 7:33
  Speaker 4: So Terminator 2 was an unbelievable, unbelievable movie. And happy
• 7:44
  Speaker 4: to talk about that a little bit more in the
• 7:45
  Speaker 4: context of what we're building.
• 7:48
  Speaker 2: But Obviously, robotics.
• 7:50
  Speaker 4: And science fiction were something that I really enjoyed.
• 7:54
  Speaker 3: Were you a gamer?
• 7:56
  Speaker 2: Yeah, I used to play.
• 7:58
  Speaker 4: So back when I was in high school, I used
• 8:00
  Speaker 4: to play StarCraft quite a bit, StarCraft I.
• 8:02
  Speaker 2: That was before StarCraft II came out.
• 8:05
  Speaker 4: So real-time strategy games, I played a lot of games
• 8:08
  Speaker 4: like that. We talked about Warcraft before. I used to
• 8:11
  Speaker 4: play that too. So that was really what I was
• 8:15
  Speaker 4: most drawn to. But, yeah, I mean, love computer games.
• 8:20
  Speaker 3: How, I mean, well, we'll get into it later. I
• 8:23
  Speaker 3: was going to ask how similar, you know, is what's
• 8:27
  Speaker 3: happening today is controlling one of those games. We'll get
• 8:29
  Speaker 3: into that in a little bit. So where did you
• 8:32
  Speaker 3: go to school? What did you go to school for?
• 8:35
  Speaker 4: So I went to college at a small liberal arts
• 8:37
  Speaker 4: school called Bowdoin College. It's in Maine. And, you know,
• 8:42
  Speaker 4: I spent four years there. I was a computer science
• 8:46
  Speaker 4: and philosophy double major as an undergrad. I started out
• 8:51
  Speaker 4: really thinking about, you know, I love computer science, like
• 8:55
  Speaker 4: I said, you know, all through high school. I also
• 8:56
  Speaker 4: really liked history and read a lot of history. And
• 8:59
  Speaker 4: so when I went to college, I was thinking about
• 9:02
  Speaker 4: maybe double majoring in computer science and history. I thought
• 9:05
  Speaker 4: it would be cool to have a more technical degree
• 9:07
  Speaker 4: and something which is more humanities-oriented with history. But I
• 9:16
  Speaker 4: quickly sort of figured out my first year that, while
• 9:21
  Speaker 4: I love history, it involved tons of reading and things
• 9:24
  Speaker 4: like this that I really like to do, but also
• 9:27
  Speaker 4: involved foreign languages. And you need to actually read about
• 9:33
  Speaker 4: history through contemporary sources and you know, in the language
• 9:38
  Speaker 4: that folks wrote in. And so it was something I
• 9:41
  Speaker 4: was not great at. I was not particularly good at languages,
• 9:46
  Speaker 4: didn't really enjoy them. And so, you know, I started to,
• 9:51
  Speaker 4: I took a couple courses in philosophy. I started out
• 9:54
  Speaker 4: with a course called Logic and Formal Systems, which was
• 9:57
  Speaker 4: really delving into, you know, formal frameworks for understanding arguments
• 10:03
  Speaker 4: and analyzing arguments. And I really started to fall in
• 10:07
  Speaker 4: love with that.
• 10:08
  Speaker 2: And so.
• 10:11
  Speaker 4: I brought that together with computer science and majored in
• 10:15
  Speaker 4: both of those areas as an undergrad.
• 10:17
  Speaker 3: Interesting. You were fascinated with the brain.
• 10:19
  Speaker 4: Too, right?
• 10:20
  Speaker 2: Yeah, yeah, yeah.
• 10:21
  Speaker 4: So, I mean, in undergrad, I was, you know, I
• 10:27
  Speaker 4: was taking computer science and, you know, took courses in
• 10:29
  Speaker 4: artificial intelligence and started to do research in robotics.
• 10:34
  Speaker 2: And this is back, you know, over 20 years ago.
• 10:36
  Speaker 3: They had artificial intelligence courses 20 years ago?
• 10:39
  Speaker 2: Yeah, it was pretty interesting.
• 10:40
  Speaker 4: So my university is a small school. There was only
• 10:43
  Speaker 4: four faculty in the computer science department there. And at
• 10:48
  Speaker 4: the time, computer science was seen as like an offshoot
• 10:51
  Speaker 4: of mathematics. And so, you know, a lot of these
• 10:55
  Speaker 4: smaller schools had combined departments with computer science and mathematics,
• 10:59
  Speaker 4: and you.
• 10:59
  Speaker 2: Take a lot of courses in both areas.
• 11:02
  Speaker 4: But at Bowdoin, two of the four professors were actually
• 11:07
  Speaker 4: folks who studied artificial intelligence. So it was something, you know,
• 11:11
  Speaker 4: it's been around for a long time. I mean, people
• 11:14
  Speaker 4: were working on aspects of AI. you know, back in
• 11:17
  Speaker 4: the 70s and 80s. But it was just starting to
• 11:21
  Speaker 4: kind of come to the forefront and be an area
• 11:25
  Speaker 4: that was really starting to accelerate around, you know, 2000
• 11:29
  Speaker 4: when I was an undergrad.
• 11:32
  Speaker 2: So I started to study AI there.
• 11:35
  Speaker 4: And then in philosophy, I was thinking about things like
• 11:38
  Speaker 4: philosophy of mind and philosophy of science. And you're just
• 11:42
  Speaker 4: kind of trying to understand how the human mind worked.
• 11:46
  Speaker 2: Well, Yeah.
• 11:49
  Speaker 3: Okay. I think that's an interesting discussion itself. Yeah. How,
• 11:52
  Speaker 3: I mean, how deep did you get into that before you,
• 11:55
  Speaker 3: you kind of switched gears? Yeah.
• 11:56
  Speaker 4: So, uh, as a, um, again, as an undergrad, I
• 12:00
  Speaker 4: double majored in computer science and philosophy. And, um, as
• 12:03
  Speaker 4: when I graduated, I was, you know, thinking about what
• 12:07
  Speaker 4: I wanted to do next. Um, I initially took a
• 12:11
  Speaker 4: job at a robotics company as an engineer, um, basically
• 12:16
  Speaker 4: working on problems related to perception and mapping and robotic systems.
• 12:21
  Speaker 4: So these were mobile robots, much smaller than the ones
• 12:26
  Speaker 4: that we.
• 12:26
  Speaker 2: Just saw outside.
• 12:28
  Speaker 4: So robots that are about this big that moved around
• 12:31
  Speaker 4: inside of buildings. And you have to determine where they
• 12:34
  Speaker 4: are and how they get from one place to another
• 12:36
  Speaker 4: and things like that. So I worked as an engineer
• 12:38
  Speaker 4: working on those sorts of problems. But I was really
• 12:41
  Speaker 4: thinking about how you know, kind of like, what do
• 12:45
  Speaker 4: I want to do next? I knew I didn't want
• 12:46
  Speaker 4: to just be, you know, kind of working as a
• 12:49
  Speaker 4: software engineer. I wanted to go back to school. And
• 12:52
  Speaker 4: the question was, like, what area should I study? So
• 12:55
  Speaker 4: I really liked computer science. I really liked philosophy. And
• 12:59
  Speaker 4: one of the things that I started to think about
• 13:01
  Speaker 4: was cognitive science, you know, just sort of how the
• 13:04
  Speaker 4: mind works. So with AI, you know, you're trying to
• 13:08
  Speaker 4: program a computer that can almost like think like a human,
• 13:12
  Speaker 4: that can perceive the world, that can understand it somehow.
• 13:16
  Speaker 2: And then in.
• 13:17
  Speaker 4: Philosophy, you're really thinking through language and, you know, by
• 13:22
  Speaker 4: writing arguments, thinking about, you know, how does the human
• 13:25
  Speaker 4: mind work? How does it contend with, you know, reality,
• 13:28
  Speaker 4: things like this. But the piece that I was missing
• 13:31
  Speaker 4: was actual neurobiology, right? Like the human nervous system, the
• 13:36
  Speaker 4: substrate of Um, of, of the mind. And so, um,
• 13:42
  Speaker 4: I decided that before I went back and entered into
• 13:46
  Speaker 4: a graduate program, I needed to.
• 13:47
  Speaker 2: Learn more about neuroscience.
• 13:51
  Speaker 4: Um, and so I, I managed to get a job
• 13:55
  Speaker 4: at Duke university, uh, in a neurobiology lab, um, studying
• 14:00
  Speaker 4: human perception. So I worked as an engineer for about
• 14:03
  Speaker 4: a year and then I went to Duke and I
• 14:05
  Speaker 4: worked there for two years. And this was really it
• 14:08
  Speaker 4: wasn't a graduate program. It was just working in a
• 14:11
  Speaker 4: neurobiology lab. And I was auditing courses on on neuroscience
• 14:17
  Speaker 4: and neurobiology while I was there trying to learn, you know,
• 14:20
  Speaker 4: how does how does the mind work?
• 14:22
  Speaker 3: I mean, wow. How the human mind perceives the world.
• 14:28
  Speaker 4: Yeah.
• 14:29
  Speaker 3: How I mean, did you did that? Is that helpful
• 14:33
  Speaker 3: in what you do today?
• 14:35
  Speaker 2: It is. It's pretty interesting.
• 14:38
  Speaker 4: So the lab that I was working in was really
• 14:40
  Speaker 4: focused on trying to understand how humans perceive the world.
• 14:45
  Speaker 4: And so let me just give you an example of
• 14:47
  Speaker 4: why this is difficult and interesting. So when you look
• 14:54
  Speaker 4: out at an environment like this room, there's light which
• 14:59
  Speaker 4: bounces off of surfaces. It comes back and it moves
• 15:03
  Speaker 4: through your eye and essentially is projected on your retina.
• 15:06
  Speaker 4: So for each of your eyes, there's a 2D projection
• 15:10
  Speaker 4: of light from the room. And the question is, how
• 15:15
  Speaker 4: do you go from that 2D projection, that 2D image,
• 15:20
  Speaker 4: to understanding what's actually out in the world, like the
• 15:23
  Speaker 4: 3D environment, the surface reflectance properties of, on things like
• 15:29
  Speaker 4: the wall or the carpet or whatever. How do you
• 15:34
  Speaker 4: sort of solve that problem? And it's called the inverse
• 15:37
  Speaker 4: optics problem. So it's the notion that you have a
• 15:40
  Speaker 4: 2D image and you're trying to kind of understand this
• 15:42
  Speaker 4: complex 3D world. And the challenge is that there's actually
• 15:47
  Speaker 4: not an easy solution to this. Because you're moving from
• 15:51
  Speaker 4: essentially like three dimensions to two dimensions, information is lost.
• 15:56
  Speaker 2: And so.
• 15:58
  Speaker 4: Another way to think about this is that an infinite
• 16:00
  Speaker 4: number of different worlds could have produced the same visual
• 16:05
  Speaker 4: image on your retina. And this manifests itself through illusions.
• 16:11
  Speaker 4: So there are certain types of illusions. There's something, for example,
• 16:14
  Speaker 4: called an Ames room, where when you look at the room,
• 16:18
  Speaker 4: it looks like a rectangular room. But in fact, it
• 16:21
  Speaker 4: has this kind of crazy shape, something you can look
• 16:24
  Speaker 4: up maybe later. The interesting thing about that is just
• 16:31
  Speaker 4: the fact that something that appears to you to be,
• 16:37
  Speaker 4: you know, like a normal rectangular room is actually something
• 16:39
  Speaker 4: completely different.
• 16:40
  Speaker 2: So that is just an.
• 16:43
  Speaker 4: Example of one of these optical illusions. Now, the interesting
• 16:46
  Speaker 4: thing about illusions is that they basically... everything you see
• 16:54
  Speaker 4: in some ways is an illusion, right? So they're not outliers.
• 16:57
  Speaker 4: It's not like every once in a while your mind
• 17:00
  Speaker 4: makes a mistake and you kind of see the world incorrectly.
• 17:04
  Speaker 4: You are always inferring some world that is not quite
• 17:08
  Speaker 4: what is actually out there. It's the rule, not the exception.
• 17:13
  Speaker 4: And so this is, it forms almost.
• 17:16
  Speaker 2: A philosophical problem.
• 17:18
  Speaker 4: It's like, if you are looking at the world, but
• 17:22
  Speaker 4: you can't actually infer what generated, you know, the images
• 17:27
  Speaker 4: that you see. How do you even, you know, how
• 17:30
  Speaker 4: do you even interact with it? Like, how do you
• 17:34
  Speaker 4: continue to exist if you're not seeing things properly? And so,
• 17:39
  Speaker 4: you know, the conclusion, one of the conclusions that we
• 17:41
  Speaker 4: came to was that really the way that you see
• 17:47
  Speaker 4: the world is whatever way is necessary to allow you
• 17:52
  Speaker 4: to continue to persist. So we kind of think about
• 17:55
  Speaker 4: this as like, you see the world in an evolutionarily
• 17:59
  Speaker 4: sort of appropriate way, in a way which informs your
• 18:02
  Speaker 4: actions so that you kind of do the right things,
• 18:05
  Speaker 4: you continue to exist, you can ultimately reproduce and continue on.
• 18:10
  Speaker 4: And so, it's just one of the problems that we
• 18:15
  Speaker 4: wrestled with.
• 18:16
  Speaker 2: Now, what does that mean?
• 18:18
  Speaker 4: It means that your perception of the world is really
• 18:20
  Speaker 4: shaped through experience. You're just perceiving the world in the
• 18:25
  Speaker 4: best possible way for you to take actions. And some
• 18:30
  Speaker 4: of these fundamental ideas actually carried through into the work
• 18:35
  Speaker 4: I did in graduate school and even some of the
• 18:37
  Speaker 4: things that we do today with the systems that we build,
• 18:42
  Speaker 4: the robotic systems that we build.
• 18:44
  Speaker 3: Very interesting.
• 18:45
  Speaker 6: Wow.
• 18:46
  Speaker 4: Wow.
• 18:49
  Speaker 3: When we first started building the Sean Ryan Show storefront,
• 18:53
  Speaker 3: we didn't have everything figured out. We had products, merch, ideas,
• 18:57
  Speaker 3: and a growing audience. But turning that into something people
• 19:01
  Speaker 3: could actually shop from and keep them coming back? That
• 19:04
  Speaker 3: takes the right setup. And that's why we built it
• 19:07
  Speaker 3: on Shopify. Shopify gives you what you need to start
• 19:10
  Speaker 3: selling without having to piece a bunch of different systems together.
• 19:14
  Speaker 1: You can build the storefront.
• 19:15
  Speaker 3: Manage products, take payments, track orders, and keep the business
• 19:19
  Speaker 3: moving from one place. And if you're just getting started,
• 19:22
  Speaker 3: you don't need to know everything on day one. Shopify
• 19:25
  Speaker 3: makes it simple to launch your store and start selling
• 19:28
  Speaker 3: in a few steps. They also have design templates and
• 19:32
  Speaker 3: AI site building tools to help you get the storefront
• 19:35
  Speaker 3: looking right, without starting from scratch. And once people are
• 19:38
  Speaker 3: ready to buy, ShopPay makes checkout fast and easy, which
• 19:43
  Speaker 3: is huge. You don't want somebody ready to support your
• 19:46
  Speaker 3: business and then lose them because you got a clunk checkout.
• 19:50
  Speaker 3: I like Shopify's product so much. It's so simple. I
• 19:54
  Speaker 3: had the CEO and founder, Toby, on my show. So
• 19:58
  Speaker 3: if you've been sitting on an idea, waiting until you
• 20:00
  Speaker 3: have it figured out, stop waiting. You can start simple.
• 20:04
  Speaker 3: Learn as you go and build it from there. If
• 20:06
  Speaker 3: you're ready to hear the... of your first sale today,
• 20:09
  Speaker 3: head over to Shopify.com slash SRS to start your free
• 20:13
  Speaker 3: trial today. That's right. Start your free trial at Shopify.com
• 20:18
  Speaker 3: slash SRS. That's Shopify.com slash SRS. So where do we
• 20:27
  Speaker 3: go from here?
• 20:28
  Speaker 4: So, you know, I think like one thing, one of the,
• 20:33
  Speaker 4: you know, when you think about building a robotic system,
• 20:36
  Speaker 4: one of the ways that we kind of think about
• 20:40
  Speaker 4: this is that when you're perceiving the environment, you're not
• 20:43
  Speaker 4: just measuring it. You're taking the context and your prior
• 20:48
  Speaker 4: experience and you're using.
• 20:50
  Speaker 2: That to.
• 20:53
  Speaker 4: Predict what you think like the world actually is. So
• 20:56
  Speaker 4: for example, if you look out at a set of trees,
• 21:00
  Speaker 4: you're not just measuring that there's, you know, some obstacles
• 21:02
  Speaker 4: out there in front of you. You're also predicting that
• 21:04
  Speaker 4: there's free space behind them that you can potentially move through.
• 21:09
  Speaker 4: And you get to that point by essentially seeing lots
• 21:15
  Speaker 4: of trees like in your past, right?
• 21:16
  Speaker 2: Like you use that prior.
• 21:17
  Speaker 4: Experience and then you can understand when you see a
• 21:20
  Speaker 4: pattern like this, that actually means that there's, you know,
• 21:23
  Speaker 4: sort of space out there behind those trees that you
• 21:27
  Speaker 4: can then leverage in order to make decisions faster, to
• 21:31
  Speaker 4: move more aggressively. That's really, a machine learning way of
• 21:36
  Speaker 4: thinking about things. You're using lots of data, lots of
• 21:39
  Speaker 4: experience to understand what you're seeing in a functional way
• 21:43
  Speaker 4: and then move, you know, use that in order to
• 21:47
  Speaker 4: move a robot more quickly or more aggressively. And those notions,
• 21:53
  Speaker 4: you know, sort of led me from neurobiology at Duke,
• 21:59
  Speaker 4: like really kind of thinking about data and machine learning
• 22:03
  Speaker 4: as fundamentally, you know, interesting things towards my graduate education,
• 22:08
  Speaker 4: which I pursued after that at Carnegie Mellon University.
• 22:10
  Speaker 3: Wow. But I've not heard everybody talk about that, that
• 22:16
  Speaker 3: it's gotten into the machine learning AI stuff. That's pretty fascinating.
• 22:21
  Speaker 2: Yeah.
• 22:22
  Speaker 4: And I think, like, another thing which is pretty interesting
• 22:24
  Speaker 4: here that I'll also highlight, you know, I was originally
• 22:28
  Speaker 4: went to Duke to try to understand stuff the brain
• 22:32
  Speaker 4: and how it works. And my hope was that by
• 22:36
  Speaker 4: understanding that, that would help me to maybe better understand,
• 22:40
  Speaker 4: you know, artificial intelligence or how to build machines and
• 22:44
  Speaker 4: things like this. What I pretty quickly realized was that,
• 22:50
  Speaker 4: you know, neuroscience is really hard, right? People have been
• 22:53
  Speaker 4: studying the human brain for almost 300 years.
• 22:55
  Speaker 2: And progress is slow.
• 22:56
  Speaker 4: It's very difficult. difficult to understand how the brain works.
• 23:02
  Speaker 4: We don't have a great grasp of it even now.
• 23:06
  Speaker 4: We can describe a lot about it, but not really
• 23:08
  Speaker 4: understand it functionally.
• 23:10
  Speaker 2: And so.
• 23:13
  Speaker 4: One of the lessons from that was that, you know,
• 23:20
  Speaker 4: I almost came away with the opposite conclusion. So instead
• 23:23
  Speaker 4: of thinking about the brain as something that would help
• 23:26
  Speaker 4: me to build better machines or build a better AI,
• 23:31
  Speaker 4: I almost think that focusing on artificial intelligence and mathematics
• 23:37
  Speaker 4: and probability and statistics and information theory and robotics helps
• 23:43
  Speaker 4: to provide a framework that people may ultimately understand the
• 23:48
  Speaker 4: brain through. So it almost works the other way, that
• 23:52
  Speaker 4: you have to really understand kind of core principles of perception, planning, control,
• 24:00
  Speaker 4: like these areas which are fundamental in AI and robotics
• 24:03
  Speaker 4: to understand ultimately, you know, what the nervous system might
• 24:09
  Speaker 4: be doing and be able to describe it. Wow.
• 24:12
  Speaker 3: Wow. So you worked at, went to NVIDIA too as well,
• 24:19
  Speaker 3: didn't you? That's right.
• 24:20
  Speaker 6: Yeah.
• 24:21
  Speaker 4: So I, before starting Overland AI I, I worked for
• 24:26
  Speaker 4: about five years at NVIDIA. And this is while I
• 24:31
  Speaker 4: was a professor. So after Carnegie Mellon, I got my
• 24:35
  Speaker 4: PhD there in machine learning. I worked on robotics problems.
• 24:40
  Speaker 4: And then I was a professor at Georgia Tech for
• 24:44
  Speaker 4: five years, and then University of Washington for seven years
• 24:48
  Speaker 4: after that. And one of the things which is really
• 24:50
  Speaker 4: cool about being a a faculty member or a professor
• 24:55
  Speaker 4: who is running a research lab is that you can
• 24:58
  Speaker 4: also work in industry. So I had a research lab
• 25:01
  Speaker 4: which was focused on robotics and machine learning, and I
• 25:04
  Speaker 4: had a number of PhD students who are working in
• 25:06
  Speaker 4: that lab. But you can take 20% of your time
• 25:10
  Speaker 4: and work in industry simultaneously. And as part of that,
• 25:14
  Speaker 4: I spent about five years working at NVIDIA on machine
• 25:18
  Speaker 4: learning and robotics.
• 25:19
  Speaker 3: How was it working there?
• 25:20
  Speaker 2: That was awesome.
• 25:21
  Speaker 4: I mean, I think I joined around 2018. So it
• 25:25
  Speaker 4: was before NVIDIA was really kind of like a tier one,
• 25:30
  Speaker 4: let's say, tech company. I think Google and Microsoft were
• 25:35
  Speaker 4: really up there sort of defining state of the art.
• 25:40
  Speaker 4: But when I went to NVIDIA, I think there were
• 25:41
  Speaker 4: a lot of good people that they were hiring. And
• 25:44
  Speaker 4: people had recently seen the power of GPUs, right, like
• 25:48
  Speaker 4: this massive parallel processing. and I was thinking about that
• 25:52
  Speaker 4: in the context of robotics. How can you parallelize tasks?
• 25:56
  Speaker 4: How can you use it, not just these sorts of
• 25:59
  Speaker 4: chips and this type of technology, not just for perceiving
• 26:02
  Speaker 4: the environment, but also controlling vehicles? So an example of
• 26:09
  Speaker 4: this actually is carried through from work that I was
• 26:13
  Speaker 4: doing originally at Georgia Tech and then NVIDIA, now to Overland,
• 26:17
  Speaker 4: where When that vehicle is out in terrain, so when
• 26:21
  Speaker 4: our uncrewed vehicles are out there and looking at terrain,
• 26:24
  Speaker 4: they're evaluating tens of thousands of possible trajectories that they
• 26:28
  Speaker 4: might take, ranking each one of them, determining, like, is
• 26:33
  Speaker 4: this a good one or a bad one, and then
• 26:35
  Speaker 4: choosing how to drive after doing that. It's doing that
• 26:38
  Speaker 4: about 10 times a second. And so how do you
• 26:41
  Speaker 4: get that to work? Well, you can use NVIDIA GPUs
• 26:44
  Speaker 4: to parallelize these tasks and... evaluate many trajectories simultaneously, and
• 26:51
  Speaker 4: then decide how you're going to move based on that.
• 26:54
  Speaker 3: Very, very interesting. Do you still, do you miss being
• 26:58
  Speaker 3: a professor?
• 26:59
  Speaker 4: Well, yeah, I'm currently, I've got a 5% appointment at
• 27:05
  Speaker 4: the University of Washington, which means that I'm there, you know,
• 27:09
  Speaker 4: every couple of weeks working with students.
• 27:13
  Speaker 2: You know, I, I like teaching. I like interacting with students.
• 27:16
  Speaker 4: I think that's one of the great benefits of being
• 27:20
  Speaker 4: a professor is just engaging with students and people who
• 27:25
  Speaker 4: want to learn. So that part's fantastic. And I miss
• 27:29
  Speaker 4: doing that. I haven't been teaching recently since I've been
• 27:33
  Speaker 4: spun off the company. But I think working in industry
• 27:37
  Speaker 4: also allows you to really scale your ideas more. Um,
• 27:42
  Speaker 4: you know, there's only so much you can do in a,
• 27:44
  Speaker 4: in a smaller research lab. Yeah.
• 27:48
  Speaker 3: And so in 2018, the army research laboratory spotted you
• 27:54
  Speaker 3: at an IEEE conference, demoing machine learning. What is an IEEE?
• 28:00
  Speaker 2: It's IEEE.
• 28:02
  Speaker 4: So it's, it's a association for electrical engineers, but it's
• 28:05
  Speaker 4: one of the major types of con conferences that, um, uh,
• 28:11
  Speaker 4: that folks publish in. So when you're a professor, one
• 28:15
  Speaker 4: of your main goals is to publish papers, right? And
• 28:18
  Speaker 4: those scientific papers further human knowledge. And so in computer science,
• 28:24
  Speaker 4: the way that you do this is you publish papers
• 28:26
  Speaker 4: at a lot of conferences. You work with your graduate student,
• 28:32
  Speaker 4: you develop a new technology, you then tell the world
• 28:36
  Speaker 4: about it, right? You publish it in a paper and
• 28:39
  Speaker 4: you do this on a pretty um, pretty, you know,
• 28:43
  Speaker 4: fast iterative basis. So, uh, you know, that's, uh, this
• 28:47
  Speaker 4: is one of the, the major conferences in, in robotics.
• 28:50
  Speaker 4: And we had, you know, had some work there. Um,
• 28:53
  Speaker 4: and you know, some folks in the army were seeing
• 28:56
  Speaker 4: what we were doing and had some cool ideas of how,
• 28:59
  Speaker 4: of how we could potentially like take that fundamental research
• 29:02
  Speaker 4: and start to apply it to, uh, to army problems.
• 29:05
  Speaker 3: Right on, right on. And then the DARPA racer. Yeah.
• 29:11
  Speaker 3: the crucible that birthed Overland AI? What was that? What's the.
• 29:15
  Speaker 2: Yeah, yeah.
• 29:17
  Speaker 4: So, okay, so I was, you know, originally working at
• 29:23
  Speaker 4: Georgia Tech and doing some work on ground vehicle autonomy.
• 29:27
  Speaker 4: So the way that this started out, we took one
• 29:31
  Speaker 4: fifth scale vehicle, so these kind of smaller remote control vehicles,
• 29:35
  Speaker 4: we put computers and sensors on them and made them autonomous.
• 29:39
  Speaker 4: And we were trying to race them as fast as possible.
• 29:41
  Speaker 4: So we were using data that we were collecting while
• 29:47
  Speaker 4: we were driving these cars to learn what's called a policy.
• 29:52
  Speaker 4: You think about it as like AI, essentially, like for
• 29:56
  Speaker 4: the that could perceive the world and try to drive
• 30:00
  Speaker 4: really quickly. And these vehicles were doing things like drifting
• 30:03
  Speaker 4: around turns and things like this. So they learned to
• 30:06
  Speaker 4: do this, which is one of the things which is cool. Like,
• 30:08
  Speaker 4: the vehicle's out there, it's trying to drive faster and faster,
• 30:11
  Speaker 4: and then it's learning how to do things like drift
• 30:13
  Speaker 4: in order.
• 30:14
  Speaker 2: To drive even faster.
• 30:15
  Speaker 3: Oh, shit.
• 30:16
  Speaker 2: So that's what the army was looking at.
• 30:18
  Speaker 3: Now, is it actually learning, or are you programming that
• 30:21
  Speaker 3: into it?
• 30:21
  Speaker 4: So it's actually learning. So you start by bootstrapping it. Like,
• 30:25
  Speaker 4: you have a human demonstrate You know, this is how
• 30:28
  Speaker 4: the vehicle should drive. And then it tries to replicate
• 30:33
  Speaker 4: what the human does. And as it does that, sometimes
• 30:36
  Speaker 4: it makes mistakes, sometimes it does well, but it's kind
• 30:39
  Speaker 4: of grading itself.
• 30:40
  Speaker 2: And then it will start to experiment.
• 30:42
  Speaker 4: Like if I, you know, accelerate a little bit here
• 30:45
  Speaker 4: or I brake a little bit there, does this make me,
• 30:48
  Speaker 4: you know, faster or slower? And as it does that,
• 30:52
  Speaker 4: it learns how to drive faster and faster and learns
• 30:55
  Speaker 4: on its own. And so You're programming in the ability
• 31:00
  Speaker 4: to learn, but then it's looking at its own behavior,
• 31:05
  Speaker 4: learning from that and driving, you know, figuring out how
• 31:08
  Speaker 4: to drive faster.
• 31:09
  Speaker 3: Now, is this like on a track?
• 31:11
  Speaker 2: Yeah, exactly.
• 31:13
  Speaker 6: Yeah.
• 31:13
  Speaker 4: So we started out by driving on a sand track
• 31:18
  Speaker 4: and we were able to achieve these like really fast
• 31:21
  Speaker 4: lap times where the vehicle's you know, basically drifting through
• 31:24
  Speaker 4: turns and things like that. And so when I started
• 31:26
  Speaker 4: to work with the Army, the question was, can you
• 31:29
  Speaker 4: take these fundamental principles and apply them to larger vehicles
• 31:33
  Speaker 4: and figure out how to drive aggressively through all sorts.
• 31:36
  Speaker 2: Of different terrain?
• 31:37
  Speaker 4: So not just on tracks, but, you know, through forests
• 31:41
  Speaker 4: and deserts and beaches and things like this.
• 31:44
  Speaker 3: So, excuse me, what I was going to ask is,
• 31:47
  Speaker 3: I mean, if this, if, okay, if we have a
• 31:50
  Speaker 3: race car, it's going around a track over and over again,
• 31:52
  Speaker 3: and it is learning, and then you switch up the track,
• 31:57
  Speaker 3: will it be able to take what it learned on
• 32:01
  Speaker 3: track A and apply it to track B immediately? Do
• 32:07
  Speaker 3: you understand? Basically what I'm saying is how do we
• 32:09
  Speaker 3: go from a track to high-speed chase in the middle of,
• 32:15
  Speaker 3: I don't know, Manhattan?
• 32:17
  Speaker 4: So this is kind of like a fundamental challenge in
• 32:21
  Speaker 4: machine learning. If you learn, for example, how to drive
• 32:25
  Speaker 4: really fast on an oval track where you're always going
• 32:29
  Speaker 4: in one direction, you'll learn how to drive really fast
• 32:33
  Speaker 4: while always turning left, essentially, let's say. But then if
• 32:37
  Speaker 4: you go on another track which has right turns, will
• 32:41
  Speaker 4: it be able to generalize and be able to perform
• 32:45
  Speaker 4: well on a track like that? The short answer is,
• 32:50
  Speaker 4: you know, not without some sort of work. So generally speaking,
• 32:54
  Speaker 4: you want to collect data in a wide variety of
• 32:58
  Speaker 4: environments that are inclusive of the types of places that
• 33:02
  Speaker 4: you might want to be driving in the future. So,
• 33:04
  Speaker 4: for example, you might have a complex dirt track with
• 33:08
  Speaker 4: left and right turns and some wider turns and some
• 33:10
  Speaker 4: sharper turns and things like this. If you train on
• 33:13
  Speaker 4: a track like that, it's very easy to then race
• 33:17
  Speaker 4: on like an oval track. Because you've seen all of
• 33:20
  Speaker 4: the things that you're likely to see.
• 33:24
  Speaker 2: So that becomes easy.
• 33:25
  Speaker 4: But if you, for example, train on a dirt track
• 33:28
  Speaker 4: and now you have to drive on asphalt, you may
• 33:30
  Speaker 4: not be able to do that super well. And so
• 33:33
  Speaker 4: you need to collect new data as you move to
• 33:35
  Speaker 4: that new type of environment or new type of problem,
• 33:39
  Speaker 4: incorporate that into your learning algorithm, and then it will
• 33:45
  Speaker 4: start to do better on that new type of track.
• 33:48
  Speaker 4: And that matters even now when we think about where
• 33:52
  Speaker 4: we want to drive our vehicles because if you only train,
• 33:55
  Speaker 4: for example, in the desert, you're not going to be
• 33:57
  Speaker 4: able to necessarily drive well through a forest or vice versa.
• 34:01
  Speaker 4: And so you want to really train these systems, have
• 34:06
  Speaker 4: them collect data and learn from as wide a set
• 34:09
  Speaker 4: of environments as possible. So that they're able, when they
• 34:13
  Speaker 4: encounter a new environment, to still perform well, that there'll
• 34:16
  Speaker 4: still be aspects of that environment that they've seen before.
• 34:19
  Speaker 3: So I guess what my question is, will it get
• 34:22
  Speaker 3: to the point where it runs the route for the
• 34:29
  Speaker 3: very first time, like a brand new route for a
• 34:32
  Speaker 3: very first time? It could be off-road. It could be
• 34:35
  Speaker 3: in the middle of a city. But we're talking left turns,
• 34:38
  Speaker 3: right turns, heavy braking.
• 34:41
  Speaker 6: Yep.
• 34:42
  Speaker 3: Fast acceleration, drifting, all of that stuff, will it be able,
• 34:48
  Speaker 3: will it eventually learn what it needs to learn to
• 34:50
  Speaker 3: be able to have a complicated new route run it
• 34:56
  Speaker 3: the first time and it will run it perfectly? It
• 34:59
  Speaker 3: will go as fast as the machine is capable of.
• 35:03
  Speaker 4: Yeah, it's certainly possible. And like, that's what we're always
• 35:06
  Speaker 4: striving for. So When we're putting, you know, uncrewed ground vehicles,
• 35:10
  Speaker 4: autonomous vehicles in new environments, they're already performing really well
• 35:16
  Speaker 4: because it's seen many aspects of that before. And you're
• 35:19
  Speaker 4: trying to get it to perform, you know, optimally, right?
• 35:22
  Speaker 4: Like that is the goal, like potentially faster than a
• 35:26
  Speaker 4: human driver, even on environments that it's never seen before.
• 35:30
  Speaker 4: Every time our autonomous vehicles are out in the world,
• 35:35
  Speaker 4: they are... Basically, we've trained them so that we don't
• 35:42
  Speaker 4: pre-map environments or anything like that. It's as if they're
• 35:44
  Speaker 4: encountering that environment for the first time. And they're constantly learning.
• 35:49
  Speaker 4: So the more environments that we encounter, the more data
• 35:51
  Speaker 4: we get, the better and better the system gets. Then
• 35:55
  Speaker 4: the goal is always kind of moving towards that optimal movement.
• 36:00
  Speaker 3: How close are you to that goal?
• 36:02
  Speaker 2: Uh, depends, depends on the environment.
• 36:04
  Speaker 4: I mean, I think we can, we can already drive
• 36:06
  Speaker 4: faster than, than humans in some environments.
• 36:09
  Speaker 2: So, um, yeah, it's, it's pretty cool.
• 36:13
  Speaker 4: Wow.
• 36:14
  Speaker 3: Very interesting. And so how long did you spend at DARPA?
• 36:19
  Speaker 2: Oh, so, yeah.
• 36:20
  Speaker 4: So I started work with, um, army research lab, you know, again,
• 36:24
  Speaker 4: around like 2018. Um, and in 2020, 19 or so,
• 36:32
  Speaker 4: I started to talk with folks at DARPA. They were
• 36:35
  Speaker 4: really interested in developing, essentially rebooting ground autonomy for defense.
• 36:43
  Speaker 4: So here's what the situation was. DARPA, back in 2004
• 36:47
  Speaker 4: and 2005, had something called the Grand Challenge or Grand Challenges.
• 36:52
  Speaker 4: So these were challenges where they were trying to incept
• 36:56
  Speaker 4: ground autonomy. So this actually goes back to the NDAA
• 37:01
  Speaker 4: in 2001. This is basically where Congress decided in 2001,
• 37:06
  Speaker 4: they were like, by 2015, we want a third of
• 37:09
  Speaker 4: all military ground vehicles to be autonomous. So thinking about
• 37:13
  Speaker 4: this over 25 years ago, the problem was no one
• 37:17
  Speaker 4: knew how to do that. It wasn't like there were
• 37:20
  Speaker 4: autonomous vehicles driving all over the place and this.
• 37:24
  Speaker 2: Would be easy. No one knew how to make these
• 37:26
  Speaker 2: vehicles autonomous.
• 37:27
  Speaker 3: Where did the idea come from?
• 37:29
  Speaker 4: Well, I think, you know, there had been, you know,
• 37:32
  Speaker 4: work in robotics where people were moving vehicles in simpler
• 37:37
  Speaker 4: environments autonomously. And so the notion was like, well, what
• 37:42
  Speaker 4: if we could do this on the battlefield, right? What
• 37:44
  Speaker 4: if we could do it in these more complex environments?
• 37:47
  Speaker 4: If you could take the warfighter out of the vehicle,
• 37:50
  Speaker 4: you can imagine, you know, that not just provide safety,
• 37:54
  Speaker 4: but potentially, you know, tactical things that you can do.
• 37:57
  Speaker 2: We can maybe talk about that in a little bit.
• 37:59
  Speaker 4: But in order to make those vehicles autonomous, you needed
• 38:04
  Speaker 4: to know how to do it, right?
• 38:05
  Speaker 2: And they didn't.
• 38:06
  Speaker 4: And so DARPA, one of the things that's great about
• 38:08
  Speaker 4: DARPA is that it is an organization that is designed
• 38:13
  Speaker 4: to just like create new things, right? Like you have
• 38:17
  Speaker 4: some crazy challenge, DARPA is out there and can create
• 38:22
  Speaker 4: a program, pull together some of the best minds in
• 38:25
  Speaker 4: the US to focus like really focus on it for
• 38:28
  Speaker 4: up to about four years and try to solve a problem.
• 38:33
  Speaker 4: And so in 2004 and 2005, they came up with
• 38:37
  Speaker 4: the DARPA Grand Challenge where they were trying to race
• 38:40
  Speaker 4: vehicles from Barstow, California to Prim, Nevada. It was about
• 38:47
  Speaker 4: 135 miles on dirt roads. And they just laid down
• 38:53
  Speaker 4: a challenge. They said, like, if you can do this,
• 38:54
  Speaker 4: you get prize money. And all sorts of teams came
• 38:57
  Speaker 4: together to attack this problem. So there were university teams
• 39:02
  Speaker 4: from places like Carnegie Mellon and Stanford and MIT and
• 39:06
  Speaker 4: so on.
• 39:07
  Speaker 2: And there were.
• 39:09
  Speaker 4: Industry teams like Oshkosh had a team. And there were
• 39:12
  Speaker 4: just people who were trying to put together autonomous vehicles
• 39:14
  Speaker 4: in their garage, like just build robots. And they went
• 39:17
  Speaker 4: out there and raced. And in 2004, the first challenge,
• 39:22
  Speaker 4: no one made it beyond seven miles. Like that was
• 39:24
  Speaker 4: the you know, so it was, CMU's vehicle, I think, made.
• 39:29
  Speaker 2: It that far. It got stuck, caught fire. It was
• 39:32
  Speaker 2: like a whole thing.
• 39:34
  Speaker 4: The next year, though, in 2005, I think five teams
• 39:38
  Speaker 4: competed this challenge. They made it the entire 134 miles.
• 39:42
  Speaker 4: Those teams, like the folks from those teams, After that
• 39:45
  Speaker 4: challenge was over, there were a few other ones. There
• 39:48
  Speaker 4: was something called the Urban Challenge and some other DARPA
• 39:51
  Speaker 4: programs which followed.
• 39:53
  Speaker 3: Up on this.
• 39:53
  Speaker 4: But many of those people then moved into industry and
• 39:59
  Speaker 4: started the self-driving car projects and companies that then turned
• 40:04
  Speaker 4: into companies like Waymo or Aurora Innovation and so on.
• 40:11
  Speaker 4: So these autonomous driving company, commercial companies. So by 2012,
• 40:18
  Speaker 4: let's say, a lot of work was being done in
• 40:20
  Speaker 4: the commercial sector, bootstrapped off of this DARPA work, right?
• 40:24
  Speaker 4: So the military started this whole thing because they wanted
• 40:27
  Speaker 4: autonomous vehicles. People started to build autonomous vehicles because of
• 40:31
  Speaker 4: this DARPA program, but then basically just went off into
• 40:37
  Speaker 4: industry and were working on like robo-taxis and autonomous trucks
• 40:40
  Speaker 4: and things like this. So by 2019, coming back to
• 40:45
  Speaker 4: my story, DARPA was in a position where they were like, well, cool,
• 40:51
  Speaker 4: we have autonomous taxis. There's been a lot of progress
• 40:54
  Speaker 4: in this area, but where's our autonomous tanks? Where are
• 40:59
  Speaker 4: our autonomous military vehicles? The whole point of this was
• 41:02
  Speaker 4: initially to support the military. And so DARPA Racer was
• 41:08
  Speaker 4: a program that got stood up. It started in 2021.
• 41:12
  Speaker 4: to reboot autonomy for defense. So specifically to take a
• 41:16
  Speaker 4: lot of the learnings that had been produced over the
• 41:18
  Speaker 4: previous 20 years or so for the on-road autonomous driving
• 41:23
  Speaker 4: industry and work which had been done in robotic perception
• 41:28
  Speaker 4: and robotic vehicle control, bring that back together and focus
• 41:34
  Speaker 4: on defense problems. And so that meant trying to drive
• 41:39
  Speaker 4: much faster you know, larger and faster vehicles off-road, what
• 41:43
  Speaker 4: we call complex natural terrain.
• 41:46
  Speaker 2: So, you know, no roads at all, right?
• 41:49
  Speaker 4: Like through deserts, through forests, through snow, things like this,
• 41:54
  Speaker 4: and contested terrain. So thinking about, you know, how do
• 41:57
  Speaker 4: you move when you, not only do you not have infrastructure,
• 42:01
  Speaker 4: which is there to help you like roads or road
• 42:04
  Speaker 4: networks or signs or things like this, but infrastructure, which
• 42:08
  Speaker 4: might be in the environment, which is there to defeat you,
• 42:11
  Speaker 4: to stop you. And so that's what the DARPA Racer
• 42:15
  Speaker 4: program was. So I had already been doing work at
• 42:20
  Speaker 4: Georgia Tech and then University of Washington, working with the
• 42:24
  Speaker 4: Army on developing off-road ground vehicle autonomy. And I then
• 42:29
  Speaker 4: put together a team, it's called a performer team, to
• 42:34
  Speaker 4: attack these problems for the military through DARPA starting in 2021.
• 42:40
  Speaker 3: So did they recruit you from the IEEE conference?
• 42:44
  Speaker 4: So the way that that worked was, like, through the
• 42:46
  Speaker 4: research I was doing and the publications I was, you know,
• 42:49
  Speaker 4: putting out to the world, Army saw that the.
• 42:52
  Speaker 2: Technology that we were.
• 42:54
  Speaker 4: Army Research Lab saw that the technology we were developing
• 42:57
  Speaker 4: might be really helpful for the types of problems they
• 42:59
  Speaker 4: wanted to solve. I then started working with army. So
• 43:03
  Speaker 4: the way that that works is when you're running a
• 43:04
  Speaker 4: university research lab, um, you have a bunch of PhD students,
• 43:08
  Speaker 4: you know, they're doing research, you're publishing it, you're making
• 43:11
  Speaker 4: it publicly available for other scientists to see.
• 43:13
  Speaker 3: Okay.
• 43:14
  Speaker 4: But you need funding to, to run that lab. And
• 43:17
  Speaker 4: so people pay you to essentially do research. They pay
• 43:20
  Speaker 4: your lab to do research. So army, um, us army
• 43:24
  Speaker 4: was one of the organizations that fund like started to
• 43:27
  Speaker 4: fund my research. When you fund research, you can say like, hey,
• 43:31
  Speaker 4: here are the problems we want you to solve. We'll
• 43:33
  Speaker 4: give you, you know, this much money to solve them
• 43:35
  Speaker 4: with your PhD students. And then you provide those solutions
• 43:39
  Speaker 4: back to the Army. So that's what I was doing
• 43:43
  Speaker 4: when I was working at Georgia Tech and University of
• 43:45
  Speaker 4: Washington was my lab was partially funded by the U.S.
• 43:50
  Speaker 6: Army.
• 43:51
  Speaker 4: Then I worked on problems that were interesting to them.
• 43:54
  Speaker 4: We provided those solutions back to the Army. And then
• 43:57
  Speaker 4: we started to work with DARPA, which is at the
• 44:00
  Speaker 4: time Department of Defense, but Department of War-level organization. They
• 44:06
  Speaker 4: saw the work we were doing with the U.S. Army,
• 44:09
  Speaker 4: and then they decided to fund my lab at a
• 44:11
  Speaker 4: much higher level to attack these problems of how do
• 44:16
  Speaker 4: you drive vehicles off-road and all sorts of different types
• 44:20
  Speaker 4: of terrain at high speeds.
• 44:22
  Speaker 2: Wow.
• 44:23
  Speaker 3: So you've been– You've really been at the cutting edge
• 44:25
  Speaker 3: of this thing the entire time. Yeah.
• 44:28
  Speaker 4: So we've been working on these problems for more than
• 44:31
  Speaker 4: 10 years, more than a decade. And my lab was
• 44:35
  Speaker 4: doing a lot of that work in academia before we
• 44:38
  Speaker 4: spun out Overland.
• 44:40
  Speaker 3: Wow. Wow. Well, before we get into Overland, let's take
• 44:44
  Speaker 3: a quick break. There's been a lot of attention lately
• 44:49
  Speaker 3: on how important sleep is, not just for recovery, but
• 44:53
  Speaker 3: for your brain, hormones, immune system, and for your performance.
• 44:57
  Speaker 3: And the more research comes out, the clearer it gets.
• 45:01
  Speaker 3: If your sleep is off, everything else gets harder. That's
• 45:04
  Speaker 3: why I use Helix. I've had my Helix mattress for
• 45:08
  Speaker 3: a while now, and it's been a real upgrade from
• 45:10
  Speaker 3: what I was sleeping on before. Helix has over 20
• 45:13
  Speaker 3: mattress models, so you can find one that actually fits
• 45:17
  Speaker 3: how you sleep instead of guessing your way through it.
• 45:19
  Speaker 3: And if you sleep hot, they've got cooling upgrades that
• 45:22
  Speaker 3: help you stay comfortable through the night, especially during the summer.
• 45:27
  Speaker 3: For me, Helix feels high quality. durable, and comfortable. It's
• 45:31
  Speaker 3: not just a mattress that shows up in a box.
• 45:33
  Speaker 3: It feels like something that's built to last. Helix ships
• 45:37
  Speaker 3: straight to your door for free in the US, and
• 45:40
  Speaker 3: you get 120 nights sleep trial, a limited lifetime warranty,
• 45:44
  Speaker 3: and they're happy with Helix guarantee. So returns and exchanges
• 45:48
  Speaker 3: are simple if it's not the right fit. Helix is
• 45:51
  Speaker 3: also the most awarded mattress brand. Tested and reviewed by
• 45:55
  Speaker 3: experts like Forbes and Wired, Go to helixsleep.com slash SRS
• 46:01
  Speaker 3: for up to 30% off. That's helixsleep.com slash SRS for
• 46:06
  Speaker 3: up to 30% off. Make sure you enter our show
• 46:08
  Speaker 3: name after checkout so they know we sent you. helixsleep.com
• 46:13
  Speaker 3: slash SRS. Do you ever wonder what it takes to
• 46:18
  Speaker 3: make an episode of The Sean Ryan Show? In this
• 46:20
  Speaker 3: exclusive studio tour, I'm taking you behind the scenes for
• 46:24
  Speaker 3: an in-depth look at every part of the operation, From
• 46:28
  Speaker 3: the editing room in the main studio to the spaces
• 46:30
  Speaker 3: where we film range day content and more, you'll see
• 46:34
  Speaker 3: how the show comes together, meet some of the people
• 46:36
  Speaker 3: behind it, and get a closer look at the work
• 46:39
  Speaker 3: that happens off-camera. When you become a paid member of
• 46:42
  Speaker 3: the SRS Patreon community, you get more than just the podcast.
• 46:46
  Speaker 3: Watch new episodes early alongside other members, join monthly live
• 46:50
  Speaker 3: shows with guest Q & As, and submit questions, just
• 46:54
  Speaker 3: like you see on the show, for upcoming guests online.
• 46:57
  Speaker 3: on the Protector tier. You'll also unlock exclusive Range Day videos,
• 47:01
  Speaker 3: behind-the-scenes content, and premium ambience videos that you're not going
• 47:05
  Speaker 3: to find anywhere else. Join the Patreon community today and
• 47:14
  Speaker 3: get access to the full experience. All right, let's talk
• 47:20
  Speaker 3: about the Textron M5.
• 47:24
  Speaker 2: Sure, yeah, so.
• 47:26
  Speaker 4: As part of the DARPA Racer program, so this is
• 47:29
  Speaker 4: DARPA program where we're trying to develop ground vehicle autonomy
• 47:35
  Speaker 4: for the military, we started out working with Polaris Racer side-by-sides.
• 47:40
  Speaker 2: So these are very capable vehicles. They can go very fast.
• 47:44
  Speaker 4: They've got four wheels, about 3,000 pounds, about the size
• 47:47
  Speaker 4: of a small SUV, something like this.
• 47:51
  Speaker 2: And you know, in the.
• 47:54
  Speaker 4: First part of the DARPA racer program, uh, we were just,
• 47:59
  Speaker 4: there were multiple teams and we were essentially like racing
• 48:03
  Speaker 4: these Polaris racers. So trying to record the best possible
• 48:07
  Speaker 4: times through a wide variety of different terrains and, um,
• 48:11
  Speaker 4: you know, test scenarios, uh, as we progressed through the program,
• 48:17
  Speaker 4: you know, the, the, DARPA program manager didn't want us
• 48:21
  Speaker 4: to just be using those small vehicles. So we moved
• 48:25
  Speaker 4: up to the Textron M5. The Textron M5 is basically
• 48:29
  Speaker 4: a robotic combat vehicle prototype. It's about 12 tons. So
• 48:37
  Speaker 4: it is much larger.
• 48:38
  Speaker 3: You went from a side-by-side to a 12-ton vehicle?
• 48:42
  Speaker 2: Yeah, it's a 12-ton tracked vehicle.
• 48:44
  Speaker 3: Wow.
• 48:45
  Speaker 4: So it's about the size of a 113. So it
• 48:48
  Speaker 4: looks like a tank. It's sort of like a small tank,
• 48:52
  Speaker 4: I guess, but pretty heavy vehicle, tracked. It's electric, so
• 48:58
  Speaker 4: really aggressive. The ability to essentially go from, you know,
• 49:03
  Speaker 4: zero to 60, like super, think about like a Tesla,
• 49:06
  Speaker 4: but like, you know, like a tank version of that.
• 49:09
  Speaker 3: Wow.
• 49:10
  Speaker 4: So really, really awesome vehicle to work with. And we're
• 49:13
  Speaker 4: putting autonomy on that vehicle and then driving that through
• 49:16
  Speaker 4: off-road terrain.
• 49:17
  Speaker 3: Okay. And this was a competition, correct?
• 49:21
  Speaker 2: Yeah. So DARPA Racer, it started out with three teams.
• 49:26
  Speaker 4: There was our team and there were two other teams
• 49:27
  Speaker 4: that we were competing against. By the first summer, about
• 49:32
  Speaker 4: a year into the program, we had by far the
• 49:37
  Speaker 4: best technology available. Um, and pretty quickly, uh, the other
• 49:41
  Speaker 4: teams essentially dropped out and we were the only ones left.
• 49:45
  Speaker 4: We were focused on not only driving on the, on
• 49:48
  Speaker 4: the side-by-sides, but also the, the much larger vehicle. So
• 49:50
  Speaker 4: we're the only team that actually worked with, with these.
• 49:53
  Speaker 4: And there aren't that many of them. There was something like,
• 49:55
  Speaker 4: you know, four or five Textron M5s ever produced. Uh,
• 49:59
  Speaker 4: but you know, we had all of them and we
• 50:00
  Speaker 4: were just like smashing through terrain and in these, these
• 50:03
  Speaker 4: huge tracked vehicles.
• 50:04
  Speaker 2: Pretty awesome.
• 50:06
  Speaker 3: And so what is, what is the. What is kind
• 50:10
  Speaker 3: of the point of the competition? Do they buy the
• 50:12
  Speaker 3: technology from you, or do they.
• 50:15
  Speaker 2: Fund the technology? That's a great question.
• 50:18
  Speaker 3: How does it work?
• 50:19
  Speaker 4: Yeah, like, this is a really good question, actually, because
• 50:25
  Speaker 4: the way that DARPA normally works, you know, they have
• 50:27
  Speaker 4: these programs where they're trying to develop new technology. In
• 50:30
  Speaker 4: this case, develop ground vehicle autonomy.
• 50:34
  Speaker 2: For the U.S.
• 50:35
  Speaker 4: Military, so for our soldiers and Marines, right? Um, but
• 50:40
  Speaker 4: often in a DARPA program, even if the program is
• 50:44
  Speaker 4: very successful, like ours was, how do you actually take
• 50:48
  Speaker 4: that technology and then get it into the hands of
• 50:51
  Speaker 4: the warfighter?
• 50:51
  Speaker 2: Like, how do you actually do that? Um, it's a good.
• 50:55
  Speaker 4: Question and very few people succeed. So the, this is
• 51:01
  Speaker 4: known as like transition, transitioning, like a new capability that's
• 51:04
  Speaker 4: developed from DARPA. to the actual warfighter is something like
• 51:08
  Speaker 4: 4% of technologies ever actually get transitioned. So just because
• 51:14
  Speaker 4: you succeed in developing a new capability doesn't mean that
• 51:17
  Speaker 4: the warfighter gets that capability. This is actually why we
• 51:22
  Speaker 4: created Overland AI. We were like in this program and
• 51:28
  Speaker 4: we had this new capability and we were just like,
• 51:31
  Speaker 4: this is incredible. drive extremely fast off road. We can
• 51:35
  Speaker 4: do it in a variety of different types of vehicles.
• 51:38
  Speaker 4: We have the thing that DARPA wanted us to produce.
• 51:44
  Speaker 4: We actually did it. Now, how can the warfighter benefit?
• 51:48
  Speaker 4: And so we felt that the most promising way of
• 51:56
  Speaker 4: making that transition was to spin off a company that
• 52:00
  Speaker 4: would then take the capability that we developed as part
• 52:03
  Speaker 4: of this DARPA program, commercialize it. So turn it into
• 52:07
  Speaker 4: a robust, reliable, commercial autonomy stack, and then sell that
• 52:13
  Speaker 4: back to the Army and Marine Corps so that they
• 52:15
  Speaker 4: could use it on their programs, so on their vehicles,
• 52:22
  Speaker 4: and so that warfighting units could start to work with
• 52:25
  Speaker 4: this technology and start to integrate.
• 52:28
  Speaker 2: It into tactics. So that's why we created Overland AI.
• 52:32
  Speaker 4: We, like we, we thought the army, like this is
• 52:35
  Speaker 4: just going to disappear if we don't do this.
• 52:37
  Speaker 3: So why would they fund the, did they fund the research?
• 52:41
  Speaker 2: Yeah.
• 52:42
  Speaker 4: So DARPA funded all of this, like, like they did
• 52:45
  Speaker 4: to develop the capability, but just because DARPA.
• 52:47
  Speaker 3: And they just shelf it?
• 52:49
  Speaker 4: So, so then what has to happen is like the
• 52:51
  Speaker 4: army and the Marine Corps or whoever, like one of
• 52:55
  Speaker 4: the services has to then put up money to be
• 52:59
  Speaker 4: able to then like buy and transition that technology because.
• 53:04
  Speaker 2: From DARPA or from you?
• 53:07
  Speaker 4: Yeah, so the way that DARPA works is that DARPA,
• 53:09
  Speaker 4: like these DARPA programs are generally four years long.
• 53:13
  Speaker 2: They have a time limit. And that's part of the point.
• 53:17
  Speaker 4: Like it gives you a very focused window in time
• 53:20
  Speaker 4: to just do whatever it takes to try to develop
• 53:23
  Speaker 4: this new technology. But at the end of those four years,
• 53:26
  Speaker 4: that program is over and you have to find- someone
• 53:30
  Speaker 4: who is then going to support the continuation of that
• 53:34
  Speaker 4: technology to move it into warfighting units. So you've created it.
• 53:44
  Speaker 4: Now someone has to continue to transfer it, to get
• 53:48
  Speaker 4: it onto military vehicles, to push it into.
• 53:53
  Speaker 6: Units.
• 53:54
  Speaker 4: And that's where a lot of these programs die because
• 53:56
  Speaker 4: you create the technology, but you can't.
• 54:00
  Speaker 2: That no one picks it up like when it's over.
• 54:02
  Speaker 3: I would think there would be. So, so this is,
• 54:06
  Speaker 3: I mean, what would you call this? Like an incubator?
• 54:08
  Speaker 3: Like you get, you get, you get picked to, to
• 54:11
  Speaker 3: get funding, to be an incubator. And then DARPA doesn't even,
• 54:14
  Speaker 3: they don't own the actual.
• 54:16
  Speaker 2: Well, they do. They, they, they, they own the IP.
• 54:18
  Speaker 2: So also on the IP. Yeah.
• 54:20
  Speaker 4: So the way that this works is like when, when,
• 54:24
  Speaker 4: so DARPA pays for the development of, you know, this
• 54:27
  Speaker 4: core technology and, In this case, I'm running a lab
• 54:33
  Speaker 4: at the University of Washington. So they pay the University
• 54:36
  Speaker 4: of Washington to develop the IP. When they do that,
• 54:42
  Speaker 4: that IP belongs to both DARPA and the University of Washington, right?
• 54:46
  Speaker 4: Because DARPA pays for a university to develop it. Both
• 54:53
  Speaker 4: parties essentially own that IP. But once you have that IP,
• 54:57
  Speaker 4: in this case, it is like, It is raw technology.
• 55:01
  Speaker 4: It is research-grade code that can demonstrate this capability, that
• 55:08
  Speaker 4: will allow a vehicle to drive fast, but it's not
• 55:11
  Speaker 4: going to be reliable in the same way that a
• 55:17
  Speaker 4: Waymo is, where you have a whole team of engineers
• 55:21
  Speaker 4: that is making production-quality code.
• 55:25
  Speaker 2: They own that, like they own the.
• 55:28
  Speaker 4: Basic capability and they own that IP. But that then
• 55:34
  Speaker 4: has to be taken and turned into essentially a commercial
• 55:40
  Speaker 4: production ready piece of software.
• 55:44
  Speaker 2: It's almost like a rough draft.
• 55:45
  Speaker 3: I would think there would be.
• 55:47
  Speaker 4: So, but then someone has to kind of pay for
• 55:49
  Speaker 4: that transition. And that's often, again, where these programs end
• 55:54
  Speaker 4: up having problems where Like you develop a new technology,
• 55:58
  Speaker 4: but then who adopts it? Who actually transitions it to
• 56:03
  Speaker 4: the warfighter?
• 56:04
  Speaker 3: Are these classified? I mean, are these classified? I would
• 56:07
  Speaker 3: think there would be venture capital firms galore just surrounding
• 56:11
  Speaker 3: these projects.
• 56:12
  Speaker 5: Well, yeah, yeah.
• 56:13
  Speaker 2: So sometimes, yes, right?
• 56:15
  Speaker 1: Only 4% make it to the warfighter?
• 56:17
  Speaker 4: Right.
• 56:18
  Speaker 3: 96% of projects. It's crazy, right?
• 56:20
  Speaker 4: It's like, oh, yeah. But in some sense, that's actually
• 56:24
  Speaker 4: what happened here, right? Because we said, okay, we're going
• 56:28
  Speaker 4: to take this IP. We license it, right? So that
• 56:32
  Speaker 4: IP is licensed to Overland AI. We created this company
• 56:36
  Speaker 4: to essentially take that IP and then turn it into
• 56:39
  Speaker 4: a commercial autonomy stack that is reliable.
• 56:45
  Speaker 2: And we keep building on it. We keep improving it.
• 56:48
  Speaker 4: And then we can transition that back to the warfighters.
• 56:52
  Speaker 4: So Overland AI got its start by focusing on the
• 56:56
  Speaker 4: software portion of this, right? Take these good ideas that
• 56:59
  Speaker 4: were developed during DARPA. We licensed that IP at Overland.
• 57:05
  Speaker 4: We build on top of it. We turned it into
• 57:07
  Speaker 4: a commercial autonomy stack. And now we have this software
• 57:10
  Speaker 4: that you can put onto a vehicle to make that
• 57:13
  Speaker 4: vehicle autonomous. And we had worked with you know, different
• 57:17
  Speaker 4: types of vehicles. We discussed like the side-by-side or like,
• 57:20
  Speaker 4: you know, the big Textron M5, um, that we can
• 57:22
  Speaker 4: make autonomous. Um, you know, we really wanted to develop software,
• 57:26
  Speaker 4: which would work on any vehicle, uh, that the military had.
• 57:31
  Speaker 4: And so Overland, um, was stood up to do that.
• 57:36
  Speaker 4: And to your point about VCs, you know, we're a
• 57:38
  Speaker 4: VC backed company, right? So they're essentially making that bet
• 57:41
  Speaker 4: that you're saying it's like, okay, we can step in,
• 57:44
  Speaker 4: provide additional funding to turn this into a mature product
• 57:48
  Speaker 4: that can then be transitioned over to the warfighter. But
• 57:53
  Speaker 4: now you're in like defense tech territory, right? Where you
• 57:56
  Speaker 4: now have to fight all of these battles to go
• 57:59
  Speaker 4: from this good core idea or good core IP to
• 58:05
  Speaker 4: actually getting it procured, right?
• 58:09
  Speaker 2: And that takes.
• 58:12
  Speaker 2: You know, time and, and, and, and effort.
• 58:14
  Speaker 4: And, and so like our story as, as a defense
• 58:17
  Speaker 4: tech company is, is really similar to a lot of
• 58:20
  Speaker 4: other stories, except that we got this start by, you know,
• 58:24
  Speaker 4: doing all of this, you know, sort of initial research
• 58:26
  Speaker 4: to accept a new technology that the military really wanted.
• 58:31
  Speaker 3: Very interesting. And so, so the DARPA competition ends, you
• 58:36
  Speaker 3: win and you, all the information with you and develop
• 58:41
  Speaker 3: Overland AI. Right.
• 58:42
  Speaker 4: So we started Overland AI. And Overland actually was formed
• 58:51
  Speaker 4: before the DARPA competition completed. So we became part of
• 58:56
  Speaker 4: the competition as well. We then As the competition was
• 59:02
  Speaker 4: going on, we were developing new technology for moving really
• 59:07
  Speaker 4: quickly through a lot of different biomes, so different types
• 59:09
  Speaker 4: of environments. Um, and then we were really trying to think,
• 59:13
  Speaker 4: and this is the Overland AI started in, in 20,
• 59:15
  Speaker 4: like December, 2022. Um, so three and a half years ago,
• 59:21
  Speaker 4: and we were initially focused on, okay, we have this software,
• 59:25
  Speaker 4: we're refining it, we're turning it into a commercial piece
• 59:28
  Speaker 4: of software. How do you then get traction with the
• 59:34
  Speaker 4: Army and Marine Corps? Not just DARPA, but the people
• 59:38
  Speaker 4: who ultimately need to use this technology. And the problem
• 59:43
  Speaker 4: was that even though we had this great piece of software,
• 59:47
  Speaker 4: it has to go on a vehicle. So it was like,
• 59:49
  Speaker 4: what vehicle are we going to put it on? And
• 59:51
  Speaker 4: we started to talk to companies and units that had
• 59:55
  Speaker 4: vehicles which could be made autonomous, but there just weren't
• 59:57
  Speaker 4: that many of them out there. We had something which
• 1:00:00
  Speaker 4: could be very effective, but we didn't have vehicles to
• 1:00:03
  Speaker 4: put it on. So we started out, you know, again
• 1:00:07
  Speaker 4: with a software stack, but we quickly realized that we
• 1:00:09
  Speaker 4: had to actually also build the hardware so that we
• 1:00:13
  Speaker 4: could get this idea of autonomous vehicles into the hands.
• 1:00:17
  Speaker 2: Of warfighters faster. So we just started to build the
• 1:00:20
  Speaker 2: vehicles too.
• 1:00:23
  Speaker 4: And that's what resulted in this ultra vehicle that you've
• 1:00:29
  Speaker 4: seen outside where this is an autonomous vehicle that we
• 1:00:36
  Speaker 4: built and we put our software on so that we
• 1:00:39
  Speaker 4: could start to put many different types of payloads on
• 1:00:43
  Speaker 4: the vehicles and start to work with warfighting units to
• 1:00:47
  Speaker 4: really integrate autonomous systems into their concepts of operations. So our,
• 1:00:54
  Speaker 4: our path really is initially research and development, starting with
• 1:00:58
  Speaker 4: DARPA that then we, we worked with the defense innovation
• 1:01:02
  Speaker 4: unit and got a prototype contract with them. We then
• 1:01:07
  Speaker 4: started to work with army applications lab and, you know, winning,
• 1:01:12
  Speaker 4: you know, these Sibbers, these small contracts with the Marine
• 1:01:14
  Speaker 4: Corps and the army, and we were doing really well
• 1:01:16
  Speaker 4: with those.
• 1:01:19
  Speaker 2: And, you know, we, Then.
• 1:01:21
  Speaker 4: Last year, we revealed like the ultra vehicle, fully autonomous,
• 1:01:25
  Speaker 4: you know, vertically integrated vehicle that you could push into hands.
• 1:01:28
  Speaker 2: Of war fighters.
• 1:01:30
  Speaker 4: We started to push this technology into units who were
• 1:01:34
  Speaker 4: constantly testing them. And then this led to a production contract.
• 1:01:39
  Speaker 4: So we just recently won the first production contract for
• 1:01:43
  Speaker 4: autonomous ground vehicles in the US military.
• 1:01:46
  Speaker 2: And that's with the Marine Corps.
• 1:01:48
  Speaker 4: So it's really this whole process of- And it didn't
• 1:01:52
  Speaker 4: take that long, right? It's about three, three and a
• 1:01:54
  Speaker 4: half years where we went from like university lab R
• 1:01:58
  Speaker 4: and D through prototype fielding with war fighters production contract. Wow.
• 1:02:03
  Speaker 3: So we felt three and a half years.
• 1:02:04
  Speaker 6: Yeah.
• 1:02:05
  Speaker 2: Yeah.
• 1:02:07
  Speaker 4: Wow.
• 1:02:08
  Speaker 3: I mean, so who won the contract or who, who, who,
• 1:02:12
  Speaker 3: who are you contracted to?
• 1:02:13
  Speaker 4: Yeah.
• 1:02:13
  Speaker 3: Yeah.
• 1:02:14
  Speaker 2: With, with, with, with the Marine Corps.
• 1:02:16
  Speaker 4: So the Marine Corps is, Buying a whole set of
• 1:02:21
  Speaker 4: the autonomous vehicles that we produce, it's part of their
• 1:02:24
  Speaker 4: ground-based air defense program. And so they'll initially start using
• 1:02:30
  Speaker 4: those vehicles for autonomous resupply of basically.
• 1:02:37
  Speaker 2: Air defense systems.
• 1:02:38
  Speaker 4: So trying to make sure that they get enough ammunition
• 1:02:42
  Speaker 4: as they're shooting down drones.
• 1:02:43
  Speaker 3: Can I ask how many you're going to manufacture for them?
• 1:02:50
  Speaker 4: Yeah, so we're manufacturing in the first tranche about 15
• 1:02:55
  Speaker 4: of these in the next year or so. So that's
• 1:02:59
  Speaker 4: what we're starting out with. And one of the things
• 1:03:03
  Speaker 4: which is pretty interesting about these vehicles, they're modular. So
• 1:03:08
  Speaker 4: a lot of units want to start using them for
• 1:03:10
  Speaker 4: resupply where they're just kind of putting supplies on them
• 1:03:12
  Speaker 4: and moving those supplies back and forth. But we've been
• 1:03:15
  Speaker 4: doing a lot of work with other units like 82nd Airborne,
• 1:03:20
  Speaker 4: 173rd Airborne, other warfighting units to put other types of
• 1:03:24
  Speaker 4: payloads on board the platforms as well. And so one
• 1:03:28
  Speaker 4: of the things that we've been talking to Marines about
• 1:03:30
  Speaker 4: is putting sensors and effectors, so basically kinetic counter UAS
• 1:03:38
  Speaker 4: payloads on the vehicles and thinking about how to disaggregate them.
• 1:03:45
  Speaker 4: So one of the ways that you can use autonomous
• 1:03:49
  Speaker 4: vehicles is by essentially putting your sensors and putting your
• 1:03:56
  Speaker 4: counter UAS, for example, payloads on vehicles and moving them
• 1:04:00
  Speaker 4: away from the places that you're trying to defend, right?
• 1:04:03
  Speaker 2: So you're disaggregating, you're dispersing.
• 1:04:05
  Speaker 4: You're essentially making it much harder for an adversary to
• 1:04:11
  Speaker 4: be able to take everything else, like all of the
• 1:04:14
  Speaker 4: pieces out when you're defending an area.
• 1:04:17
  Speaker 3: Interesting. So they're going to be using this for logistics.
• 1:04:21
  Speaker 3: They're going to be using this for defense. And I
• 1:04:23
  Speaker 3: would imagine there's going to be an offensive commotant.
• 1:04:26
  Speaker 2: Yeah, yeah.
• 1:04:27
  Speaker 4: So the way that we're thinking about this is, so logistics.
• 1:04:33
  Speaker 4: So you can think about resupply and casualty evacuation. A
• 1:04:37
  Speaker 4: lot of people, when they think about autonomous vehicles, this
• 1:04:39
  Speaker 4: is like the first thing which comes to mind because
• 1:04:42
  Speaker 4: they think about a vehicle and they're like, oh, okay,
• 1:04:44
  Speaker 4: like I can put stuff in it and I can.
• 1:04:46
  Speaker 2: Move it, right? Like that's like what a vehicle does.
• 1:04:48
  Speaker 4: But I like to think about these vehicles more like
• 1:04:52
  Speaker 4: robotic systems. They can sense, they can track, they can
• 1:04:57
  Speaker 4: move on their own on the battlefield. And I think
• 1:05:00
  Speaker 4: like the real product market fit here is that you
• 1:05:03
  Speaker 4: want to move these vehicles out in front of the warfighters, right?
• 1:05:07
  Speaker 4: So you want to be using them for things like intelligence, surveillance,
• 1:05:11
  Speaker 4: and reconnaissance, being able to move them through bad weather.
• 1:05:15
  Speaker 4: They can persistently stay in a location for a long
• 1:05:19
  Speaker 4: period of time because they're just on the ground. So
• 1:05:21
  Speaker 4: they might be more effective than drones for some of
• 1:05:24
  Speaker 4: these things. You can use them for defense or strike capabilities.
• 1:05:31
  Speaker 4: So you can put kinetic payloads on them. You can
• 1:05:33
  Speaker 4: put drones on them. launch them off of the vehicle
• 1:05:37
  Speaker 4: so you can move them into areas that might be
• 1:05:39
  Speaker 4: too risky for a human, but they can potentially hold
• 1:05:43
  Speaker 4: ground or try to take terrain in those areas. Breaching
• 1:05:47
  Speaker 4: is a major thing that we're working on. So we're
• 1:05:49
  Speaker 4: working with a number of different combat engineering units on
• 1:05:54
  Speaker 4: removing people from breaching operations, which are just exceptionally dangerous.
• 1:05:59
  Speaker 4: And then air defense, as I was discussing, is essentially
• 1:06:03
  Speaker 4: putting sensors and shooters in different locations and moving them autonomously,
• 1:06:08
  Speaker 4: reconfiguring what that defensive position might look like.
• 1:06:13
  Speaker 3: I mean, I could think of a whole slew of things,
• 1:06:16
  Speaker 3: but let's run out back and take a look at
• 1:06:18
  Speaker 3: this thing.
• 1:06:19
  Speaker 2: Yeah, sounds great.
• 1:06:22
  Speaker 3: You guys know my schedule. I'm in the studio all day.
• 1:06:26
  Speaker 3: I'm on the road and I've got kids. The last
• 1:06:28
  Speaker 3: thing I have time for is standing around a gym
• 1:06:31
  Speaker 3: wondering what to do next. That's why I'm on the
• 1:06:34
  Speaker 3: ladder app. I've got a real coach in my ear
• 1:06:36
  Speaker 3: on every set. What's next? What weight?
• 1:06:39
  Speaker 6: Why?
• 1:06:40
  Speaker 3: It's a new plan every week that builds on the last.
• 1:06:43
  Speaker 3: And it goes where I go. The garage with dumbbells
• 1:06:46
  Speaker 3: or a hotel room on the road. 30 to 45
• 1:06:49
  Speaker 3: minutes and I'm done for less than a dollar a day.
• 1:06:52
  Speaker 3: I've trained my whole life in the teams. Nothing we
• 1:06:56
  Speaker 3: did was random. Somebody who knew more than you wrote
• 1:06:59
  Speaker 3: the program and you got results. That's the difference between
• 1:07:03
  Speaker 3: working out and training. Ladder puts that kind of real
• 1:07:07
  Speaker 3: programming in your ear. Remove the guesswork with Ladder and
• 1:07:11
  Speaker 3: get a real coach in your ear telling you exactly
• 1:07:14
  Speaker 3: what to do for every workout. No thinking. Everything planned
• 1:07:19
  Speaker 3: for you. If you're a guy who's been meaning to
• 1:07:21
  Speaker 3: get back to it, This is the way back. If
• 1:07:23
  Speaker 3: you have an iPhone, head to ladder.fit slash SRS and
• 1:07:27
  Speaker 3: take a quick quiz to get matched with your coach
• 1:07:30
  Speaker 3: and the right plan for you. Use my link and
• 1:07:33
  Speaker 3: get a free seven-day trial with no credit card and
• 1:07:37
  Speaker 3: $ 10 off your first month if you join. All right,
• 1:07:43
  Speaker 3: we're out here with Byron Boots, CEO and founder of
• 1:07:45
  Speaker 3: Overland AI. Get ready to take a look at this
• 1:07:48
  Speaker 3: beast here. What are we looking at?
• 1:07:51
  Speaker 4: So this is an Ultra, one of our autonomous vehicles.
• 1:07:54
  Speaker 4: So let me just kind of show you what it is.
• 1:07:58
  Speaker 4: You can tell it's an off-road vehicle. It's got long
• 1:08:01
  Speaker 4: travel suspension and big wheels, but it's fully autonomous. So
• 1:08:05
  Speaker 4: it has sensors up in the front of the vehicle.
• 1:08:08
  Speaker 4: There's stereo cameras. There's actually three of them, like one here,
• 1:08:11
  Speaker 4: one there, one on the other side.
• 1:08:14
  Speaker 5: LiDAR.
• 1:08:15
  Speaker 4: So this allows it to see, you know, in this
• 1:08:17
  Speaker 4: whole area in front of the vehicle.
• 1:08:20
  Speaker 1: What degree?
• 1:08:22
  Speaker 4: So actually you can see 360 degrees around it. So
• 1:08:26
  Speaker 4: like you have these sensors in the front and there's
• 1:08:27
  Speaker 4: also sensors in the back. So it can just kind
• 1:08:29
  Speaker 4: of see this whole area. And then if you come
• 1:08:33
  Speaker 4: along this way, this is a payload deck. So you
• 1:08:37
  Speaker 4: can put a wide variety of payloads on it. You
• 1:08:40
  Speaker 4: can see there's all these attachment points. So it makes
• 1:08:42
  Speaker 4: it easy to integrate new payloads onto the vehicle. Below
• 1:08:46
  Speaker 4: this deck you have compute, so that's where your computer
• 1:08:50
  Speaker 4: systems are, your power, batteries, the alternator, like all this
• 1:08:55
  Speaker 4: is below the deck. And then back here, you've got comms,
• 1:09:01
  Speaker 4: so satellite comms. We can integrate tactical mesh comms as
• 1:09:07
  Speaker 4: well on the vehicle. Walking around back here, you can
• 1:09:12
  Speaker 4: see there's more sensors, so stereo cameras and LIDAR, again.
• 1:09:17
  Speaker 1: What does the LIDAR do?
• 1:09:19
  Speaker 4: So LIDAR basically allows you to see depth in the
• 1:09:25
  Speaker 4: area around the robot. So think about it as like
• 1:09:28
  Speaker 4: a depth sensor. It tells you the distance to surfaces
• 1:09:31
  Speaker 4: in the terrain.
• 1:09:33
  Speaker 3: Okay. Yeah.
• 1:09:35
  Speaker 1: What engine are we running in here?
• 1:09:38
  Speaker 4: I think it's... You have to ask Chad. It's about,
• 1:09:43
  Speaker 4: I think, 115 horsepower engine.
• 1:09:45
  Speaker 5: So this is based... This whole vehicle is actually.
• 1:09:48
  Speaker 4: Based on a Polaris M-Razor commercial side-by-side vehicle.
• 1:09:54
  Speaker 3: Oh, okay. Okay?
• 1:09:55
  Speaker 4: So the engine, drive chain, chassis, that all comes from Polaris.
• 1:10:00
  Speaker 4: We take out the seats from the vehicle, take off
• 1:10:03
  Speaker 4: the roll cage and everything, and then transform it into
• 1:10:06
  Speaker 4: this autonomous platform.
• 1:10:09
  Speaker 3: Right on. And so.
• 1:10:13
  Speaker 3: What kind of stuff would you be mounting on.
• 1:10:17
  Speaker 4: Here? So you can put all sorts of different payloads
• 1:10:20
  Speaker 4: on here. This particular vehicle takes about 1,000 pounds of payload.
• 1:10:25
  Speaker 1: 1,000 pounds.
• 1:10:26
  Speaker 4: So we've done everything from fairly straightforward things like comms,
• 1:10:31
  Speaker 4: making a communication node, to electronic warfare payloads, to remote
• 1:10:36
  Speaker 4: weapon stations. So you can mount a machine gun on
• 1:10:40
  Speaker 4: the vehicle.
• 1:10:41
  Speaker 3: Nice. So if you mount a weapon system on here,
• 1:10:46
  Speaker 3: is that all run through the autonomous stack? Is that
• 1:10:49
  Speaker 3: what you call it?
• 1:10:50
  Speaker 4: So the autonomous stack controls the vehicle. It allows an
• 1:10:54
  Speaker 4: operator to essentially tell the vehicle where they want the
• 1:10:58
  Speaker 4: vehicle to go. So it will say something like, you know,
• 1:11:02
  Speaker 4: go 10 kilometers to this location, orient in this way.
• 1:11:06
  Speaker 4: and then they will access the payload through the comms network.
• 1:11:11
  Speaker 4: So the human is still in the loop whenever you're
• 1:11:14
  Speaker 4: using the payload, but the movement of these types of
• 1:11:19
  Speaker 4: vehicles is autonomous. And what that allows you to do
• 1:11:24
  Speaker 4: is instead of having to just remote control the vehicle,
• 1:11:27
  Speaker 4: essentially drive it through its sensors, the human can just say,
• 1:11:32
  Speaker 4: I want one vehicle to go to this location, I
• 1:11:34
  Speaker 4: want two vehicles to go to this location, And so on.
• 1:11:37
  Speaker 4: So it really allows for force multiplication where a single
• 1:11:40
  Speaker 4: operator can control many different platforms and then access the
• 1:11:43
  Speaker 4: payloads on those platforms.
• 1:11:45
  Speaker 2: And we'll show you how to do that.
• 1:11:46
  Speaker 3: It's like the old computer games like Warcraft.
• 1:11:49
  Speaker 2: It's exactly like that.
• 1:11:50
  Speaker 4: So that's actually how we think about it.
• 1:11:54
  Speaker 1: Kill this thing.
• 1:11:55
  Speaker 4: I think our vision of this is a single operator
• 1:11:58
  Speaker 4: could control potentially... hundreds of different assets on the battlefield.
• 1:12:02
  Speaker 4: And you're going to do that through an interface, which
• 1:12:04
  Speaker 4: is a little bit like a real-time strategy game, like
• 1:12:07
  Speaker 4: Warcraft or Starcraft, where it allows you to select the units,
• 1:12:11
  Speaker 4: tell them where to go, execute the payloads, and so on.
• 1:12:16
  Speaker 6: Wow.
• 1:12:17
  Speaker 1: So I just....
• 1:12:20
  Speaker 3: I mean, with the changing landscape of the battlefield now,
• 1:12:26
  Speaker 3: there's a first land... like autonomous land vehicle I've seen.
• 1:12:32
  Speaker 3: So you can mount like an Epirus direct-to-energy weapon on
• 1:12:36
  Speaker 3: this thing. So let's say, I mean, Enduro's got their
• 1:12:40
  Speaker 3: stuff coming out. Shield AI's got that new expat they
• 1:12:44
  Speaker 3: came out with.
• 1:12:46
  Speaker 5: You can mount counter UAS.
• 1:12:48
  Speaker 3: So what I'm asking is, I don't know if we're
• 1:12:51
  Speaker 3: there yet. Maybe we're already there and this is old news.
• 1:12:53
  Speaker 3: But with all these companies, you know, like yours... the submarines,
• 1:12:59
  Speaker 3: the Saronic, you.
• 1:13:01
  Speaker 5: Know, Overland AI.
• 1:13:05
  Speaker 3: When we do go to a full-scale war, are you
• 1:13:08
  Speaker 3: going to be, like, is the operator or the ground
• 1:13:12
  Speaker 3: force commander or just the commander.
• 1:13:14
  Speaker 5: Of the entire operation, are they going to be controlling.
• 1:13:20
  Speaker 3: Shield AIs, expats, Overland AIs, ground vehicles, Saronics, surface warfare vehicles, submarines, drones,
• 1:13:30
  Speaker 3: all of it? All of it on the same system?
• 1:13:33
  Speaker 4: So the military is working on this right now, finding
• 1:13:36
  Speaker 4: ways to integrate all of these different pieces into the
• 1:13:40
  Speaker 4: same sort of system so that you have a unified
• 1:13:44
  Speaker 4: view of the battlefield. Now the way that that may
• 1:13:46
  Speaker 4: play out I think you're likely to see, there are
• 1:13:51
  Speaker 4: single systems where you can see all of the different pieces,
• 1:13:55
  Speaker 4: and then there will likely be systems which will allow
• 1:13:58
  Speaker 4: the war fighters to actually be controlling some subset of
• 1:14:01
  Speaker 4: them on the battlefield. But it will all have to
• 1:14:03
  Speaker 4: be linked together to provide that overall awareness of what
• 1:14:08
  Speaker 4: you're pushing out there.
• 1:14:10
  Speaker 3: And so if you have 100, 200 of these things
• 1:14:12
  Speaker 3: right here, what is the name of this?
• 1:14:14
  Speaker 2: This is an Ultra.
• 1:14:15
  Speaker 1: The Ultra, yeah.
• 1:14:16
  Speaker 3: 100, 200, 300 Ultras out here, and they've got surface-to-air
• 1:14:22
  Speaker 3: missiles for air defense. They've got, I don't know,. 50
• 1:14:27
  Speaker 3: calibers for other ground vehicles and anti-personnel and rocket launchers
• 1:14:32
  Speaker 3: and grenade launchers and drones. Are they all going to
• 1:14:36
  Speaker 3: read off each other, or is somebody going to have to.
• 1:14:39
  Speaker 4: So the way that this is going, we're taking... So
• 1:14:45
  Speaker 4: the basic idea is that you start with just being
• 1:14:49
  Speaker 4: able to move one vehicle at a time. So you
• 1:14:53
  Speaker 4: can say, I want this vehicle to go to this location.
• 1:14:55
  Speaker 4: You just let it go, execute a payload there. We
• 1:14:59
  Speaker 4: already can do that really well. So now we're starting
• 1:15:02
  Speaker 4: to build up coordination where you can move multiple vehicles
• 1:15:06
  Speaker 4: at once and they coordinate in order to achieve a task.
• 1:15:10
  Speaker 5: And then we'll keep building on that.
• 1:15:11
  Speaker 4: So the basic idea is that, you know, over like
• 1:15:15
  Speaker 4: as we build out the technology and as we field
• 1:15:17
  Speaker 4: more and more of it, you're going to have a
• 1:15:20
  Speaker 4: situation where a single operator will be able to move
• 1:15:24
  Speaker 4: multiple vehicles into formations, have them, you know, send them
• 1:15:28
  Speaker 4: to achieve a particular task. And they're going to go
• 1:15:31
  Speaker 4: out there and just do it. And that's part of
• 1:15:35
  Speaker 4: what's called mission autonomy or orchestration. So platform autonomy is
• 1:15:40
  Speaker 4: basically the autonomy that lives on the vehicle that allows
• 1:15:44
  Speaker 4: it to see the terrain, see where vehicles and people
• 1:15:49
  Speaker 4: are out in that terrain and move through it. And
• 1:15:51
  Speaker 4: then you tie that together where vehicles are coordinating with
• 1:15:56
  Speaker 4: each other.
• 1:15:57
  Speaker 5: And that's mission autonomy. They're coordinating.
• 1:16:01
  Speaker 4: autonomously to actually conduct like a full mission and then
• 1:16:05
  Speaker 4: all of that like all of these different autonomous systems
• 1:16:08
  Speaker 4: and uncrewed systems which run the battlefield will be pulled
• 1:16:12
  Speaker 4: into command and control systems which allow people to see
• 1:16:15
  Speaker 4: everything which is is.
• 1:16:17
  Speaker 3: Out there wow i got a i got a million
• 1:16:21
  Speaker 3: questions for you but it's it's hot as shit out
• 1:16:24
  Speaker 3: here and humid so yeah yeah let's see uh Let's see.
• 1:16:27
  Speaker 1: We'll do that inside.
• 1:16:28
  Speaker 5: That sounds great.
• 1:16:30
  Speaker 3: Let's see what this thing can do.
• 1:16:31
  Speaker 5: Do you want to drive it?
• 1:16:32
  Speaker 3: Yes, I want to drive it.
• 1:16:34
  Speaker 4: Okay, so the upper part of this interface here, this
• 1:16:38
  Speaker 4: is where the vehicle is on a satellite map.
• 1:16:42
  Speaker 5: You can see the name of the vehicle.
• 1:16:44
  Speaker 4: And then down here, it gives you a view of
• 1:16:47
  Speaker 4: what the vehicle sees.
• 1:16:48
  Speaker 5: So over here, this is the front camera. on the vehicle.
• 1:16:53
  Speaker 5: And then this.
• 1:16:54
  Speaker 2: Is actually like the AI view of the vehicle.
• 1:16:56
  Speaker 4: You don't have to necessarily spend too much attention like
• 1:16:59
  Speaker 4: looking at this, but, um, the, uh, magenta.
• 1:17:03
  Speaker 2: Areas are lethal. So those are areas.
• 1:17:06
  Speaker 4: That, you know, the vehicle, you don't want to move
• 1:17:08
  Speaker 4: the vehicle into as you're, as you're, teleoperating it, you
• 1:17:11
  Speaker 4: can actually move it, um, anywhere that you want.
• 1:17:13
  Speaker 3: So this is that big pond that we're digging down there.
• 1:17:16
  Speaker 4: Uh, this is, this is going to be about 50
• 1:17:18
  Speaker 4: meters around. So it's only seeing up here on this, um, this,
• 1:17:22
  Speaker 4: this little area, what we can do is give you
• 1:17:25
  Speaker 4: the controller. And the way that this works is when
• 1:17:27
  Speaker 4: you pull, like you have to hold this, uh, down.
• 1:17:30
  Speaker 4: So your lower left and you hit a, and that
• 1:17:35
  Speaker 4: moves you into autonomy. And then you can move this
• 1:17:39
  Speaker 4: joystick moves the vehicle forward.
• 1:17:43
  Speaker 3: Holy shit.
• 1:17:46
  Speaker 4: So that's basically moving it forward and then you can
• 1:17:49
  Speaker 4: turn using this so we can like turn it to
• 1:17:51
  Speaker 4: the right.
• 1:17:51
  Speaker 6: And.
• 1:17:58
  Speaker 4: You know, when the vehicle is here in front of you,
• 1:18:00
  Speaker 4: it's tempting to look at the vehicle, but you should
• 1:18:01
  Speaker 4: actually look at this, which is what the vehicle can see.
• 1:18:06
  Speaker 4: And then you can control it.
• 1:18:08
  Speaker 2: That's right, yeah.
• 1:18:09
  Speaker 5: And you can control it beyond line of sight.
• 1:18:11
  Speaker 4: So you could do this from like 5,000 miles away. Wow.
• 1:18:15
  Speaker 5: So left down, forward. And then push forward.
• 1:18:27
  Speaker 3: It's weird not looking at the vehicle.
• 1:18:30
  Speaker 2: Yeah.
• 1:18:31
  Speaker 3: I want to look over there. Holy shit. How do.
• 1:18:38
  Speaker 5: You go backwards?
• 1:18:45
  Speaker 4: Okay. so to go backwards, hold down this, so both
• 1:18:54
  Speaker 4: that and this, and then move that backwards.
• 1:18:57
  Speaker 2: There you go.
• 1:19:06
  Speaker 3: Oh, there we are. It will automatically stop. It won't
• 1:19:14
  Speaker 3: run us over.
• 1:19:17
  Speaker 2: Let's not test it.
• 1:19:18
  Speaker 6: All right.
• 1:19:18
  Speaker 3: Yes. Yeah, there it is. Slowing for a person.
• 1:19:24
  Speaker 2: Yeah, yeah.
• 1:19:27
  Speaker 1: Well, that's not going to work in war, Byron.
• 1:19:29
  Speaker 3: I'm just kidding.
• 1:19:30
  Speaker 5: You can turn that off.
• 1:19:43
  Speaker 3: Dude, this is crazy. So would somebody be looking at
• 1:19:55
  Speaker 3: this or like a VR headset or does it, I
• 1:19:58
  Speaker 3: guess you could probably do no matter.
• 1:19:59
  Speaker 5: You can do whatever.
• 1:20:01
  Speaker 4: Um, like generally you're looking at this, uh, you can
• 1:20:03
  Speaker 4: look through the other sensors on board the vehicle as well. Um,
• 1:20:07
  Speaker 4: so you can look at the rear sensors or off
• 1:20:09
  Speaker 4: to the sides. You can access the payloads on the
• 1:20:15
  Speaker 4: vehicle through the interface up here.
• 1:20:17
  Speaker 3: Wow.
• 1:20:18
  Speaker 6: All right, and then if you want to.
• 1:20:31
  Speaker 3: Oh, shit. There's a tree.
• 1:20:33
  Speaker 5: Yeah, maybe back up.
• 1:20:34
  Speaker 3: In case you can't tell, I'm not a gamer.
• 1:20:39
  Speaker 4: I'm not great at that either, but we like to
• 1:20:41
  Speaker 4: control it actually through this interface and I'll show you
• 1:20:43
  Speaker 4: that in a moment. I think one of the challenges
• 1:20:47
  Speaker 4: for teleoperation is that you don't have like a vehicle sense, right?
• 1:20:52
  Speaker 4: Like you can't feel like how the vehicle is moving
• 1:20:54
  Speaker 4: like you would when you're driving. And now you're trying
• 1:20:57
  Speaker 4: to interpret what the vehicle can see like through its
• 1:21:00
  Speaker 4: own sensors.
• 1:21:01
  Speaker 2: You should probably stop.
• 1:21:04
  Speaker 3: Yeah.
• 1:21:07
  Speaker 4: What it's, it's actually much safer to put it into
• 1:21:10
  Speaker 4: autonomy and just tell it where you want it to go.
• 1:21:12
  Speaker 4: And it will find a way to get there without,
• 1:21:14
  Speaker 4: you know, hitting anything while staying safe.
• 1:21:17
  Speaker 5: Yeah.
• 1:21:17
  Speaker 2: So you can do that.
• 1:21:19
  Speaker 3: That's awesome.
• 1:21:20
  Speaker 2: All right.
• 1:21:20
  Speaker 3: So I did almost hit that damn tree.
• 1:21:24
  Speaker 5: Okay.
• 1:21:25
  Speaker 3: Let's send it to go get a pizza.
• 1:21:27
  Speaker 2: That sounds great.
• 1:21:29
  Speaker 5: I'm not sure. I'm not sure they'll, uh, let me
• 1:21:32
  Speaker 5: do this, uh, on, on the road.
• 1:21:35
  Speaker 3: Can we go over there?
• 1:21:35
  Speaker 1: Can we go over here and watch it?
• 1:21:40
  Speaker 3: So before you guys run the route, how is that
• 1:21:43
  Speaker 3: thing determining what is a human being and what's a
• 1:21:45
  Speaker 3: tree and what's a vehicle and what's a rock?
• 1:21:48
  Speaker 4: Yeah, it's looking at the whole environment around it and
• 1:21:52
  Speaker 4: then it's determining where it can drive and where it
• 1:21:55
  Speaker 4: can't drive. So that's the first part, this notion of traversability,
• 1:21:58
  Speaker 4: like what part of the terrain is traversable. And then
• 1:22:00
  Speaker 4: on top of that, there's a semantic understanding of the terrain,
• 1:22:04
  Speaker 4: which means you can understand that, you know, it can
• 1:22:07
  Speaker 4: see a person or a car, you know, determine what
• 1:22:11
  Speaker 4: that is, where they are relative to it. And right
• 1:22:14
  Speaker 4: now it's, it's running a safety system, which essentially says like, don't,
• 1:22:18
  Speaker 4: you know, stop if you get too close to a
• 1:22:19
  Speaker 4: person or a vehicle.
• 1:22:22
  Speaker 3: Right on. I mean, how does it differ? Does it
• 1:22:24
  Speaker 3: pick up like body temperature or how.
• 1:22:26
  Speaker 5: Does it look? It looks and so it's just using
• 1:22:30
  Speaker 5: a camera and silhouette. Yeah.
• 1:22:32
  Speaker 4: And saying like, this looks like a person, this looks
• 1:22:34
  Speaker 4: like a vehicle and it's able to pick them out.
• 1:22:37
  Speaker 3: Cool.
• 1:22:37
  Speaker 4: Yeah.
• 1:22:39
  Speaker 3: And this is full autonomy, what we're getting ready to
• 1:22:41
  Speaker 3: do here.
• 1:22:42
  Speaker 5: Yep.
• 1:22:49
  Speaker 3: Holy shit.
• 1:22:52
  Speaker 6: Oh, we're in the way.
• 1:22:53
  Speaker 4: Let's move back away from it. So, yeah, as it's
• 1:23:09
  Speaker 4: moving through the terrain, it's also tracking where all the
• 1:23:11
  Speaker 4: people are, right, where the vehicles are, where they are
• 1:23:13
  Speaker 4: relative to it, and then deciding how to drive.
• 1:23:19
  Speaker 6: It picked him up.
• 1:23:21
  Speaker 5: It picked him up, yeah.
• 1:23:22
  Speaker 3: It's actually pretty wild that it picked him up in
• 1:23:25
  Speaker 3: the middle of all those weeds and trees and shit.
• 1:23:27
  Speaker 6: Yep.
• 1:23:39
  Speaker 4: You can see one of the benefits of, of this is,
• 1:23:41
  Speaker 4: you know, you can just tell it, I want you
• 1:23:43
  Speaker 4: to go to again, like this location and you can
• 1:23:46
  Speaker 4: let it go and it will, will do its thing. Um,
• 1:23:50
  Speaker 4: so I was in the run, basically came up the hill,
• 1:23:52
  Speaker 4: you know, went around the field and then, uh, stopped there.
• 1:23:56
  Speaker 4: So very short run for what it normally does, but
• 1:23:58
  Speaker 4: you can get a taste of kind of like how
• 1:24:00
  Speaker 4: it moves.
• 1:24:01
  Speaker 3: That is sick. All right. Well, thanks for showing us, uh,
• 1:24:05
  Speaker 3: what the ultra can do and let's go wrap up
• 1:24:07
  Speaker 3: the interview.
• 1:24:08
  Speaker 5: Yeah, thanks for letting us come out here and show it.
• 1:24:10
  Speaker 1: Thanks for bringing that badass thing.
• 1:24:15
  Speaker 3: There's nothing more frustrating than putting in the work and
• 1:24:18
  Speaker 3: still feeling like your body is not responding the way
• 1:24:21
  Speaker 3: it used to. You're training. You're trying to eat right.
• 1:24:24
  Speaker 3: You're trying to sleep. But energy still drops off. Recovery
• 1:24:29
  Speaker 3: takes longer. And you don't have that same drive that
• 1:24:32
  Speaker 3: you used to have. That's why I started taking Mars Men.
• 1:24:36
  Speaker 3: What I like about Mars Men... is that it's not
• 1:24:39
  Speaker 3: forcing hormones into your body. It's designed to help your
• 1:24:42
  Speaker 3: body use more of what it's already producing. As men
• 1:24:46
  Speaker 3: get older, SHBG can bind up testosterone. So even if
• 1:24:51
  Speaker 3: your levels look fine on paper, a lot of it
• 1:24:53
  Speaker 3: may not be available for your body to actually use.
• 1:24:57
  Speaker 3: Mars Men uses clinically dosed ingredients to help free up
• 1:25:00
  Speaker 3: bound testosterone, plus vitamin D and zinc to support healthy
• 1:25:04
  Speaker 3: T levels, energy, stamina, and recovery. And since taking it,
• 1:25:09
  Speaker 3: I've noticed steadier energy and better recovery. It's not like
• 1:25:12
  Speaker 3: a caffeine spike. It's more consistent drive throughout the day.
• 1:25:16
  Speaker 3: Mars Men is made in the USA, third-party tested, and
• 1:25:20
  Speaker 3: backed by a 90-day money-back guarantee. For a limited time,
• 1:25:24
  Speaker 3: our listeners get 50% off for life, plus free shipping. And...
• 1:25:30
  Speaker 3: Three free gifts at mengotomars.com. That's mengotomars.com for 50% off
• 1:25:36
  Speaker 3: and three free gifts when you check out. After you purchase,
• 1:25:41
  Speaker 3: they'll ask you where you heard about them. Please support
• 1:25:44
  Speaker 3: our show and tell them our show sent you. Okay,
• 1:25:50
  Speaker 3: so after taking a look at this thing, I.
• 1:25:54
  Speaker 5: This.
• 1:25:56
  Speaker 3: This just has.
• 1:25:58
  Speaker 3: So many different capabilities that could be used for so
• 1:26:00
  Speaker 3: many different things. I mean, even just keep, I mean,
• 1:26:03
  Speaker 3: you admit, it sounds like the first thing that the
• 1:26:05
  Speaker 3: military is interested in is, is, is logistics, um, which,
• 1:26:10
  Speaker 3: I mean, just keeping major supply routes open. Yeah. I mean,
• 1:26:18
  Speaker 3: constant detection. Could these things detect, I mean, when we
• 1:26:21
  Speaker 3: were up there, it detected, I mean, I know you
• 1:26:24
  Speaker 3: guys have that safety feature on it, but it detected, uh, uh,
• 1:26:27
  Speaker 3: one of my camera guys who was, I mean, really
• 1:26:30
  Speaker 3: wasn't even that close to it. It looked like he
• 1:26:32
  Speaker 3: was maybe 25 yards, 30 yards from it in, in
• 1:26:35
  Speaker 3: the middle of a bunch of trees and it picked
• 1:26:37
  Speaker 3: him up like that. Could this thing pick up IEDs?
• 1:26:44
  Speaker 3: Is there a way that you could, that it could
• 1:26:46
  Speaker 3: detect IEDs?
• 1:26:47
  Speaker 2: I mean, yeah.
• 1:26:48
  Speaker 4: So you, you can put a variety of different types
• 1:26:50
  Speaker 4: of sensors on the vehicle. So, um, you can certainly
• 1:26:56
  Speaker 4: detect like people and vehicles, and then you can also,
• 1:26:59
  Speaker 4: you know, we've been putting tethered.
• 1:27:02
  Speaker 2: Drones on these.
• 1:27:03
  Speaker 4: So, you know, the drone can go up in the air,
• 1:27:06
  Speaker 4: be able to look down on the ground. So if
• 1:27:09
  Speaker 4: you have, you know, ground penetrating radar, other types of sensors,
• 1:27:14
  Speaker 4: you can certainly detect things like IEDs or obstacles in
• 1:27:18
  Speaker 4: the environment that then can be identified and reduced, right? So,
• 1:27:24
  Speaker 4: that's certainly something they can do. And to your point,
• 1:27:28
  Speaker 4: you know, I think you can, of course, move things around,
• 1:27:30
  Speaker 4: but just the ability to push these forward and sense
• 1:27:35
  Speaker 4: the environment is, you know, I think will be game-changing, right?
• 1:27:40
  Speaker 2: You don't have to put a person out there to
• 1:27:41
  Speaker 2: do it.
• 1:27:42
  Speaker 3: Do you see these integrating with human forces that are,
• 1:27:50
  Speaker 3: you know, Do you see these as a forward observations
• 1:27:54
  Speaker 3: platform for maybe a slew of tanks or MRADs or
• 1:28:00
  Speaker 3: whatever we're using and then kind of pushing that information
• 1:28:05
  Speaker 3: back to a larger force or is it going all autonomous?
• 1:28:11
  Speaker 4: I think the– so there's a number of statistics out
• 1:28:15
  Speaker 4: there which are pretty interesting. So you've got like Ukrainian
• 1:28:18
  Speaker 4: ground commanders basically saying that they're going to replace something
• 1:28:23
  Speaker 4: like 80% of infantry with uncrewed ground vehicles in the
• 1:28:28
  Speaker 4: near term.
• 1:28:30
  Speaker 2: Those statistics are interesting.
• 1:28:32
  Speaker 4: I think the reality will be a little bit more subtle.
• 1:28:35
  Speaker 4: So these are machines that can integrate with human formations.
• 1:28:42
  Speaker 4: You should think about them as giving humans more capability
• 1:28:46
  Speaker 4: on the battlefield. So I think the way that they'll
• 1:28:51
  Speaker 4: actually work with the US military is that we'll have
• 1:28:55
  Speaker 4: formations with humans in them. They will be pushing these
• 1:28:58
  Speaker 4: vehicles out in front, as you suggested, in order to
• 1:29:02
  Speaker 4: be able to get a better understanding of the terrain
• 1:29:05
  Speaker 4: that they're going to be moving into. But you can
• 1:29:08
  Speaker 4: use them for all sorts of things, right? So it's
• 1:29:09
  Speaker 4: not just sensing what's out there. You can use them
• 1:29:12
  Speaker 4: to create, like, diversions, right? Like, you can use them,
• 1:29:17
  Speaker 4: you know, to, you can use them, you know, for
• 1:29:21
  Speaker 4: strike capability. You can use them to create that kind
• 1:29:24
  Speaker 4: of counter UAS bubble to protect humans. I think about
• 1:29:28
  Speaker 4: them as, you know, something that will be like a
• 1:29:31
  Speaker 4: force multiplier.
• 1:29:33
  Speaker 3: I mean, you're saying that Ukrainian military is saying that
• 1:29:38
  Speaker 3: it will replace up to 80% of infantry, 80% of
• 1:29:46
  Speaker 3: the infantry.
• 1:29:47
  Speaker 2: Yeah.
• 1:29:48
  Speaker 3: And you're saying the U.S. is saying, well, maybe not
• 1:29:53
  Speaker 3: the U.S., but you see a subtle integration. Why would
• 1:29:56
  Speaker 3: we do a subtle integration? I mean, I know on
• 1:29:58
  Speaker 3: one hand I'm going to piss the warfighter off because
• 1:30:00
  Speaker 3: they love fighting wars. On the other hand, I'm thinking
• 1:30:03
  Speaker 3: about my toddlers at home, you know, with all the
• 1:30:07
  Speaker 3: wars that we're involved in and re-involved in, and I
• 1:30:11
  Speaker 3: don't want my kids going over there having to do it.
• 1:30:14
  Speaker 2: I get it.
• 1:30:15
  Speaker 3: You know, and so why do you see a subtle
• 1:30:18
  Speaker 3: integration of this? Why wouldn't we go full scale and say, hey,
• 1:30:21
  Speaker 3: you know, like we don't need– Private Joe driving this
• 1:30:29
  Speaker 3: MRAP into hostile territory. I don't want to say a
• 1:30:35
  Speaker 3: wasted life, but it would be a wasted life if
• 1:30:38
  Speaker 3: he got killed when we have the capability of going
• 1:30:41
  Speaker 3: full autonomous right now.
• 1:30:43
  Speaker 2: I think we'll eventually get there.
• 1:30:45
  Speaker 4: So one of the value propositions for ground autonomy is
• 1:30:50
  Speaker 4: that you're really focused on the....
• 1:30:55
  Speaker 2: One of the hardest domains, right, in the military, right?
• 1:30:57
  Speaker 4: So just kind of stepping back, there are a lot
• 1:31:01
  Speaker 4: of folks who are focused on air power or naval
• 1:31:04
  Speaker 4: power or missiles. But at the end of the day,
• 1:31:08
  Speaker 4: wars are ultimately won on the ground. You know, that's
• 1:31:11
  Speaker 4: where people live, right? We don't live in the sea.
• 1:31:13
  Speaker 4: We don't live in the air.
• 1:31:14
  Speaker 2: We live on the ground. And that's, you know, where
• 1:31:17
  Speaker 2: we have to fight.
• 1:31:18
  Speaker 5: And, you know, I think.
• 1:31:21
  Speaker 4: When you look at the US and our casualty rates,
• 1:31:28
  Speaker 4: something like more than 80% of combat-related casualties since World
• 1:31:33
  Speaker 4: War II are infantry. So of course you want to like,
• 1:31:37
  Speaker 4: you know, I think like focusing on the ground makes
• 1:31:40
  Speaker 4: a lot of sense because we want to bring technology
• 1:31:42
  Speaker 4: to the warfighter that will save those lives, right? Yes,
• 1:31:49
  Speaker 4: we want to take humans out of harm's way. And
• 1:31:53
  Speaker 4: I think, you know, in Ukraine, we're seeing technology, you know,
• 1:31:59
  Speaker 4: increasingly being pushed out in front of the warfighter. And
• 1:32:02
  Speaker 4: I think they're very optimistic about how they want to
• 1:32:05
  Speaker 4: use them. And ultimately, we do want to do that, right?
• 1:32:08
  Speaker 4: We want to pull, we want to essentially reduce risk
• 1:32:13
  Speaker 4: on the battlefield, right? Pull warfighters away from these points
• 1:32:16
  Speaker 4: of contact. have more people who are maybe operating teams
• 1:32:20
  Speaker 4: of robots that are far away from a location where
• 1:32:25
  Speaker 4: they could get hit. But in the meantime, I think
• 1:32:29
  Speaker 4: that the way they're going to be integrated is in
• 1:32:32
  Speaker 4: a way that helps warfighters on the ground to stay safer,
• 1:32:36
  Speaker 4: to potentially give them more options, more tools, without totally
• 1:32:42
  Speaker 4: pulling them out. So to your point, yes, we want
• 1:32:45
  Speaker 4: to ensure that we ultimately want to pull as many
• 1:32:53
  Speaker 4: people out of harm's way as possible. But these are
• 1:32:57
  Speaker 4: still machines that aren't quite as creative and adaptable as humans.
• 1:33:03
  Speaker 2: And so I think.
• 1:33:05
  Speaker 4: What we're really trying to do is find the places
• 1:33:07
  Speaker 4: where we can maximize the ability for them to be
• 1:33:13
  Speaker 4: able to take risk while also allowing our humans to
• 1:33:19
  Speaker 4: do what they do best and do maybe more specialized
• 1:33:24
  Speaker 4: roles on the battlefield.
• 1:33:26
  Speaker 2: So I think it will be a process.
• 1:33:28
  Speaker 3: I mean, just thinking about my time in Iraq, Afghanistan alone,
• 1:33:33
  Speaker 3: I mean, I have right– IED threat was huge. And
• 1:33:37
  Speaker 3: then here is a key statistic here. 44% of U.S.
• 1:33:41
  Speaker 3: killed in action from 2006 to 2021 were killed by IEDs. 44%. Now,
• 1:33:48
  Speaker 3: I mean, if these things did have IED detection capabilities,
• 1:33:57
  Speaker 3: it's got to be better than humans. I mean, there's
• 1:34:03
  Speaker 3: almost 50% of people that were killed would still be
• 1:34:06
  Speaker 3: here today.
• 1:34:07
  Speaker 5: Yeah.
• 1:34:07
  Speaker 3: You know, just with... Just in that application.
• 1:34:11
  Speaker 4: I think there's a lot you can do there. I mean,
• 1:34:13
  Speaker 4: there's sensors you can put on the vehicles.
• 1:34:15
  Speaker 3: The human sensor, like I was just saying, would caught
• 1:34:18
  Speaker 3: my guy out in the weeds trying to, in a
• 1:34:21
  Speaker 3: bunch of trees. I mean, rolling down Iraq, I mean,
• 1:34:24
  Speaker 3: there were sniper problems. There was the guy, the trigger
• 1:34:27
  Speaker 3: men of IEDs. I mean, I would think that it
• 1:34:31
  Speaker 3: would pick them up immediately. I mean, they can't hide
• 1:34:34
  Speaker 3: from the machine. Yeah. How is it sensing them?
• 1:34:37
  Speaker 2: Yeah.
• 1:34:38
  Speaker 4: So the machine has cameras on it, right? And depending
• 1:34:42
  Speaker 4: on the loadout for it, you can have sort of
• 1:34:47
  Speaker 4: normal RGB cameras. You can have thermal cameras, right? Like
• 1:34:53
  Speaker 4: there's a lot that you can do to detect people
• 1:34:56
  Speaker 4: and vehicles in the environment. So if I was just
• 1:34:59
  Speaker 4: using cameras out there, I was able, as you said,
• 1:35:01
  Speaker 4: just be able to pick someone up even in the
• 1:35:04
  Speaker 4: woods and in the weeds. But you know, you should
• 1:35:08
  Speaker 4: think about, like, there's a lot of different ways that
• 1:35:10
  Speaker 4: these can be used out, again, out in front of
• 1:35:13
  Speaker 4: the warfighter. You know, you can have convoys of autonomous
• 1:35:18
  Speaker 4: vehicles that don't have people on them, right, that can
• 1:35:20
  Speaker 4: move very quickly and move supplies back and forth. Even
• 1:35:23
  Speaker 4: if they're targeted, at least no human is being killed.
• 1:35:26
  Speaker 4: You can have them, autonomous vehicles, out in front of
• 1:35:31
  Speaker 4: human convoys as well, right? So, really making sure that
• 1:35:35
  Speaker 4: there is no one out there that, you know, moving
• 1:35:38
  Speaker 4: first through the terrain. So generally speaking, there's a lot
• 1:35:42
  Speaker 4: of ideas in Army and Marine Corps about how to
• 1:35:46
  Speaker 4: use these vehicles out in front, again, like leading convoys
• 1:35:50
  Speaker 4: or it's like a protective onion almost, right? You can
• 1:35:52
  Speaker 4: think about a whole set of vehicles that are surrounding
• 1:35:57
  Speaker 4: higher value assets, right? whether they're people or tanks or
• 1:36:03
  Speaker 4: Bradley fighting vehicles or XM-30, which is a replacement for that,
• 1:36:07
  Speaker 4: having autonomous vehicles that can provide sensing and protection for
• 1:36:14
  Speaker 4: those formations. So we think about them as adding to
• 1:36:17
  Speaker 4: some of the things that we're developing.
• 1:36:20
  Speaker 3: How many of these scenarios have you kind of war-gamed
• 1:36:23
  Speaker 3: or run scenarios on out at wherever?
• 1:36:26
  Speaker 2: Uh, quite, quite a few.
• 1:36:27
  Speaker 4: I mean, I think people are, once they start to
• 1:36:29
  Speaker 4: see the vehicle and they start to think about it,
• 1:36:31
  Speaker 4: they are coming up with lots of ways of, of
• 1:36:35
  Speaker 4: potentially using them. Uh, and as we, you know, one
• 1:36:38
  Speaker 4: of the things that we've been trying to do in
• 1:36:39
  Speaker 4: the last year is work with a lot of different units.
• 1:36:42
  Speaker 4: So we've worked with over 20 different, um, different units
• 1:36:46
  Speaker 4: integrating this technology into the unit. Uh, for example, the
• 1:36:50
  Speaker 4: 82nd airborne, uh, we worked with 382 for, almost six months,
• 1:36:56
  Speaker 4: just embedded with them up through one of their training
• 1:36:58
  Speaker 4: rotations down at JRTC in Louisiana. And we started out
• 1:37:04
  Speaker 4: just doing resupply. They were like, this is kind of
• 1:37:06
  Speaker 4: the obvious use case for this. Let's use this to
• 1:37:10
  Speaker 4: support our logistics to allow us to move faster.
• 1:37:13
  Speaker 2: And we were able to do that.
• 1:37:14
  Speaker 4: And so then they were like, okay, like, can we
• 1:37:16
  Speaker 4: get cameras and drones and things like that on it?
• 1:37:19
  Speaker 2: The answer is yes.
• 1:37:20
  Speaker 4: And so We started to add different types of payloads
• 1:37:23
  Speaker 4: to it. They started to use them in more creative ways.
• 1:37:27
  Speaker 4: And by the end of this, they were using these
• 1:37:29
  Speaker 4: autonomous vehicles in this force on force exercise, not just
• 1:37:35
  Speaker 4: for resupply, but also for intelligence surveillance and reconnaissance.
• 1:37:39
  Speaker 2: There were snipers who were using the.
• 1:37:40
  Speaker 4: Vehicles as decoys. There's just a lot of things you
• 1:37:44
  Speaker 4: can start to do. The first step is really just
• 1:37:50
  Speaker 4: getting it in the hands of the warfighter, right? Let
• 1:37:52
  Speaker 4: them think about the problems they're trying to solve, experience
• 1:37:56
  Speaker 4: the technology, and then start to iterate on potentially new
• 1:38:03
  Speaker 4: tactics and things like this that you can do with
• 1:38:06
  Speaker 4: the vehicles.
• 1:38:07
  Speaker 3: Now, are these truly autonomous, or is there an operator
• 1:38:12
  Speaker 3: in the rear that's going to have to be behind
• 1:38:16
  Speaker 3: each one of these things with a remote control?
• 1:38:18
  Speaker 2: Yeah, this is a great question.
• 1:38:22
  Speaker 4: So, you know, and this is sometimes kind of glossed over.
• 1:38:27
  Speaker 4: A lot of people say that they can do autonomous things.
• 1:38:30
  Speaker 4: and people think about uncrewed ground vehicles. People talk about
• 1:38:33
  Speaker 4: UGVs all the time in Ukraine versus autonomous vehicles. So
• 1:38:36
  Speaker 4: what are the differences between these things, right? So remote
• 1:38:39
  Speaker 4: control basically means you have a controller like in your
• 1:38:42
  Speaker 4: hand and you're looking at the thing and you're driving it.
• 1:38:46
  Speaker 4: So think about like a toy car, right? Like that's
• 1:38:48
  Speaker 4: a remote control Now, that's fine, except you have to
• 1:38:54
  Speaker 4: actually look at the terrain yourself, look at the vehicle
• 1:38:57
  Speaker 4: and determine where it's going to go. That obviously doesn't
• 1:38:59
  Speaker 4: work beyond line of sight or in areas where you
• 1:39:02
  Speaker 4: can't see the terrain very well. So the next sort
• 1:39:06
  Speaker 4: of evolution of this is teleoperation. And people talk about
• 1:39:10
  Speaker 4: teleoperation a lot with ground vehicles. This is basically where
• 1:39:13
  Speaker 4: you take a remote control and you look through the
• 1:39:16
  Speaker 4: vehicle's own sensors, whether it's like camera or thermal sensors.
• 1:39:21
  Speaker 4: And then you drive the vehicle based on that sensory
• 1:39:24
  Speaker 4: feedback that you're getting over a network. So you can
• 1:39:26
  Speaker 4: go beyond line of sight because you just see what
• 1:39:29
  Speaker 4: the vehicle sees. You don't have to see the vehicle itself.
• 1:39:32
  Speaker 4: So you can teleoperate something from like across the world
• 1:39:35
  Speaker 4: if you have a really good connection to that vehicle,
• 1:39:39
  Speaker 4: if you have a good satellite connection or radio connection
• 1:39:42
  Speaker 4: to the vehicle. And a lot of the work which
• 1:39:44
  Speaker 4: is being done in Ukraine is teleoperated uncrewed ground vehicles,
• 1:39:48
  Speaker 4: where you'll actually have a set of people, like up
• 1:39:51
  Speaker 4: to five people, who are controlling each vehicle and essentially
• 1:39:56
  Speaker 4: driving it through the terrain using, you know, like a
• 1:39:59
  Speaker 4: handheld controller. Or, you know, they're looking through potentially either
• 1:40:03
  Speaker 4: the vehicle's cameras or the cameras from a drone overhead,
• 1:40:06
  Speaker 4: and they're just trying to drive this thing around. The
• 1:40:09
  Speaker 4: thing which is hard about that is that It occupies
• 1:40:12
  Speaker 4: at least one person's attention at all times because the
• 1:40:17
  Speaker 4: person's making all.
• 1:40:17
  Speaker 2: Of those decisions. But it also is.
• 1:40:22
  Speaker 4: Something where if your comms gets disrupted, the thing's just dead, right?
• 1:40:26
  Speaker 4: Like it requires a human to drive it and the
• 1:40:28
  Speaker 4: human can no longer connect to it. So one of
• 1:40:32
  Speaker 4: the strategies is you cut the comms to the vehicle,
• 1:40:35
  Speaker 4: you use EW or whatever to cut the comms, and
• 1:40:37
  Speaker 4: then you strike it because it's just a sitting duck.
• 1:40:42
  Speaker 4: Autonomy is extremely powerful because it allows the vehicle to
• 1:40:47
  Speaker 4: sense the environment, represent it, plan through the environment all
• 1:40:51
  Speaker 4: on board, like at the edge, makes its own decisions.
• 1:40:54
  Speaker 4: And it doesn't require a human to essentially drive it.
• 1:40:59
  Speaker 4: So when you have an autonomous vehicle, you can tell
• 1:41:03
  Speaker 4: it where you want it to go and it will
• 1:41:06
  Speaker 4: just go there. And that means that you can focus
• 1:41:09
  Speaker 4: your attention on something else. That might be another autonomous vehicle.
• 1:41:12
  Speaker 4: It might be another task. So for example, you can
• 1:41:15
  Speaker 4: send a vehicle back to resupply you, but you don't
• 1:41:18
  Speaker 4: have to be focused on it anymore.
• 1:41:20
  Speaker 2: You can do your job.
• 1:41:22
  Speaker 4: The second thing about autonomy, which is important to understand,
• 1:41:26
  Speaker 4: is because the vehicle is making decisions based on its
• 1:41:30
  Speaker 4: own sensing and own onboard compute, if your comms get cut,
• 1:41:35
  Speaker 4: the thing's going to keep going and continue the mission.
• 1:41:40
  Speaker 4: And so one of the things we're starting to see
• 1:41:41
  Speaker 4: out of the Russian-Ukraine conflict is that the Russians are
• 1:41:46
  Speaker 4: starting to add autonomy to ground vehicles to deal with
• 1:41:50
  Speaker 4: the contested comms situation. Because it's so hard for them
• 1:41:53
  Speaker 4: to maintain comms, autonomy allows the vehicles to continue to
• 1:41:57
  Speaker 4: move without human oversight. So that's another piece of this.
• 1:42:03
  Speaker 4: So we think about this as kind of force multiplication
• 1:42:06
  Speaker 4: and resilience to... a contested comms environment.
• 1:42:10
  Speaker 3: Well, now, so another thing with autonomy to correct me
• 1:42:13
  Speaker 3: if I'm wrong is the, let's say you have a
• 1:42:16
  Speaker 3: hundred of these, you say have 500 of these.
• 1:42:19
  Speaker 6: Yeah.
• 1:42:19
  Speaker 3: And then you have all these other, you know, we
• 1:42:21
  Speaker 3: were talked about at the beginning, all these other companies
• 1:42:23
  Speaker 3: that are doing, you know, we shield AI with their,
• 1:42:27
  Speaker 3: with their expat. We got Saronic with the, with the
• 1:42:30
  Speaker 3: autonomous boats. We got Enduro making stuff. We have a
• 1:42:34
  Speaker 3: mock making stuff with all these companies are jumping all over,
• 1:42:37
  Speaker 3: you know, Lots of autonomous stuff. But just overland AI alone.
• 1:42:44
  Speaker 3: I mean, let's say that there's 500 of these vehicles. Yeah.
• 1:42:49
  Speaker 3: We're talking, I don't know, the invasion of Fallujah. Yeah.
• 1:42:53
  Speaker 3: These will all be able to communicate with each other
• 1:42:57
  Speaker 3: and accomplish the mission without each one being micromanaged. They'll
• 1:43:04
  Speaker 3: all read off each other, communicate with each other, know
• 1:43:06
  Speaker 3: what everything is doing, correct?
• 1:43:08
  Speaker 2: Yeah.
• 1:43:09
  Speaker 4: So think about that. So this is, this is our
• 1:43:12
  Speaker 4: vision for, for the company. So you should think about
• 1:43:15
  Speaker 4: it as like, imagine that there's five people that are
• 1:43:19
  Speaker 4: controlling something like 500 vehicles and like each of those
• 1:43:24
  Speaker 4: five people, you know, maybe each one's controlling a hundred
• 1:43:27
  Speaker 4: vehicles and those vehicles are coordinating with each other to,
• 1:43:32
  Speaker 4: to complete that mission. So we think about the autonomy,
• 1:43:35
  Speaker 4: which is onboard the vehicle. This is, what we call
• 1:43:38
  Speaker 4: platform autonomy. It's how each individual vehicle analyzes terrain, makes decisions,
• 1:43:44
  Speaker 4: decides how to drive. Then there's a notion of mission autonomy,
• 1:43:49
  Speaker 4: where multiple vehicles can coordinate with each other. And the
• 1:43:53
  Speaker 4: real idea here is to make it really easy for
• 1:43:56
  Speaker 4: one person to have an immense effect on the battlefield,
• 1:44:00
  Speaker 4: to control many, many different vehicles and the payloads on them.
• 1:44:05
  Speaker 4: And so going back to earlier part of the conversation,
• 1:44:09
  Speaker 4: we view this as something like a gamer who's playing
• 1:44:13
  Speaker 4: a real-time strategy game, something like StarCraft, where you're controlling
• 1:44:18
  Speaker 4: 200 different units. Can a person do that in the
• 1:44:21
  Speaker 4: real world with real platforms like our Ultra platform? And
• 1:44:27
  Speaker 4: that's what we're building up towards. So that whole idea
• 1:44:32
  Speaker 4: is single operator, massive force through potentially hundreds of different
• 1:44:38
  Speaker 4: vehicles and payloads. Wow.
• 1:44:42
  Speaker 3: And so will it move to the point where, you know,
• 1:44:44
  Speaker 3: I just brought up all these other companies. Will it
• 1:44:46
  Speaker 3: move to the point where these Overland AI vehicles are
• 1:44:52
  Speaker 3: communicating with Shield AI's expat with Saronix boats? I mean,
• 1:44:59
  Speaker 3: will it get to the point where the entire battle
• 1:45:01
  Speaker 3: space is coordinated under one brain?
• 1:45:07
  Speaker 4: So, uh, you know, the, the, the short answer is
• 1:45:11
  Speaker 4: that like, yes, the ground vehicles will be communicating with, uh, um,
• 1:45:16
  Speaker 4: like aerial vehicles and the different payloads and they'll be
• 1:45:20
  Speaker 4: coordinating in.
• 1:45:21
  Speaker 2: Order to, to complete a task.
• 1:45:24
  Speaker 4: Um, I think we still, you know, one of the
• 1:45:27
  Speaker 4: things which is really important to keep in mind is again,
• 1:45:30
  Speaker 4: like these are all just tools. So, um, you really
• 1:45:34
  Speaker 4: want to enable, um, an operator to achieve their objective
• 1:45:39
  Speaker 4: as effectively as possible. And that means, you know, not
• 1:45:44
  Speaker 4: just kind of handing everything over to an AI brain,
• 1:45:47
  Speaker 4: but having, you know, essentially like AI assistants who are
• 1:45:54
  Speaker 4: allowing a human to choose courses of action, to coordinate,
• 1:46:02
  Speaker 4: you know, ground vehicles or payloads or, um, aerial vehicles
• 1:46:06
  Speaker 4: more effectively and, and so on.
• 1:46:08
  Speaker 3: Let's talk about, let's talk about African lion. What happened there? Okay.
• 1:46:14
  Speaker 4: So, um, African lion is, is an exercise. It's, um,
• 1:46:18
  Speaker 4: one of the largest exercises, uh, in Africa. So it's
• 1:46:20
  Speaker 4: a place where, um, us war fighters work with, with
• 1:46:24
  Speaker 4: partner nations and things like that, um, to, uh, essentially, um,
• 1:46:30
  Speaker 4: train and, and demonstrate capabilities.
• 1:46:33
  Speaker 2: We were part.
• 1:46:34
  Speaker 4: Of African Lion working with 173rd Airborne. So the 173rd
• 1:46:39
  Speaker 4: Airborne had two of our ultra vehicles with different types
• 1:46:43
  Speaker 4: of payloads on them. One vehicle had a Crow's remote
• 1:46:47
  Speaker 4: weapon station with an M240 machine gun on it. Another
• 1:46:51
  Speaker 4: vehicle had a rocket propelled breaching system on it. And
• 1:46:58
  Speaker 4: they were using these vehicles for a um, breaching operation. Um,
• 1:47:04
  Speaker 4: so essentially what happened was, uh, they sent one vehicle
• 1:47:08
  Speaker 4: forward with machine gun on it, providing security. And then
• 1:47:11
  Speaker 4: a second vehicle, um, was quickly following it, moved to
• 1:47:15
  Speaker 4: a breach point, uh, essentially applied the payload. So the, um,
• 1:47:21
  Speaker 4: it's a explosive line charge, um, something like a Miklik,
• 1:47:24
  Speaker 4: if people know what that is, but essentially an explosive
• 1:47:27
  Speaker 4: rope that gets launched out in front of the vehicle
• 1:47:30
  Speaker 4: and produces a it blows up and produces like a
• 1:47:34
  Speaker 4: safe corridor that can then be proofed with another vehicle.
• 1:47:38
  Speaker 4: So they're using this to reduce obstacles out in front
• 1:47:43
  Speaker 4: of the force as part of an assault. And the
• 1:47:47
  Speaker 4: thing which is really important about this is that it's
• 1:47:49
  Speaker 4: an extremely dangerous operation. Like when you're doing things like breaching,
• 1:47:53
  Speaker 4: every area of the defensive obstacle belts are being watched
• 1:47:59
  Speaker 4: by an adversary And as you move forward to try
• 1:48:03
  Speaker 4: to reduce those obstacles, the combat engineers, everyone's targeting them.
• 1:48:08
  Speaker 4: So even in a successful breaching operation, you're expecting something
• 1:48:11
  Speaker 4: like 50% casualty rate. So if you're able to do
• 1:48:17
  Speaker 4: that with autonomous vehicles, like the way that 173rd Airborne
• 1:48:22
  Speaker 4: was demonstrating, that's taking, in their estimate, up to almost
• 1:48:27
  Speaker 4: two platoons, about 40 people out of that extremely dangerous situation.
• 1:48:34
  Speaker 2: You're just sending the machines.
• 1:48:35
  Speaker 4: Forward to do that, to work that problem, to create
• 1:48:38
  Speaker 4: the breach in the obstacle belt, and then they're able
• 1:48:42
  Speaker 4: to move through. So they're very excited about the technology.
• 1:48:46
  Speaker 4: But it's just an example of, like, warfighters using our vehicles,
• 1:48:50
  Speaker 4: putting payloads on them, coordinating as part of their unit's
• 1:48:55
  Speaker 4: movement maneuver.
• 1:48:57
  Speaker 2: So pretty exciting to see.
• 1:48:59
  Speaker 3: This is fascinating stuff.
• 1:49:01
  Speaker 6: Yeah. Wow.
• 1:49:02
  Speaker 3: The game has changed a lot.
• 1:49:04
  Speaker 2: Yeah.
• 1:49:04
  Speaker 8: 40 years. Yeah.
• 1:49:05
  Speaker 3: Holy shit. Here's the pitch. It takes two minutes, and
• 1:49:14
  Speaker 3: you'll never think about it again. Download it, hit one button,
• 1:49:18
  Speaker 3: close the app. That's the entire job. No manual, no settings,
• 1:49:24
  Speaker 3: no tech background.
• 1:49:25
  Speaker 1: Nothing for you to get wrong.
• 1:49:27
  Speaker 3: And then, it just runs. Blocking the trackers, the phishing sites,
• 1:49:33
  Speaker 3: and the surveillance ads across every app on the device.
• 1:49:37
  Speaker 1: All day, quietly, without anybody having to do anything.
• 1:49:42
  Speaker 3: It's first week out, it hit number six on Apple's
• 1:49:45
  Speaker 3: top downloaded productivity chart, built by former US intelligence professionals.
• 1:49:52
  Speaker 3: I'm a co-founder. Glacier, download it on the App Store.
• 1:49:56
  Speaker 3: And remember, privacy isn't paranoia, it's protection.
• 1:50:07
  Speaker 1: All right, so we're back from the break.
• 1:50:10
  Speaker 3: And, you know, one thing I'm kind of wondering is
• 1:50:15
  Speaker 3: what other, I mean, you had mentioned Russia was doing
• 1:50:17
  Speaker 3: some of this, correct? What is, what about China? Are they,
• 1:50:20
  Speaker 3: they seem to be pretty on the ball with just
• 1:50:23
  Speaker 3: about everything.
• 1:50:25
  Speaker 4: Yeah, we don't have, or at least I don't have
• 1:50:28
  Speaker 4: detailed knowledge of everything that they're doing, but we do
• 1:50:32
  Speaker 4: read academic papers that come out of Chinese institutions and
• 1:50:37
  Speaker 4: They're certainly working on ground robots. I think there's several
• 1:50:40
  Speaker 4: different types. They've been working on robotic dogs and integrating
• 1:50:44
  Speaker 4: those types of robots into infantry formations. And they're also
• 1:50:52
  Speaker 4: working on off-road ground vehicle autonomy, similar to some of
• 1:50:55
  Speaker 4: the stuff that we're doing. We think we're pretty far
• 1:50:59
  Speaker 4: ahead right now, but they're working hard to catch up.
• 1:51:07
  Speaker 3: What kind of, is there any, what other kind of
• 1:51:11
  Speaker 3: weapon systems do you think we'll be putting on these Ultras?
• 1:51:17
  Speaker 4: I think for any of, you know, you think about
• 1:51:21
  Speaker 4: uncrewed ground vehicles and autonomous ground vehicles, you can put,
• 1:51:27
  Speaker 4: it depends on the size of the vehicle, you can
• 1:51:28
  Speaker 4: put like almost anything on them, right?
• 1:51:30
  Speaker 3: Yeah.
• 1:51:32
  Speaker 4: Any type of vehicle that you have today, which can
• 1:51:35
  Speaker 4: carry anything from a very small vehicle, might be a
• 1:51:38
  Speaker 4: couple hundred pounds, up to tens of thousands of pounds,
• 1:51:42
  Speaker 4: you can put on a potentially autonomous vehicle. So that's,
• 1:51:47
  Speaker 4: you think about smaller things, might be smaller remote weapon stations,
• 1:51:52
  Speaker 4: might just be sensors, radar, optics, things like this. But
• 1:51:58
  Speaker 4: much larger vehicles, Large missile systems, right? So ship interdiction missiles,
• 1:52:04
  Speaker 4: anti-air missiles, all sorts of things.
• 1:52:08
  Speaker 3: You could put pretty much anything on this.
• 1:52:10
  Speaker 2: Pretty much anything.
• 1:52:11
  Speaker 3: Are you familiar with Epirus?
• 1:52:12
  Speaker 2: Yes.
• 1:52:13
  Speaker 3: So you could put Leonidas mounted on one of these
• 1:52:15
  Speaker 3: damn things and then you have drone defense for an
• 1:52:21
  Speaker 3: entire fucking battalion.
• 1:52:23
  Speaker 4: Yeah, usually when we're thinking about drone defense, you know,
• 1:52:26
  Speaker 4: it's going to be a layered defense. So you're going
• 1:52:28
  Speaker 4: to want maybe something like, you know, EPRIS' system. But
• 1:52:32
  Speaker 4: you might also want kinetic air defense that might be drones, like,
• 1:52:40
  Speaker 4: you know, drone interceptors. It might be machine guns. But
• 1:52:44
  Speaker 4: you can imagine a whole set of these, you know,
• 1:52:47
  Speaker 4: many different vehicles with different defensive systems. creating that layered defense.
• 1:52:53
  Speaker 4: And we think it'll probably be something like this so that,
• 1:52:56
  Speaker 4: you know, even if drones are getting through one type
• 1:52:58
  Speaker 4: of defense, they're getting hit by another.
• 1:53:00
  Speaker 2: I mean, even if you lose.
• 1:53:02
  Speaker 4: Some of your vehicles or systems, there are more to
• 1:53:04
  Speaker 4: take their place.
• 1:53:06
  Speaker 3: Let's talk about the technical mode, the team, and the
• 1:53:09
  Speaker 3: peer threat. Sure.
• 1:53:13
  Speaker 4: So the way that we think about the technology that
• 1:53:16
  Speaker 4: we're developing, we really are thinking about it as.
• 1:53:22
  Speaker 4: Developing an autonomous version of core battlefield functions, right? So
• 1:53:26
  Speaker 4: like either the ISR or breaching as examples, where you
• 1:53:33
  Speaker 4: have multiple vehicles with payloads.
• 1:53:37
  Speaker 2: That are performing a task.
• 1:53:42
  Speaker 4: Now, in order to do any of that, in order
• 1:53:44
  Speaker 4: to have a set of autonomous vehicles with payloads that's
• 1:53:48
  Speaker 4: doing something complicated, Think about like first principles, like what
• 1:53:52
  Speaker 4: do you have to do first? Well, you have to
• 1:53:54
  Speaker 4: be able to carry stuff and move it, you know,
• 1:53:56
  Speaker 4: like move it from one place to.
• 1:53:57
  Speaker 2: Another in the environment.
• 1:54:00
  Speaker 4: And when you look at our history as a company,
• 1:54:05
  Speaker 4: we started with what we think is the hardest problem
• 1:54:08
  Speaker 4: there first, which is being able to understand terrain and
• 1:54:13
  Speaker 4: move through that terrain.
• 1:54:15
  Speaker 2: Given that basis, you can then build off of that.
• 1:54:19
  Speaker 4: So, you know, One of our strongest technical moats is
• 1:54:23
  Speaker 4: the fact that we're the best in the world at
• 1:54:26
  Speaker 4: being able to see and understand terrain and move vehicles
• 1:54:30
  Speaker 4: through it. We can put payloads pretty much anywhere that
• 1:54:34
  Speaker 4: a vehicle can drive, right? And so once you have that,
• 1:54:40
  Speaker 4: as you now combine this to say, instead of just
• 1:54:44
  Speaker 4: moving one vehicle or one type of payload, now I'm
• 1:54:47
  Speaker 4: moving multiple vehicles, multiple payloads. that foundation allows you to
• 1:54:52
  Speaker 4: build up those capabilities. So we think that that's our,
• 1:54:55
  Speaker 4: you know, one of our biggest technical moats is we
• 1:54:58
  Speaker 4: have just this incredible team coming from, you know, deep
• 1:55:02
  Speaker 4: tech places like Waymo and Cruise and, you know, self-driving
• 1:55:07
  Speaker 4: car companies, some of the top artificial intelligence labs, you know,
• 1:55:12
  Speaker 4: they're working for us focused on these problems and creating,
• 1:55:16
  Speaker 4: you know, this AI essentially that allows you to, to
• 1:55:20
  Speaker 4: uh move payloads in the environment so um that's technical
• 1:55:24
  Speaker 4: mode i can't quite remember the the other questions but
• 1:55:27
  Speaker 4: that's um really a foundation of everything.
• 1:55:33
  Speaker 3: I mean we had one thing that i didn't ask
• 1:55:36
  Speaker 3: i think we started talking about it out there and
• 1:55:38
  Speaker 3: then i said we would come back in and discuss
• 1:55:40
  Speaker 3: it because it was so damn hot out there but Yeah.
• 1:55:43
  Speaker 3: What does it look like, you know, when we're talking
• 1:55:45
  Speaker 3: about controlling all these autonomous vehicles, especially when it comes
• 1:55:49
  Speaker 3: to hundreds? I mean, and it's not remote control. So
• 1:55:55
  Speaker 3: what is it? I mean, we kind of discussed a
• 1:55:57
  Speaker 3: little bit out there, but what does it look like
• 1:56:02
  Speaker 3: for the team or the person or whatever? is controlling
• 1:56:07
  Speaker 3: hundreds of these at once?
• 1:56:09
  Speaker 2: That's a great question.
• 1:56:10
  Speaker 4: So you're not going to be just like looking through
• 1:56:12
  Speaker 4: the vehicle sensor and remote controlling it because you have
• 1:56:14
  Speaker 4: hundreds of them.
• 1:56:15
  Speaker 2: How can you do that?
• 1:56:16
  Speaker 4: So the way that we have designed the software is
• 1:56:21
  Speaker 4: that you essentially have overhead maps, something like a satellite map,
• 1:56:26
  Speaker 4: and you're able to see all of the different vehicles
• 1:56:29
  Speaker 4: on that map, all of the different autonomous assets on
• 1:56:33
  Speaker 4: the map.
• 1:56:34
  Speaker 2: So that gives you.
• 1:56:36
  Speaker 4: And you can zoom out and you can look at
• 1:56:40
  Speaker 4: terrain features. You can look at where all the vehicles are.
• 1:56:43
  Speaker 4: You can see what payloads are on the different vehicles.
• 1:56:46
  Speaker 4: And then the question is, how do you now coordinate
• 1:56:49
  Speaker 4: and control hundreds of those assets? And you need to
• 1:56:53
  Speaker 4: be able to do that by selecting vehicles, grouping vehicles,
• 1:56:57
  Speaker 4: telling groups of vehicles to go to one place or another.
• 1:57:01
  Speaker 2: Telling them to execute a.
• 1:57:02
  Speaker 4: Certain payload at a particular location so that you can
• 1:57:06
  Speaker 4: kind of quickly move between sets of vehicles and essentially
• 1:57:13
  Speaker 4: tell them where to go. Again, a lot like a
• 1:57:15
  Speaker 4: real-time strategy game. We're also working on tools.
• 1:57:18
  Speaker 3: So this looks like, this is what we were talking
• 1:57:20
  Speaker 3: about at the beginning. This looks like, sounds weird, looks
• 1:57:24
  Speaker 3: like World of Warcraft.
• 1:57:26
  Speaker 2: Yeah, yeah.
• 1:57:27
  Speaker 3: You're sending a group of things giving it a task,
• 1:57:30
  Speaker 3: and then you're on to the next thing.
• 1:57:32
  Speaker 5: Yeah.
• 1:57:33
  Speaker 4: I mean, and I think, you know, one of the
• 1:57:35
  Speaker 4: reasons why we've looked to some of these types of
• 1:57:39
  Speaker 4: games is because it's one of the few places where
• 1:57:42
  Speaker 4: humans are actually controlling something like hundreds of different assets.
• 1:57:48
  Speaker 2: StarCraft is a great example of that, and WarCraft.
• 1:57:51
  Speaker 4: But The tools that people use for grouping, for moving,
• 1:57:58
  Speaker 4: for coordinating assets in those types of games are things
• 1:58:01
  Speaker 4: we can take those ideas and apply them here as
• 1:58:05
  Speaker 4: well so that a single operator can control many, many
• 1:58:09
  Speaker 4: different assets. Another thing that we're developing is things like
• 1:58:15
  Speaker 4: AI assistance, which can help to suggest particular tasks. So
• 1:58:22
  Speaker 4: you can just say like, hey, I need to move
• 1:58:25
  Speaker 4: 10 vehicles into this area, provide reconnaissance here. And it
• 1:58:31
  Speaker 4: will tell me how those vehicles are going to move,
• 1:58:37
  Speaker 4: suggest to me solutions to these problems so that I
• 1:58:40
  Speaker 4: can make decisions faster. So this is part of essentially
• 1:58:44
  Speaker 4: using AI to increase your decision advantage. Now, we're not
• 1:58:47
  Speaker 4: saying we hand over the actual decision making.
• 1:58:51
  Speaker 2: To the AI, but the AI.
• 1:58:53
  Speaker 4: Can suggest solutions that can help you to handle, you know,
• 1:58:59
  Speaker 4: more assets at once. Wow.
• 1:59:02
  Speaker 6: Wow.
• 1:59:03
  Speaker 3: I got a hot question for you. All right. Ready?
• 1:59:09
  Speaker 3: In 1941, 353 Japanese planes came off six carriers and
• 1:59:12
  Speaker 3: hit Pearl Harbor in under two hours. They sank or
• 1:59:16
  Speaker 3: crippled all eight battleships and killed over 2,400 Americans. And
• 1:59:21
  Speaker 3: here's the part people forget. We had warnings. We'd broken
• 1:59:25
  Speaker 3: their codes and Our own ambassador flagged a possible attack
• 1:59:29
  Speaker 3: a year earlier, and Congress later said the real failure
• 1:59:34
  Speaker 3: wasn't intelligence. It was imagination. Nobody could picture it until
• 1:59:40
  Speaker 3: the harbor was on fire. Today, China's arming robot dogs,
• 1:59:44
  Speaker 3: they build 90% of the world's drones, and they just
• 1:59:48
  Speaker 3: flew a mothership not long ago that launched 100 kamikaze
• 1:59:53
  Speaker 3: drones in a single swarm. The warnings are everywhere again.
• 1:59:57
  Speaker 3: Are we sleepwalking into a robotic Pearl Harbor?
• 2:00:01
  Speaker 4: I hope not. But I think it's correct. There are
• 2:00:05
  Speaker 4: warnings everywhere. We're seeing robots used in the battlefield, again,
• 2:00:11
  Speaker 4: in Ukraine. There's both by the Ukrainians and Russians. So,
• 2:00:20
  Speaker 4: you know, We and I think other defense tech companies
• 2:00:24
  Speaker 4: are certainly paying attention to this. We are trying to
• 2:00:27
  Speaker 4: understand what capabilities our adversaries have and also ensure that
• 2:00:33
  Speaker 4: the US military has similar or better capabilities. We think
• 2:00:40
  Speaker 4: about these sorts of things all the time. It's the
• 2:00:44
  Speaker 4: new technology which is going to be defining these next conflicts.
• 2:00:49
  Speaker 2: And I think it's all of.
• 2:00:50
  Speaker 4: Our jobs to ensure that the decision makers understand that
• 2:00:56
  Speaker 4: this technology exists and what it's capable of. So I
• 2:01:02
  Speaker 4: think we are certainly aware of it.
• 2:01:05
  Speaker 2: I think we're pushing the.
• 2:01:06
  Speaker 4: Military to adopt new technology and defenses against it as
• 2:01:10
  Speaker 4: quickly as possible. But you know, there may be work
• 2:01:15
  Speaker 4: to be done. Sometimes these things aren't really real to
• 2:01:19
  Speaker 4: people until, unfortunately, they actually experience them. So our best
• 2:01:25
  Speaker 4: bet is to really be watching and taking seriously what
• 2:01:29
  Speaker 4: is happening in these conflicts and what our adversaries are developing.
• 2:01:33
  Speaker 3: I mean, do you feel the Department of War has
• 2:01:36
  Speaker 3: taken this seriously? Do they understand the capabilities that you
• 2:01:39
  Speaker 3: guys have? I mean, when you... looking at that thing
• 2:01:43
  Speaker 3: out there, what it's capable of, it's very surprising and
• 2:01:49
  Speaker 3: almost alarming to me that they, you know, 15, 15, 15,
• 2:01:54
  Speaker 3: you have a contract for 15. Why don't you have
• 2:01:56
  Speaker 3: a contract for 10,000? Yeah.
• 2:02:00
  Speaker 4: It, I mean, it, it, it should be, um, you know,
• 2:02:04
  Speaker 4: I think the department of defense is trying to, or
• 2:02:07
  Speaker 4: department of war is in, in the services are trying
• 2:02:09
  Speaker 4: to move faster, but the procurement system was really designed
• 2:02:15
  Speaker 4: for a different era and a different type of war.
• 2:02:18
  Speaker 4: And that's something that I think all of the defense
• 2:02:22
  Speaker 4: tech founders who you've had on here will.
• 2:02:26
  Speaker 2: Probably agree with, right?
• 2:02:27
  Speaker 4: And something that we're all fighting hard against. Now, we've
• 2:02:32
  Speaker 4: seen the Department of War speed things up. And like
• 2:02:38
  Speaker 4: I was saying a little bit earlier, We have actually
• 2:02:41
  Speaker 4: gone from, you know, university research to production contract, even
• 2:02:45
  Speaker 4: if it's, you know, a relatively small initial one in
• 2:02:49
  Speaker 4: like three years, which is about right, maybe a little
• 2:02:53
  Speaker 4: slow for like, you know, the tech sector, but that's
• 2:02:57
  Speaker 4: lightning fast for Department of War. And we've made use
• 2:03:03
  Speaker 4: of Defense Innovation Unit and a lot of the new tools,
• 2:03:06
  Speaker 4: AppFit contracting process and things like that, that have Come
• 2:03:11
  Speaker 4: online recently.
• 2:03:12
  Speaker 2: So I, you know, I do want.
• 2:03:15
  Speaker 4: To credit the Department of War for moving faster, but
• 2:03:18
  Speaker 4: there's still so much work to be done in order
• 2:03:21
  Speaker 4: to move at the speed that we need to move at. So,
• 2:03:25
  Speaker 4: you know, it's something I think like we're all working
• 2:03:28
  Speaker 4: on and trying to.
• 2:03:29
  Speaker 3: Are other countries better customers than our own country?
• 2:03:34
  Speaker 4: For this problem for, for like our stuff or like
• 2:03:38
  Speaker 4: their own stuff or for stuff that America, not, not,
• 2:03:41
  Speaker 4: not overland AI for things.
• 2:03:43
  Speaker 3: That America, Americans like yourself are developing here in this
• 2:03:49
  Speaker 3: country or other countries, maybe Ukraine. Yeah. I mean, I think,
• 2:03:55
  Speaker 3: you know, I've, I've, I've heard of these issues happening
• 2:03:58
  Speaker 3: and we, here we have the best that the fucking
• 2:04:01
  Speaker 3: world has to offer right here. developing groundbreaking new technology
• 2:04:07
  Speaker 3: that's going to change how war is fought.
• 2:04:11
  Speaker 8: I mean, it's such an upgrade, especially with a company
• 2:04:21
  Speaker 8: like Epper.
• 2:04:22
  Speaker 3: All these companies, man. Everybody that sat across from me
• 2:04:28
  Speaker 3: in the defense tech sector, what you guys are developing is,
• 2:04:32
  Speaker 3: It's fucking incredible.
• 2:04:33
  Speaker 8: Yeah.
• 2:04:33
  Speaker 3: You know, but I see it being utilized in Ukraine.
• 2:04:40
  Speaker 3: You know what I mean? And I see it in
• 2:04:43
  Speaker 3: a– it bothers me when we're sending all this stuff.
• 2:04:49
  Speaker 3: Maybe we aren't, but I see things similar being utilized
• 2:04:54
  Speaker 3: in Ukraine, and it just seems like our country is
• 2:04:57
  Speaker 3: not taking advantage of the talent that we have here.
• 2:05:01
  Speaker 3: it's fucking scares me. Yeah.
• 2:05:03
  Speaker 1: I mean, we, there's, there's kind of a, like we
• 2:05:07
  Speaker 1: have all this, we have all this talent.
• 2:05:09
  Speaker 2: There's things to be concerned about.
• 2:05:11
  Speaker 4: And I, and I think there's, there's things to be
• 2:05:12
  Speaker 4: hopeful about. So like on the one hand, um, you know, you're,
• 2:05:18
  Speaker 4: I think you're right. I think countries like Ukraine, I mean,
• 2:05:21
  Speaker 4: they're adopting tech as fast as they possibly can.
• 2:05:25
  Speaker 2: There's just a statistic.
• 2:05:26
  Speaker 4: I was saying something like there's been 2 million casualties in,
• 2:05:29
  Speaker 4: in that conflict.
• 2:05:30
  Speaker 3: Right.
• 2:05:31
  Speaker 4: And, about a million and a half Russian and about
• 2:05:35
  Speaker 4: half a million Ukrainian. And one of the ways that
• 2:05:39
  Speaker 4: Ukraine has been able to continue to fight and hold
• 2:05:44
  Speaker 4: off the Russians is through all of this technological innovation.
• 2:05:49
  Speaker 4: It's existential for them. They have to do this or
• 2:05:53
  Speaker 4: they will lose, right? And so they will do whatever
• 2:05:57
  Speaker 4: it takes to win here. it's not existential for the
• 2:06:02
  Speaker 4: US yet. And I think that is means that adoption
• 2:06:09
  Speaker 4: is just slow. Like people aren't feeling the, um, the
• 2:06:14
  Speaker 4: pressure to, to do it. And, and I, I think that,
• 2:06:17
  Speaker 4: you know, it's unfortunate of course, because we have a
• 2:06:21
  Speaker 4: little bit of luxury, um, right now, like we are
• 2:06:24
  Speaker 4: not in a conflict like the Ukrainians. Um, so we
• 2:06:28
  Speaker 4: have the space to be able to, potentially build this
• 2:06:33
  Speaker 4: technology and transform our forces. But we don't have the urgency, right?
• 2:06:39
  Speaker 4: And so I think that's part of the problem. On
• 2:06:43
  Speaker 4: the positive side, I do think that we have the
• 2:06:46
  Speaker 4: best minds in the world here. I think we can
• 2:06:50
  Speaker 4: do it. But again, it's a matter of- We are
• 2:06:53
  Speaker 4: doing it.
• 2:06:54
  Speaker 3: It's our government that's not fucking keeping up with you guys.
• 2:06:57
  Speaker 4: But it could be done faster, right?
• 2:06:59
  Speaker 2: Like, we could accelerate this.
• 2:07:02
  Speaker 4: And, you know, that's a fight that we just have
• 2:07:06
  Speaker 4: to continue to make, right? To continue to fight to
• 2:07:12
  Speaker 4: get this tech into the government faster. And I think
• 2:07:14
  Speaker 4: going back to, you know, thinking about some of the
• 2:07:16
  Speaker 4: DARPA stuff earlier, like, we're creating incredible new technology, but
• 2:07:21
  Speaker 4: then you really got.
• 2:07:22
  Speaker 2: To fight to, like.
• 2:07:23
  Speaker 4: actually get the Army and Marine Corps and services to actually,
• 2:07:30
  Speaker 4: you know, try it, iterate, you know, ultimately, ultimately use
• 2:07:36
  Speaker 4: it and incorporate it.
• 2:07:37
  Speaker 2: Yeah.
• 2:07:37
  Speaker 3: Yeah. Yeah. If we were attacked tomorrow, do you think
• 2:07:40
  Speaker 3: we could survive a sea or ground invasion?
• 2:07:45
  Speaker 4: I mean, I think I, I'm, I think we would
• 2:07:48
  Speaker 4: absolutely survive a sea or ground invasion. I think we would,
• 2:07:52
  Speaker 4: mobilize very quickly. I think those, the urgency would be
• 2:07:56
  Speaker 4: there and, you know, we would, we would do whatever
• 2:07:59
  Speaker 4: it takes. took to, to, to win. So I'm, I'm,
• 2:08:02
  Speaker 4: you know, I think very positive and optimistic on that.
• 2:08:07
  Speaker 4: Like I, I think when we're pressed, we can, we
• 2:08:09
  Speaker 4: can do a lot. But I also think that, you know,
• 2:08:13
  Speaker 4: we could be better prepared. So.
• 2:08:17
  Speaker 3: Me too. Last thing. manufacturing. Are you guys manufacturing these yourselves?
• 2:08:23
  Speaker 4: How fast do you scale up? So we're trying to
• 2:08:26
  Speaker 4: scale as quickly as possible. We are manufacturing the ultra vehicles.
• 2:08:32
  Speaker 4: I think I mentioned before that they're based on Polaris
• 2:08:35
  Speaker 4: engines and drivetrains and things like that, but we're upgrading
• 2:08:37
  Speaker 4: the suspension. We're putting in the payload deck, adding the compute,
• 2:08:41
  Speaker 4: the sensors and everything. We've got a factory running in
• 2:08:45
  Speaker 4: Seattle that's doing this right now, but And I think
• 2:08:48
  Speaker 4: we've 5X manufacturing over the last six months or so,
• 2:08:52
  Speaker 4: and we need to do a lot more. So we're
• 2:08:55
  Speaker 4: building up that capability as fast as possible. And again,
• 2:08:59
  Speaker 4: the more support we get from the government, the more
• 2:09:03
  Speaker 4: we actually work with warfighters, the more demand there is.
• 2:09:05
  Speaker 4: And there's a ton of demand right now. So we're
• 2:09:09
  Speaker 4: scaling as quickly as we can.
• 2:09:12
  Speaker 3: Right on. Well, Byron? I really appreciate you coming. It
• 2:09:16
  Speaker 3: was an honor to interview you.
• 2:09:17
  Speaker 6: And I.
• 2:09:19
  Speaker 3: Love everything Overland AI is doing. And that thing out
• 2:09:22
  Speaker 3: there is super impressive. Thank you.
• 2:09:24
  Speaker 4: Yeah, thank you so much for having me. I really
• 2:09:26
  Speaker 4: appreciate you taking the time to learn a little bit
• 2:09:28
  Speaker 4: more about what we're doing and inviting us out and
• 2:09:31
  Speaker 4: being able to show you some of the things that
• 2:09:33
  Speaker 4: we're building.
• 2:09:34
  Speaker 2: So thank you.
• 2:09:34
  Speaker 3: Thank you. No matter where you're watching The Sean Ryan
• 2:09:51
  Speaker 3: Show from, if you get anything out of this at all, anything,
• 2:09:55
  Speaker 3: please like, comment, and subscribe. And most importantly, share this
• 2:10:03
  Speaker 3: everywhere you possibly can. And if you're feeling extra generous,
• 2:10:08
  Speaker 3: head to Apple Podcasts and Spotify and leave us a review.