Dark Factories, AI Agents, and Mass Production for Defense | Dennis DeMeyere, Autonomous Production

Okay, Dennis, welcome. Explain the need for your company to exist. Let's look at what Ukraine's done. So Ukraine has manufactured 477 tons in just one kind of like organized group of distributed cells. Yeah, 477 tons, totaling 18.7 million objects that have already been used in defense. And that's just one of their organizations of decentralized manufacturing cells. So what's the value of the defense? It's really straightforward. You make what you need where you need it. Dennis Demier is the co-founder and CEO of Autonomous Production, a company building AI-driven manufacturing systems that aim to turn physical production into software, automating everything from design and prototyping to large-scale manufacturing. So imagine you have 40 of these in a shipping container. The shipping container has 40 different machines. Each machine can make around 30 to 40 a day. Once you do all the math, what you figure out is you take 10 shipping containers and you're making 100,000 drones a week. Okay, the machines between $400 and $2,000 a pop. So unbelievably inexpensive. Full of autonomous. You'll need to put a bunch of people in there. We're putting a battery of 3D printers on the back of every forward deployed unit that we have. All right. What is this piece of shit? That is a drone. It is a printed place drone that has all of the mechanicals printed. So there's no assembly required. Is this just a demonstration of, okay, ability or this could be a product? Oh no. We don't actually make the products. We make the things that make things. So this is something that is actually in the public domain. But it's interesting about this is propellers are printed in place. And so what you have is the ability to essentially attach a PCB, a battery, a payload, and then a very small brain on this. Of course, some module for communications, then you're done. So you have three clicks with say like an automated robotic arm and you're done. You are a founder of a company called Autonomous Production. How many other co-founders? We have one other co-founder. She came to us from Raytheon where she literally spent her entire career out of college since day one at Raytheon and then came to do this. And this is a culture shock for her. You know, it was actually really funny. She is thriving. I think yeah, and she was a superstar there. But I think when you give somebody the freedom to just like go do, there's no governors just go as fast as you possibly can. Really smart people. They seem to thrive. And I see in the Bay Area or you guys are in Ohio and California, right? Yeah, we're building our. I guess split. We're building our first dark factory in Ohio. Okay. I'm in the Bay Area. And then she's in San Diego. We have some aerospace customers. And so San Diego makes the most sense for her. Oh, okay. And now dark facility is that because it's actually dark and why? That is why they call it the dark factory. The common definition is actually seven a little over seven days. So if you want to talk about the darkest dark factory, seven days where the lights are out entirely, no humans walking indoors. And it's just printing. You know, we can do seven days, but like a largest large scale dark factory is actually in Japan based on the research that I've done. And they can go seven days about a single human walking in and out. Japan. Yeah. Nice. What does that make? Do they know what they make there? I know. I want to say it was robotics. Okay. It's a lot of infrastructure. What? Okay. So you're, but you're focused on a lot of you were talking about this before we started today, but on defense production, right? We have grown a lot in the defense sector. So it's a good question. Yeah. So at the highest level, right now, manufacturing, and there's a lot of different ways to answer this, but like in a defense context, let's look at like what Ukraine's done. Yeah. Using additive manufacturing and other like basically like CNC type machines, but primarily actually additive manufacturing 477 tons, totaling 18.7 million objects that have already been used in defense. And so what's the advantage? The advantage is there's no single place that you can strike to take out manufacturing capabilities. Okay. At the same time, you can, because you're closer to where you need to be, you're closer to the actual feeder. Not kind of make the air strikes less useful, huh? Yeah. Or to a bit, look at, you know, another country that's done this quite well, but going back all the way to like the 80s, look at Russia's oil infrastructure. It's highly decentralized by design. That's smart. So decentralized manufacturing. So what's the value to defense? The idea is, you know, you actually, it's really straightforward. Yeah. And you don't need people. What was that? Um, that Rhyme Mattel quote that was going around, making fun of Ukrainians. Do you see that thing? Like find Rhyme Mattel CEO making, yeah, because it's totally wrong. And I think you'll laugh at this. And why is pulling that up? Oh, yeah, here. This is it. It's like, he calls them housewives with 3D printers. I saw this. I'm playing with Legos. Could not be any more wrong about that, in my opinion. I think he's probably wrong here. Ukraine's been putting up an amazing fight with their strategy. Yes. What? Okay. Why do they call it added manufacturing? Is it just manufacturing? It's a great question. I think if it is manufacturing, the way that we do it is the machines were agnostic. We don't really care which hardware you use. All of ours are controlled by agents that are like actually like, stateful. So they run forever. There's a fleet of them between one and a thousand. And they're fully autonomous. So if one goes down, it can transfer its work to others. But like, why is it additive? It's actually the way that you make it. And this is really fun if you think about it. So you can see, if you look really closely, you'll see that there's little lines. And so it uses FDM, which is fusing these layers together. And so it starts at the bottom. There's your bottom. And that's obviously a part of, you know, pretty strong. Yeah. So I picked that one. Try and break it. I mean, I can't break this. It's really, really. You can probably break this. She looks strong. Try it. Really try. It's like a wishbone. I don't play like the turkey, you know. You're a play wishbone. Of course. Okay. Skipping. Yeah. You have like siblings. Yes. Okay. We that's a very competitive game. The wishbone with like the turkey. We got the big piece. But like what's your strategy? Well, you could put your right before it. Just wait for that in the pool. Yes. You torque. But that, but even though you try to break that, this is an interesting one. It's part of a lead slide, basically, which is I'm sure there's a better name for it, but growing up breaking this obviously. But this is not made with polycarbonate ABS, not polycarbonate ABS. This is actually made with the cheapest, lowest-drink material that we support. Yeah. And so it's amazingly sturdy. And this is why it's working so on the Ukraine. That'll go 140 miles an hour. Well, you had the ACS swag before the show here. You're a small, I mean, where I'm, I had to say I'm the grand twink of Austin. So like, Twink's can't break this. Maybe you could. Maybe you're a jack twink. No, I mean, what's the name for a jack twink? Jack twink. It's like an order. I'm not coming with that. Not an order. It's something like that. I don't know. But no, so additive manufacturing is you build bottom up. Right. And so that, this is interesting. So then how would you make this bottom up? You actually place it upside down. Yeah. So then it comes off, drops out. Okay. The machines between $400 and $2,000 a pop. So unbelievably inexpensive, fully autonomous. You don't need to put a bunch of people in there. The UK Army did an exercise recently that I think is really instructive. And the quote at the end of it was, we're putting a battery of 3D printers on the back of every forward deployed unit that we have. Wow. It's published. It's a lot. It's it is. Wait, so and what I guess are your customers than direct government customers or you're going to work with industry both or like what's the split right now? I mean, we do have a cage code. But ultimately, we we help the people who supply. So we're a supplier to the suppliers. Okay. Yeah. That's a nice place to be. You know, because you don't really sell into the government directly is it can be quite impossible sometimes. Oh, we did it at our last company. We did straight to Cibber 2, then Cibber 3, then IDIQ. But it took years. How long did it take to win your first Cibber? I think it took us like three years. Yeah, about three years actually. Two and a half. Yeah. From the start point. Yeah, about three years. Some people would say that that's actually fast though. So I had to tell myself we were fast. Did you do Cibber? Did you go straight to Cibber 2? I mean, we pretty much skipped all. We pretty much never did any Cibbers to be honest. We had like one, but it wasn't as I was tiny and then like our product was finished. So then they were like, well, I just won't buy the product, I guess. But like I the Cibber problem, the Cibber thing's pretty broken. Like, there's a lot of great companies that should get Cibbers, but they just can't because the Cibber mills get them, you know? Yes. Although I love the Cibber mills because they're just like, I don't know what they exactly do, but they're kind of funny. I don't know either. It's almost like the white paper mills. So there are you can't call them Cibber mills though. I get offended. But I'm not the one who said I think there's one in Ohio though. I mean, there's a couple. There's a couple of companies that got in trouble relatively recently for a lot of million dollar white papers. There's a lot of yeah. There's a lot of that happening. I kind of want to I want to acquire a Cibber mill. There's like a next step in the business. That's what I want to do with all the bullfrog revenue. I mean, I've seen all the bullfrogs. I think you guys have your hands. I want to know. We need to buy a Cibber mill so I can see what's inside of there. But I mean, but it didn't the day like isn't the D.I.U. trying to fix a lot of of what you would call like the E. - I think there's a lot of attempts to make things easier, 'cause the cyber is intent is good, right? - Well, I think that the cyber mills, I don't actually know, at my suspect, there's a lot of congressional pressure with them in their states, because these are 300 to 500 million year revenue businesses. If they, and certain congressmen have invested interest in keeping the cyber mills, because their states, that's a lot of jobs in revenue and stuff in R&D. And then some people will defend cyber mills, proponents of the cyber mill, and say that's where a lot of R&D happens. Like, for other companies to pick from. So I don't know. I can see both sides of it, but-- - I did it the old way, like actually submit your own cyber. I understand that there's something like a cyber mill, but we actually had to go through the whole process, which took a lot of time. - Well, I mean, some companies are, what we would call an aspiring cyber mill, that don't have any cyber yet, but want to mill them. (laughs) For the listeners, what I mean by mill is, just the industry jargon for a company that exists only on cyber dollars. And it's kind of a funny, whoever came up with cyber mills, pretty funny, I want to know who created that. - It's a great name. But I mean, so we just need to turn the world. - There's a cyber mill up here, here we go. Submit excessive proposals. - I still can't believe that this is off the thing. - Leverages, shell companies, lacks real R&D capacity, focuses on recidivism. - The focus of sound recidivism is savage. - How do they make it through the procurement process, which is actually pretty intensive? - I, the DOW. - Well, they never reach procurement. - Oh, I guess it's true. - They're just R&D and just Sivers. - I don't know though. - Even somebody's Sivers is a lot of work. - This is kind of the dirty underbelly of defense R&D, but I even think that this is not as bad of a problem as like kind of how RDT&E works across just broadly. I don't know if you, how much experience do you have with that, but like, I think the DOW, someone from the DOW one showed me like a chart, it's like 97% of all R&D, dollars goes to like the seven major primes. Like if you add Anderl and SpaceX, then it goes like 96% of all, like, those primes at this point. - They're not primes, I'm saying like, if you add like SpaceX and Anderl and you count them too, then it's only like the major legacy primes are only getting 96%. Which means they're only getting like two, like even these new generational companies are getting like very few R&D dollars. - But that's changing. - That's changing. - Like right now that's changing. - Yeah, like we're in the midst of the change, but it has to change fast enough, you know. - There's this saying, so I left Google to go start this company before this one and I worked for this amazing CTO there. He's like the CTO of all of it. And he came out of defense before he went to Google and he told me one thing, he goes, "If you're going to go work with the DOD, here's the rule of thumb, though the world's greatest customer, if you can survive long enough for them to become your customer." And that's because of that process, that process is so slow. And so it's almost like it was designed to avoid startup like innovation in timelines. And when the idea of, we have finite cash, you're not cash will positive, it's not so negative. And so that timeline is everything. - Yeah, our fundraising strategy was like, we need to have three years of capital before we get a customer. And that was just part of the strategy. I knew that because I started in the Navy. So I knew the timeline going in. Other founders, I try to tell them when they start, like you need like three years of cash. - Yes. - And where you're going to find that is not from the customer. - That's great. Or, you know, like our approach was more be mixed, right? So we support like very large companies, so like regulated industry. - Yeah. - And then also defense. - Which is why you see a lot of the dual use mantra from the government, 'cause the government is kind of saying like, "Hey, you better find your early revenue somewhere else, 'cause you know, it's gonna take some time here." - I think I'll be, I'll push back on you a little bit and say I think it's getting better. - No, no, it is getting better. - Like they're definitely trying. - I think it's, I think the trend is in the correct direction. - And even like the drone competition, I'm not sure how close you can follow that. - I've been, I follow that. - That actually pulled up the winners. - Have you followed the winners? - Yeah, we work with them. - We went of them. - Okay, see. - You got to say it, but. - That's okay. - That's okay. - I think we're working with more than just one. But yeah, one of them right now. - Well, there's this Ukrainian winner. Obviously, Neroz has been on the podcast twice three times. But the skycutter and Ernie and Hurtov, but they're cool. - There's a couple of here that are actually really cool. But yeah. - And what they end up needing is a lot of polymer parts. - Of course. - And so I don't wanna like, chill what we do, but at the end of the day, we need to take some. - You're on here to chill. - No, really, it's like it's more, we take a machine that can produce plastic parts out of like polycarbonate ABS. So like really strong plastic parts. Way stronger. - This is different material than that, right? - Those are both the same material. - No, this is just like, the parameterization on that is very different. - Yeah, okay. - And so, we take those machines, and we'll make them run for seven days straight, just chunking out parts. So like that container that's producing parts, for example, that'll produce, if you put 10 of those containers together, 100,000 drones a week. - Okay. - So a lot of these companies are scaling up, and they're scaling up really fast. - Yeah. - And one of the problems that they're having is, so what do you know this? How long is it taking you to set up manufacturing? Like real scaled manufacturing capacity? - I mean, it's still, it's a process that takes a while. - I talked about a couple of these companies. One of them we're putting machines in, I think this week, the process is 12 to 18 months, typically if you want to do large mass manufacturing. - Yeah. - In the polymer space, even if you're gonna use somebody else, it's gonna be three to nine months most likely. And then, you know, for us, it's like a couple of days, we should come in there. We're a startup, so we'll just, we'll drop them in, be like, here you go. - So would you say that you can really help speed up FPV production? - We can, yeah, that. We can also do it for, you know, like OBD readers for cars, like we do it for, - Well, let's talk, - The polymers are a third of what we make in the US. - But let's talk about FPV production, 'cause obviously there's an FPV, there's four brushless motors, you know, there's another one, there's a big one. - Let's pull this big dead one over here, for the people on YouTube. - And now it's time to show a little bit, please, because like, you guys shot this one, right? - Okay, yeah, we shot this one. But this is a big FPV that I'm holding here, and obviously you can 3D print this stuff, right? - Absolutely. - And what about this like circuit in the middle? - So it comes down to, if you're talking about a PCB, you can 3D print the PCB now. - Okay. - And that's something that you've inspired to do. - The wire wire to the arm. - We're working on it right now, actually. They make the same form factor. - Okay. - And it all runs, it's Cartesian G code. And so if it's Cartesian G code, that's what we do. We also do the parameterization. But I mean, just like, pick this part for like one second. So like, say you're gonna like 3D print this - Yeah. - out of the chassis. - Listen to this. So like the ways that you can do polymer production, it's a 211 factorial. That's the number of different settings you have to configure to primaries. So just to give you some. - And that's more stars than. - Yeah, it's 10 to the 356. So to get that right is one of our models. - Okay. - Then to go through the process of printing it. So it's actual generative robotic code. That's another model. And you have to have this testing sim, right? Because test engineers exist. They have a job for a reason. And then you need a physics model and a material sciences model. Because after we print it, we eject it. And we can print this next to that, next to these. And then eject them all that will write it from places. - But these propellers are injection molded, right? You can usually tell me looking at it. - Yeah, probably. - Or you think they could be 3D printed? - No. - But you could 3D print propellers. - Absolutely. - Okay. So what we're really saying, and then the wiring, you just have to have a human install. - And there's really gonna be no difference if you and Neil post printing the carburetor ABS at the end of the day. So like, yeah, you could actually do all this. The wiring, so there's two types of drones, right? There's drones that you want to come home and then there's drones that are trotable. - Yeah, right. And so a trotable, you can just print. You can print the circuit. Now you're still going to need motor, can't print that right now. Although there's a group at MIT that's going really close, but you can't print that. So you gotta snap that one in, you can use a robot, I can't snap that in. You need a brain for it, right? And then you're gonna need a communications module. That can also be snapped in. And so you're getting close to the point of being able to make it all. - You're getting closer than I had thought, probably. It's closer than most have thought. - I have thought, basically, you could just do the body. And you know what's interesting though is like the throughput. It's when you start to think differently about all of this. - Yeah. - And you go like, all right, so one machine that costs $400 can make 40 to 50 chassis per day. It can pass it on to another machine that can print the PCB. You're vertically integrated. And the amount of capacity you have, it starts to look a little bit like how Ukraine is doing what they're doing. - Yeah, and you snap the brushless motors on. You don't print those yet, right? - So the windings, do you need, if it's a tritable, do you need to have four brushless motors, or do you need to have a motor that's centralized with different drivetrains, and then you have the propellers already built in that don't need to be snapped on. - Interesting. - It's a different way to think about it. - Yeah, but that's how people are thinking about it, right? - I always thought it was a lot of creativity in the space right now. - Yeah, I can tell. I didn't even really think about it. And I run a drone podcast. So what the fuck do I know? - Well, yeah, but you do something different with drones. - Well, we're trying to shoot on, but I mean, I'm trying the podcasts, we're trying to wrap them, you know? - Shoot, shoot. - I'm the, obviously, this is the greatest drone podcast in the world, so. - I would agree. - We should probably learn about what's going on here. Okay, so there's different ways to think about it. I never even really thought that you could do this without four brushless motors, but I guess that makes sense. - It's possible. - So you could probably also print the whole fixed wing chassis, right? - They're doing that already. - Yeah, they're doing that already. - Yeah, but it's a different printer, right? So those are the ones where our premise is a little bit different than to a firestorm. - Okay, yeah. - So you've been on firestorms podcast? - I haven't, but I know. - He stole my podcast name. I used to be called Drone Wars. I heard. Did you ever work that out? - Yeah, we were beefing, we were beefing for a while, but we're actually boys now. - Just check it. - That's my boy. - I just don't know how like, or brush the wound is. - We're absolute boys. - But they use like very, very, they also do, you know, they do manufacturing, they do it in a decentralized way. They use like very, and they're exceptional, but they use it. - I've toured their little setup. - Yeah, it's cool. - But, it's cool. - There's the HP printers. - Yeah, so like, you know, those are a little bit more expensive. You're talking 10s or 10s or 10s. - They are expensive. - Thousands, whereas our premise is actually based more on the idea of, - And there's different material too. - Yes, different process as well. - Yeah. - They can use centering, there's a bunch of different ways that you can do it. ours is primarily FDM, but it's based on this premise, it's just different. People will tell you it's crazy, but it's totally possible. - Yeah. - Which is, instead of using a $40,000 or a $200,000 machine, you can use $400 machines. If you know how to do it and produce something that's going to be very close, the only part about it that would be slowly different would be it may not have fully isotropic properties. - So you're going like super cheap. - Yeah, it's based on like, we can roughly deploy production capacity 37 times faster than a traditional factory. - Yeah, that's really cheap. - And fast. - Fast. - And so how do you do it with such cheap machines though? - Well, like hold on, so let's like, let's like look at cheap for a second. - Yeah. - So is it cheap? - Is it cheap? So they're using this on nuclear submarines today. And they're going from part costs that were $317,000, like $300,000. They're using this in the army today. They're using this in the Air Force. In fact, FDM printing in the Air Force and they're doing it for flight parts. And the box that they were using, they actually published it. It's not a screech Air Force base. It was like a $800 bit of hardware. And so like, what, and here's what happened. So the stuff that, you know, like you had in your lab back there, that stuff hadn't really changed in like 10, 12 years. And then about two years ago, it changed. And then about four months ago, it really changed. Like really changed on masks. Like, software or something? - No, the hardware. New manufacturing techniques that came out of Asia basically took the cost down from something that would have cost you 80 to $150,000 and put it into a $2,000 box. It's wild. And now there are US companies. - Coming out of Asia though. How, how like behind are we? - Asia. - Asia in any factoring. Pretty far behind. But here's, I'll tell you, like I'll tell you my version, my vision of how I see this all coming together, because we're part of a bigger group that's trying to build like a Shenzhen in the Midwest, specifically in Ohio. - Shenzhen in like Columbus? - Yeah. - Like where the arsenal one is or. - Yeah. - And you can be like the same kind of like the same investors behind all of that, our investors as well. - Okay. - And it's more than just that right. So it's path robotics, Gecko's out there. There's a lot of like, you know, Gecko, D-O-W Supply. - One of the Gecko founders lives near me in DC. - Really? - He just moved there for a little. He doesn't look permanently. - He makes sense. - He's given what they do. - He's a great golfer. - Is he? - Well, he's hilarious. He used to play college golfer something he said, but I went with him once and I just, he just basically lines up way right, 'cause he like, he has like a weird hook, but he's like, he's so good though. He's just like, by the way, he starts to aim. He's like aiming in the other hole. - All right, so then you're, how good are you? I'm pretty good. - Like I'm like, okay. - That's the problem. - My handicap's 14, so not good. - Pretty good. - Like I used to be better, but I'm working on the ACS now. - That's my excuse. - Like how do you find the time? - I don't right now. So I did, but in between jobs I did, between startups. I was golfing like after I saw my last company before this one. - How long did you have? - I had like nine months off. - I had two. - Oh, that's all? Did I want it more? - I got the itch. Like it came back and I needed to do this. - No, I was happy. I was actually content, but Luke Allen kept calling me every day to start this shit. And how do you feel about it? - That's the interview. - I was actually a genius idea. It was a genius idea by him. So I'm feeling great, a great job. That makes me look good, but at the time I was, you know, trying to enjoy myself and golf a little. 'Cause startups are pretty brutal. - They're all consuming. - They're just all consuming and just really hard. And like the competition's hard, the customers hard, the employees are hard. - But it's hard. - Everything's hard. - Right, like it's a true meritocracy. - It is, it is. - But anyway, just like the idea of the, like the idea of the ultimate game. - It is. - It's a pure escape. - The ultimate game. - But the idea is, you know, some manufacturing coming back to the you anywhere, not just the US. As different geographies start to invest in it again, there's this question. And what you'll find is, oh, we're just gonna have a lot of humanoid robots. So we're gonna take this like factory prototype that we've been using since like the 1800s. And like the big thing is gonna be, we're gonna replace the humans with robots. And keep the factory the same. - That's really dumb, though. - It doesn't make any sense. And so the way that we're looking at it is that the factories become data centers and the robots are formed. So the form is based on the function. What are you making? And then you're gonna fully automate it. And so you go, okay, so you're gonna fully automate these dark factories. What about the jobs? Here's a fun number for you. 40% of all skilled technicians in manufacturing are over the age of 55. So the jobs, there's just not enough people right now. In a lot of cases they do the jobs. - Well, that's gonna be a ton of jobs. - And they'll be far more production. And so those people, yeah, they'll actually, it's Jove and Sparadox. We'll end up having more people. - Yeah, I'm not go over such big things. - Yeah, I'm never concerned about automation, taking jobs. I think that's like a weird thing that has never happened. - But everyone's, you know, people are-- - This time it's different argument never. I mean, any time, any technology's been made, more jobs have been made. - Yes. - I don't really know. But maybe this time it is different in a wrong. - No, no, no, no. - Yeah, it's different. - Hey, I, you know, it's special. It's never special. Like it's kind of like the same. - Yeah, isn't it fun to be living in the one time it's different? Think about that. I mean, this time it's finally different after millennia. - In a way it is, right? In a way it is. - Well, it's certainly the things that are the same are that we are making technological progress, just like we were in the 1980s and the 1810s. And you know, it's still, there's always technological progress being made. But I guess what's different in this time is it's with computers. - I mean, I think just to like be controversial for a second, I think it's different because it's coming after some white color jobs. - Yeah, but I'm very worried about that. And in the past, traditionally, you know, it wasn't so much there. And so now it looks like it might be, but that's, there'll be new jobs that are created. That's your point. And that's, I agree with. And that's also how we see the manufacturing, but the manufacturing as it comes into new places, what makes it so great and you ask like, how advanced does it overseas? When you go there, you kind of like a really good feel for it. It's the proximity. - Yeah. - Everything is close to everything else. So your plastics are close to your metals. And in the way you get the metal noise, your final and automated assembly, you walk two more blocks, you're right there. And so that proximity matters. And that's one version. And then there's another version where, like you create, for example, you go are also fully decentralized. And so like small units of production make a tremendous amount of sense too. - Yeah, I mean, even on the, but on the jobs thing, like I do think that we could automate every manufacturing job with stuff you do in every white color job. And I still submit to you that humans will find more work because we're beast of burden. We can't not work. We could literally create a utopian society where no one has to work. And we would still end up creating jobs for ourselves because people are like, I need to do something. - Yes. - So I guarantee, I always believe in humans' ability to make up jobs for themselves. That's where I stand. - I think you're right. - They're just gonna make up something. They'll find something in the moon. They'll start going up to the moon and like doing jobs up there. Some shit. - I mean, I'll give you like some data on this because I find, to me, I find it fascinating. I'm not a huge fan of training new AI models in labs. And then get why they do it. And I understand that that's very popular right now. So we did ours. I started by putting 10 locations up across the US, Ohio, Florida, and in California. And it picked places that were very hard to do what we're doing, which was at the time we started with polymer production. We're not doing many forms of robotics. - Okay. - But polymer production and humidity don't get long. So Columbus in the summer, Florida, Southern California, the Venice Beach Boardwalk. Anyway, and they were all small in mid-size, kind of like manufacturers. And in every case, in every case, the business didn't contract and lay people off. And every single one of the 10 that we were in, they ended up expanding. - Of course. - 'Cause they were making more money than margins went up and then they hired and they grew. And so it's to your point. I think what ends up happening is the automation just makes everybody more efficient. That one you can bend. That one you might even do. - That's not great. - No, I could break this one for sure. - Yeah, yeah. - Yeah, that's a thing. Like, I think jobs, people, I think when they talk about job loss, it's really just transient stuff. Like, you know, there could be a shift of where the jobs go. And that transient is what people are afraid of, I think, 'cause what they really mean to say. Like, 'cause in the transient, you have some like, tough times. - Yeah. - That everybody can go back to college and retool. - Yeah. - Whatever that version that looks like. Not everyone's at the age where their situation were like in a Ford to do it. - I mean, at 39, I don't really want to learn anything new. At this point. Like, at some point you get a little old, like, a little like, "Ah, what's it for me?" - See, I, one of the reasons why I did another one was I love learning. - See, I do the podcast because I was like, I could just talk about what I already know. But I can't gain, but I guess I learned what the guest is. - I was gonna say, I watch you ask questions, learn all the time. - Yeah, I guess I got it, but I got it's the retention. It's the retention that goes down. - That does go down a little bit. But, you know, so I think, you know, in general, what we're uploading that to the cloud. - Yeah. - So, the retention, hopefully. - Maybe. - I'm reading like everyone's different model. - But I think, you know, what happens is, manufacturing, just the ability to make things. - Yeah. - That becomes a national priority. And honestly, it's becoming, we've already seen it in the current. conflict, it's becoming a contention that we have to deal with. It is a literal constraint on what we can do. And so the idea that we can start making things again, like really making things again, it's going to look different. It won't be the way that it's being done in other parts of the world that are very advanced. It'll be better because what we can do is we can take the best versions of hardware and then combine them with what we do, which is software. >> Yeah. >> And like RAI software is unique. >> Yeah, these will like, you know, calibration systems that, if I'm parts that, yeah, pictures. >> Yeah, they're interesting. You put like a server on it. >> Yeah. >> You think about targeting systems that are incredibly low cost. This is how they're doing it right now in other conflicts. And it's really exciting. It's working. >> So you're a permable? >> You're bullish. >> I'm perm-optimist. >> Perm-optimist. >> Because I think the alternative just, it's kind of miserable as a way to live. And I'd rather be optimistic about things because, you know, you're going to see, you're going to see the good, you're going to see the opportunity. You're going to see like something. >> Sorry. Are we going to open the straight up Hormuz or what? >> I was looking at a sensitivity analysis model on that last night. And they're modeling these different scenarios and then the impact of GDP. >> Okay. >> I really think it would be good for the world's economy. >> If we open it. >> Well, what do you think? >> Let's do a polymarket, thanks. >> This is still the good. >> This is sponsored by Polymarket. The world's greatest prediction market, not Koushi. >> Actually, what's going to be better? Which has more market share? >> Didn't want to just get in trouble. >> Wait, what's the, wait, what are the odds on Polymarket though? >> Polymarket's website is so bad. I mean, just have, it's going to rewrite it by end of April. >> They're wet. >> They're 18%. >> Oh, wait, should Hormuz return to normal? 18%. That's actually not pretty good odds, right? No, like, seems like it's staying shut. I guess what is normal? You have to always check in Polymarket what like the rules are, you know? It's normal. Yes, if IMF Portwatch poses a 7-day moving average transit calls, all right, it's fair. Because there's two sides of this. There's the infrastructure, and that takes many, many years to rebuild. But so far outside of LNG, the damage there is somewhat minimal. It's still important, but it's just somewhat minimal. And then there's transit. And that's where it's becoming very disruptive. And we've already deployed the strategic reserves, right? Yeah, so the thing here, I'm not a geopolitical expert, but this is obviously top of mind for everyone that listens to the ultimatum. You know, this is getting tense. >> Little too tense. >> I think it applies to everybody. Even if you don't, you know, you're not following defense or you're just ignoring the news. This is going to impact all economies, right? >> For sure. >> Yeah. And so it applies to everyone, whether or not you're following it. >> Yeah, I haven't seen it. I mean, I'm trying to think back to, I guess I was a little young when the first golf or desert storm, but I haven't seen it this intense. Was it this intense during the Iraq invasion? >> I am old enough. Does anyone have a few years and you think- >> Yeah, what's it feel like? >> I mean, remembering that was more like you got to sit in front of the TV and watch it like first person style. >> Yeah. >> And that was the first time I think that it ever happened, right? >> Yeah. >> But I was also pretty young at that point. This is to me about as intense as it's gotten in my lifetime. >> This seems pretty intense. >> So this was about a month old, but it was in China. They have micro drowns like, you know, about this size or smaller. 200, that a single soldier can control. But what was interesting about it was 200 controlled by one person and they moved in non-deterministic patterns. So the way that they moved, you couldn't anticipate and track. And that is the idea of everybody's got 200 little robots around them. It feels almost iron man-esque. >> That's cool. >> Like someone around them? >> Yeah, that's what it was like. If they have a video of it working and you start to go, okay, then you're right. Every tank, every vehicle needs some version of the turns against this. >> It's really interesting. I feel like the war in Iran, I've at least been falling on X. It's like, it's really fun. I don't know if you ever played this game as a kid or anything. Metal Gear Solid 2. But the Japanese, I feel like really, really called it like this. Okay, so this game, well, not drones though. This game Metal Gear Solid 2, the premise of it is that information control becomes like the most important modalities of war. >> Yeah, there's so much misinformation on X. And this was prior to social media. They came up with this concept. There's so far ahead. But the whole game is about- >> Never those, yeah. >> Yeah. >> So what's your social strategy? Where do you spend your time on social media? >> Well, X. >> Obviously outside of the podcast. >> X for investors. >> Well, for myself, well, so for posting, we post on X on my account. And then on LinkedIn, on my account. Then we have ACS accounts. That's our social strategy for the company. And then where I consume content, I mean, I shouldn't do this, but- >> That's why I was asking. >> I consume it on X and it's just like if you watch the Iran stuff on X, it's just- most of it's fake. It's an information warfare. So like you're just getting fed a ton of misinformation. And like I spot, I can spot a lot of it, but even I can get tricked. I mean, it's easy to get tricked by it. >> I don't think anybody can spot it all. If it was easy to spot Nikita or somebody over there would find a way to get it off. >> No, I mean, I'm reading- I keep seeing attacks that one, and then I have to go into the comments and check if it's fake. Like you have to like actually- To use X properly now you have to like go read the comments to see if it's fake or not. >> It's hard. I think the only way that you can even use it anymore is you know the person. >> Yeah. >> You know the person personally who's posting. >> Right. >> The internet is ruined. The internet is almost ruined now. Like as soon as it's going to be all bots, undetectable bots. >> Or Reddit's the same way. Like how much of that is actually just bots arguing with bots? >> I mean, that's how it started, right? Didn't Reddit start? Like they just were posting themselves with different accounts. >> And did you come out? >> Yeah. >> But ultimately, yeah, I think- >> Oh yeah, here's the wired article. We went back to that wired article. I really like- Doesn't- I heard Trace buying wired now. Did you guys see that? >> No. >> Yeah, so basically they're mad at wired because of their Anderl hit pieces. Which I think that Anderl's in the right here. I mean, wired's- I don't really know if wired did a great job on some of these articles, but- >> I haven't seen the hit pieces. >> Basically they wrote some ridiculous pieces about Anderl's like internal failures, which is crazy because it's not like that at all. But then I heard something funny that Tray was gonna buy wired, which I think is classic tech billionaires being like, okay, I'll just buy this whole magazine. It's happened before. >> Yeah, it's a T-R-A-E. Here, look, I support- there's a Palmer Lockheed tweet. This is actually great. And then they go on and say he's gonna talk about buying it or something. But I think that's- I support this. Because wired actually has a kid. I used to read that. I agree with them. I got the magazine delivered to the house. >> Yeah, you used to be really good. I think that's what people see with media these days, though, and tech, the media versus tech world, you know, it's always contentious, so- >> But it's always been contentious. I just wonder if there is a form of media where you don't actually get to see a human talking that you go out full trust that this hasn't been AI generated in some way. But so you're right, there's information, right? And then there's also just quite frankly, like technology and manufacturing. And that's what's changed. And whether it's like regulated big industry, like, you know, building your own things in your country and take the macroeconomic situation out of it for a second, just, you know, what's happening in the world right now is people are starting to go, all right, domestically, wherever you are, we need more manufacturing production. And then what we're seeing in defense, it's not terribly different. And the technology is changing so fast that nobody can keep up. And that's not just, you know, hardware used to be kind of removed from it. Not anymore. Yeah. It's not like what you're doing. No, hardware is all part of it. It's all robotics is just the new thing. And it's all these countries, I mean, the cost of air striking has gone down. Like, look at Iran, they're, they're able to air strike still. And like they're not like a world superpower, but they're regional. Yes, they're causing havoc. And the cost of air defense has gone way up because it's like so hard to stop these things. It's almost like they've had decades to think about. We wanted to cause asymmetric havoc. Yeah. If we were attacked, how would we do it? How crazy is that there's like a straight of water that's like, oh, that transports 20% of the world's energy. And it's just like right in the worst spot for us. Oh, and it's also got mountains on. It's like, it's like, it's, it's like, you know, it's like, if the world's a board game of risk, like, obviously like America's in a sweet spot. We have two oceans and a lot of resources, but we also have this annoying thing. This massive dependency. I mean, it's just, it's like you couldn't arrow like it's like you couldn't have drawn the board of the map any better of the board game for Iran, like the, like where they lay in the board game. Is that kind of nuts? It's like, that's like territory you definitely would want to take. Well, what's interesting is, you know, so what happens when it's closed? Well, all of these large ships, they start sitting just outside of that area, waiting for safe passage, but safe passage is, it may or may not be available. And so then what eventually ends up happening is what is happening now, which is they start coning down production because there's no way to get really live map right now. Looks like it. I don't know if like, do you think there's just like ships waiting on both sides? They're just like sitting there. I mean, that's what it looks like right there. What are they doing? Are they just like hanging out playing backgammon, like bridge on the ship when I was on the submarine? What did you do when you were on a sub just waiting? Play a lot of spades. All right. We're like video. Like, could you watch video like, what are their movies? Yeah. Well, military guys love movie. There's always tons of movies and deployments and stuff. What's like when you were doing it, I'm just curious. What was the movie that was? Well, favorite movies of military guys are like point break. It's like a classic. I love. I mean, how do you know running man? I don't know. I mean, break. Run for red October. Yeah. Yeah. It's like a movie. Like, so you're right. sitting there, but because they're just sitting there, there's no transit. Because there's no transit, the oil fees start pulling down production. Yeah. And so it's, but why is that the case? It's kind of asymmetric again, right? It's a small area. Yeah. You can operate in very small cells. And yeah, it's very, very expensive to chase down these very small, relatively inexpensive cells, yeah, defense capacity. Yeah, it's just so interesting like watching this conflict play out right now, with just how much information warfare there is. That's really the thing that surprised me the most, but just it's just so much that you cannot tell. It's the fog of war. Where do you get your truth from? I just really kind of don't. Like I'm meaning like I kind of just think high level about the issue and try to, I think the, I think New York Times or Wall Street Journal, like, even if they have like a liberal bias or a Republican bias, I think that the speed at which they report is slow enough that they're trying to do more facts. You know what I mean? So like if you go to the front page of the Wall Street Journal, they have a process. Well, though to speak, here's the timeline of the last 48 hours and they'll list the things that happened. Yes. And like I can, you can kind of fall back and trust that, I think. I mean, I don't know where they're getting their information from, but isn't it interesting that we have a live map up that will tell you at least logistically? And they, you know, you can get access to information in real time. Yeah, you can. There's like live maps and stuff. Yeah. But even then, isn't this just a distribution format? So like if you think about it, well, I think because of no one giving them that information and that's probably part of a negotiation process. Yeah. Right. And so it's hard to, it's hard to know. Yeah, I don't know. I mean, if you've ever read Chomsky's manufacturing consent. No. Okay. It's like a, it's like a famous book, but basically the premise is that the way the headlines are written, it's like always to manufacture consent. Like the way they spin the headlines, you know, manufacturing population. Because that one New York, any like talks about New York times and they're Wall Street Journal, MSNBC, CBS. Like just the way they write stuff. Yeah. You know, like they'll call something a regime. We'll call it a regime versus like a, like a, and a government. Like we'll use the word, you know, instead of using government, we'll say regime stuff like that. Anyway, he's like a left wing, like more like, like, yes, like Russian type guy. I don't know. No, he's like more like socialist, but he's a very interesting book. Yeah. And, uh, yeah, basically that's the premise of it. And that's his famous work. And if you can, if you read, if you read media, like mainstream media with like a, like a scrutinist eye, you can see it everywhere. Yeah, I tried to, I'm like you, I tried to skip past the headline. I also want to get to like the primary data. So wherever like the source of truth is, it's usually like in paragraph eight of an art, like a story. I could do like, I just, I'm like naturally trying to scroll past like, who has time for that? All the narrative stuff. Yeah. And just like down to like what happened in the facts. I still think it'd be great if somebody could just had, so it's like a great startup idea for someone out there, like just make the app that has literally just sourced, like double sourced facts and nothing else. No editorial. Just tell us exactly what's happening. The thing is like even the facts get disputed, you know, like what? Yeah. I mean, you can, you can bend sources. You can selectively choose sources. Yeah. So anyways, I'm really interested in the world, the changing world of information warfare. Like I know, I don't work on that. Obviously we work on robotics. But I think it's, I think it's, I think it's up my area either. But yeah, two robotics guys talking about information warfare, like we know something. I mean, it's softer for a while at Google, but I mean, I'm finding hardware to be a lot more fun. Atoms are just really challenging. Bits are simple. Atoms are hard. And you're doing this right now. And then like, yeah, actually doing real things with atoms at scale. Well, with my VC had on and solve problem. My V, yeah, my VC hat on mom and investor. I like, I prefer this world for entrepreneurs. I think that the world of software is changing so fast that just to be a pure software company is quite hard right now. Like, because you could, you're one open AI or cloud code work really so away from getting just smashed out. And I think that's just risky. So that's why I mean, we kind of saw that early, like when we started this and we're like, let's get more robotics because like robotics is just a, it's a much bigger mode. If you build something that's meaningful, like if it takes years, like when I see start up say like, oh, you went from like drawing to production in 71 days and it's like some robotic product. I'm like, that's not a moat. Like 70, if you can do it 71 days, someone else could do it 71 days. It's amazing. But like, yeah, I don't, I don't tell, I actually don't even think short time. I think it's actually a cut. It's like more of a contrarian take, but I don't think short timelines in robotics is a good thing. Like I think if you're building a serious robot, it takes a, it takes time. Building like, yeah, a new serious robot. Yeah, absolutely. I think there's only two modes left anymore. And even the second is questionable data. Yeah, data, data is a moat. Hardware is a moat. Yeah, that's why like Google is in such a good position. YouTube, I mean, they have so much data. They do. Like, even if they suck at making LLMs for a while, they just have so much data. But like people will ask you like, what's your moat? And it's like, well, think of it this way. Open AI with their lead and their amount of capital didn't have a moat. Who has a moat? Yeah, it's either hardware or it's unique data. Like to do what we're doing on the material science and stuff, I got to unplug anything. But the data doesn't exist to model this. Like if you want to break adhesion on something like this, imagine, you know, this comes out of a nozzle at 225 degrees Celsius. It hits a plate that 65 degrees Celsius. The adhesion is heavy. Yeah. Like you got a crank on it as a person. So how are you going to use this highly sensitive device to break the adhesion on this? Physics of material science. Where do you get the data for it? Well, you have a, we have a person in Virginia and a battery of machines running 24 or seven, trying different combinatorial, different combinatorial versions, permutations of those two, like kind of factors, plus different geometries. And then you get a force meter and you're like measuring Newton force required. And then you start to figure things out using physics like, okay, so then if I can create a five degree delta in temperature, I can reduce the Newton force required by 75%. It gets that's physics. That's the science. Well, that's a fair. The job is to say like you're trying to, there's certain things you can't make plastics do. There's certain things you can't make electrons do. And you're trying to figure out the boundaries there and create, it's more like art. Like robotics has like art. It is a little bit. It's like art. There's science and physics back. But like there's like, you're making 800, I always say it. We're making 800 to 1000 little design decisions in our design. It's not like one aha moment. Same with you guys. Like, there's, there's a, the process is the mode. It's never done. Yeah. It's the thing. It's, it's never done. Yeah, especially in the spaces in the domains of this large, like yours. Yeah. It's a continuous. This will be never done. Yeah. It's a continuous space, meaning it's endless and you can always get smaller, right? So in that, in that world, your work is never done. There is new sensors, everything, just always. It's just constantly iterating. Speaking of LLM modes, though, I do think that I've been using Cloud co-work and open a chat, chattch ebt a lot. And I, chat, ebt, I noticed I use more for when I need it to know about me more. Because it like, it's memory. It's memory is a mode like, because I've talked to it the most. So it has like full context of my past converse. So if I'm like, Hey, should I, the question if I'm like, Hey, should I rebrand the name of the podcast? Yes. It will go and give me a full essay on who I am as a person and why I should or should not do it. Because it, like it, and it makes very good points because it knows all my previous chats of like kind of like the vibe I'm going for. Yeah, it actually knows who you are. So like, it's pretty good. And so Cloud co-work, I've just started using co-work for work tasks like, build me a list of 50 supply chain VPs. And it's so good at that. And then I'm like, and now write an email to them all and find their emails on hunter.io. Or you know, whatever. So like, but that's one of like enterprise workflow mode versus, I think, chat, ebt, is kind of getting like a consumer data mode. I mean, that's definitely the case, right? So open it up. And I think the enterprise workflow modes probably the better one. They're like 80 20 ones, like, you know, heavily API based, right? And then the other one's going to be like heavily B to C based. And that's just like the data on it. What's interesting is even what you're talking about the memory, have you seen the posts recently about ways that you can kind of like prompt hack it to give you everything that it knows about you so that you can then give that to another model to read that. Even that mode goes away. Yes. I have also been finding I've asked it. I go, pretend you know nothing about me, comma. And then I write because I want like a fresh, yes, slate. So there's that too. Do you ever do the, so like what one of my favorite hacks that our teams are doing right now is you think you know what you want to ask. So then you type that out and then you say at the end, I'm pretty sure I've done a really bad job of asking this question. Can you rewrite it the right way and make it design its own prompt and the prompt that puts together is an order of magnitude more effective than the one that you did. And so it's just, you know, take my bad thinking, turn it into your thinking and then go do the task. Have you done, like, anything where you have a separate machine set up and that's now your personal assistant, if you will. And it's constantly working around the clock. Yes. And now you can, there's a good way to isolate that too. Because now there's a new product. I forget what it's called, but you can interact with an open laptop in your house from text message. You got it on here at dispatch. Yes. You can like text. It's great. It's so good. So the way that we run our robots is actually the same way. That's called the computer using agent. So a full OS, C U A. And we had to build one like seven, eight months ago, which sounds insane. And in reality, and it's all, you know, we could put this in here because we can see the CLI timestamped. We actually had it working like seven months ago. Probably should have stopped there and made that the startup. But it did. But so you'd be like, well, why would you build that? The robots that we're controlling don't have APIs. Yeah, you're just talking to them. They have old school. They're like industrial robotics. They have old school windows apps. And you basically have LLM interacting with them. We had to train our own. And then you have to create a, it's a framework, you might call it a harness now, but it's the thing that moves the mouse and the keyboard. So it actually like runs and then it's got its own VM fully unattended and there's up to a thousand of these at a time one for each robot It's a real agent like a BS agent and it's literally clicking in a Windows app. How does it learn to make PCBs? What do you mean like I don't know how do you try to you how do your engineers like get it to like oh just the polymer side is been so hard We're just getting started. So the polymer side, but it's possible. It's inx. Oh, we already we already do it. Yeah, just but I mean It's like 10 to the 356 meaning like it's just it's non-determinism or nothing Yeah, right like you're just not gonna throw enough compute at linear algorithms for all of this But the physics and the material sciences that stuff is the same way. Yeah, the models don't exist. So you can't use wrappers You have to use your own the only thing that we can actually get rid of is That follow that full operating system once the new stuff gets good enough Well, ditch ours, which is based on our framework I take the use case of like even just like putting a chip on to the PCB like how do you train it to do that? That's hard like that's pretty hard. Yeah, it's like how small the resistors are yeah I mean at the end of the day if it's like pick in place for example or even was that even possible you think? It's possible. How do they do it today? It's great at the at the big fab shop PCB fab shops. I've never been in a big PCB But the ones that they're doing today where you're actually like How do they surface mount components on PCB is that scale? Yeah, it's just like ask chat GBT Or someone it's highly comfortable Like how do they surface mount components on PCBs at scale? Let's see I have to SMT surface mount technology It's a Pickin pickin place. Okay, reflow soldering of familiar with Pickin place is generally like G code Cartesian robot. Oh, yeah, these machines here. Oh, yeah, yeah, these like that like all PCB Machine it's like a conveyor belt And I'm like you want to see something these are crazy But like now think about how much smaller they're getting if you want to see something wild if you look up 3d printing PCB Now look at the size of the machines that you were just looking at and then if they show one of the pieces a hardware I don't know if they're gonna show like one of the actual printers. They are tiny But they can print this material. Yeah, they get yeah, they can do the traces very local copper traces very low complexity. Yeah, yeah Low complexity meaning like two layer boards. Yeah, couple layers. Yeah, couple layers. That's it That's getting better and better and so like but the idea would be it's kind of like injection molding for making plastics the presses are tons I mean like literally tons are massive Never gonna move them. They're just huge And then the idea that you could Make something that has the exact same material properties or close very close proximate material properties on a device that you can pick up with one hand and You can beat the unit economics. Yeah of the injection molder Everything just changed forever. Do you think it's a trillion dollar market? That's crazy beating the union economic some injection molding Be crazy. I mean, it's that's happening. That's what you're doing That's what we just you're deploying capital for the fact that All right, so we had me that term. She had we had me the wire and fell took a seat Do you to your point though? It's like I say you're to prove it out. I mean, so do you think everything's 3d printed in the future? I know there's always gonna be a volume where our injection molding will have better unity economics Okay, so whether it's like a you know a million it's three million or it's 10 million plus it'll always be there You know very high high volume stuff. What about like my base motor assembly? Like this big steel thing I showed you earlier. I think you have a couple polymer parts on this the size really matters to So if you want to do like really high throughput So here's here's a thing about manufacturing the holy grail of manufacturing and I'm ripping off you on us But the holy grail of manufacturing is the cost of production becomes roughly equivalent to the cost of the inputs Okay at volume and so that's you to do that for us we had to automate four different human roles But if you can as you become more and more autonomous you have to go like all the way up the chain of the engineers themselves that are working on these things Yeah, as you automate all of those then yes the cost of production Become roughly equivalent to the cost of the inputs and at that point so like we can only do it for polymers right now We're working on the next couple of categories, but yeah at that point a lot makes sense now it gets to a size thing So 3d printers that are really fast it can chunk out parts roughly 350 cubic millimeters So about this big this big about this big Those are like little factories that used to cost thousands and now cost a couple hundred bucks and they'll do it They'll do it You can 3d print steel right you could also 3d print something the size of this table and you can get something that will print large Format for a couple thousand. I only saw the 3d printing that the whole the ship hall Believe it or not. Is that company nearby? I don't know that was cool though. I just saw one on the on the expressway all the way here really yes Is it was really cool? I was like oh wow like yeah, I told the Uber driver I go I Think that's the one that I just saw in a video like I don't know It might have been seronic. Oh, the one on the video I think was like in the lab or something wasn't a company That could have been that but but you might have said seronics But they just raised today now it's nine nine billion post money 9.25. I was just winged the bell Oh by the way I tried to get you a gong and also a cowbell because I saw on the interview with Patrick You're like we really need a gong Tommy's drums in case your order and does not have them nor does any music store within 10 miles of here Oh, yeah, that was going to bring you was a gong you could ring the gong Wow, I guess that brings gifts you bought donuts though. I did bring donuts But I'm no guest ever brought us to give I'm sending you a gong because you do need a gong. Yeah, okay Like I know which one I want to get you to a gong will be sweet. It's awesome. We're gonna be smashing that gong right here Wait a week. It'll be your office Yeah, I think that like the other sales team. I've been on like a sales board. They're gone. It's fine It's everybody should have one. Yeah, I'm surprised you can't get one in a place like Austin Yeah, I tried how how I feel like a music store should have them though. You do not Tommy's drums Guitar center. I literally called five places and I went to one place that does like crystals and like meditation stuff Which was really we're in a gong shortage. Oh, I thought they might have one for like the meditation stuff. Yeah, oh like a sound bath Yeah, nope. Oh man. I've ever done a sound bath before. There's our fun. Yeah It's really relaxing is it yeah, okay, wait, I'm nervous. I got nervous. I get I was gonna ask Nars during the sound bath. Yeah, I just felt like what are you what are they expecting of me like I'm just sitting here and I'm Be hearing this Okay, so how do you do your sales? What do you sell seem right now? We have eight engineers. We've got about 20 companies that we're working with You see a word of mouth. We have no marketing our website Is a hilarious one sentence website with the sign up form we haven't had any problem with people coming to us The space that we're operating in like Ben Horowitz gave this great talk and I'll Say this in 30 seconds both my favorite thing told all the speed run teams Which is like the injuries in Horowitz by combinator given this big talk. It's in October It's this past October you can look up on YouTube it goes you guys can do anything you can do anything But there's one thing I would recommend that you don't do 3d printing you goes it's clearly the future but the material sciences complexity Like the outpaces Moore's law and so that took us about a year and it was really hard, right? So you ignored Ben Horowitz the goat. Yeah, he's amazing. They're all amazing But like it's you know, I just I enjoy really really hard problems You know my favorite speed run company is the one where you can gamble on your credit card debt Have you seen that one what no? It's so funny. So basically the premise of this the premise is basically if you have credit card debt You want to pay it down and you can bet to pay it down and get more credit card debt if you lose the bet though That's a real speed run company. Oh, there it is covered. Yeah, this is so good. So instead of paying down your debt You're gonna bet on your debt. This is this is so good Oh, this is great wait Yeah, you see you have some big you have this little way hold on scroll up. You got this big Louis Vuitton purchase there You need to pay that down better gamble What are you doing spending for granted Louis Vuitton if you're already buried in credit card debt and then you're gambling on your debt itself Yeah, yeah, actually insane You know, so I feel like sometimes we're going into like a full degenerate society like betting on polymarket betting on Sports betting on narcredit card debt. It's betting on if there's gonna be a nuclear strike like we're gonna we're going to hell What do they say like easy easy times Not great hard man that went to that saying yeah, yeah, it's kind of like gambling times create fucked up society But it's also been like about like a 20 year bull run basically with a couple of small bumps in the meantime So it's been this time of abundance, and I think what we're gonna see now is that everything changes and except the boron You know like most equity or asset holders are wealthy well the disparities the boron has been missed by most people You know honestly that it is the very unpopular thing But we started with small manufacturers small mid size ones and the reason why was yeah trying to like help out little guy I beat the idea behind this and also these materials by the way That's really interesting about this and so I'm all about like the abundance side of this I've and truly so like you're like my a permable. I don't know if I am, but I'm really about you know about it Yeah, I've got two little kids and I think about this now a lot more than I did before I had kids These materials are all made The inputs are from American corn Specifically corn sugars and unlike you know like if you're toothbrush if you throw that away It's gonna be in a landfill 400 years later. It's like micro plastics, you know roughly the size of like joking a total or whatever Yeah, and a hundred years if this is in a landfill this turns to minerals Okay, so now we also do you know hardcore polycarbonate ABS that you know good Neil and his toxic fumes Like it would be support the full spectrum, but you know in a lot of cases you can get away if you think about the strength of something like that It's still yeah, it's a consumable right like That's this is funny strong. Yeah, exactly and you might get you know relative to like aluminum You might get 95% of the life But if you're using like a polycarbonate ABS you don't get the deformation when you have something like very very jarring and high pressure Yeah over and over and over again Whereas polycarbonate ABS comes back to its original shape much better than metals ever well Let's see you have a two little kids How one one your old three-year-old and you have two wives next to two rings there. That's how two rings. Oh Or two or I don't know if you're a Mormon or if it was like a thing. No, I Not but two wives two wives two kids. Is that how they do it? No, I have no idea. I just thought it was fun I'd be interesting if they did that's the that's not the whoop though. It's the or ring Yeah, it's gotten really good enough. I'll stick out the apple watch I like having all the sensors if you know this if you start up to work Long hours. Yeah, you have to stay on top of your health and You know more data is better and I'm looking at you now. It's like well I'm pale. I'm here's like happy. Just like taped on Wow, I just I have a raw dog life bro. That's amazing. Well, I and smoke a lot of nicotine What are you gonna plane like do you do you broad dog planes? Actually, that's a thing now you know that right yeah, no just sitting there for five hours and staring ahead like this honest No, so I'm like very narcoleptic on planes. I instantly sleep like I just go. Oh, that's a gift. It's a gift It's a gift. That's amazing. I can't stay awake on a plane. I barely ever see takeoff I really ever make it to the takeoff. Oh my gosh It's a super power. You could be on the Avengers. I know. Yeah, it's like my what's he do you fall asleep? That's my party trick. It just passes out on planes. You're like the only founder. I've ever heard of that It's really convenient easily. It's really convenient for traveling because we travel a lot of founders So it's time travel. It's time travel is what it is actually yeah, I make up better for it Popcorn long. All right. Get out of here. Row at the popcorn. Long. This is what I have to Apparently it's rare and irreversible That's cool. I mean, that's you know the irreversible part. Just give it time. Honestly I'm gonna make so much money with this like we're gonna pay our way out of that right? There's a care for all this shit. That's what I think actually, you know Yeah, medicines another really interesting area. We've gone down talk about going on some tangents, but yeah medicine cheap I mean those are yeah, that's like two bullfrogs sales No problem you guys are crushed so I feel like But here's the thing so like go back your abundance thing right so like take it so take it all back home and you go all right Are we going to be in a place about abundance or is it going to become like a really weird idiotic or see type thing and And you know as the so you think about health health gets better cheaper faster We can start to like build things maybe we've even built them like a something more sustainable way Maybe defense becomes a bit more autonomous and we have the idea of a little bit less human life at risk like maybe Maybe there is a case to be optimistic. Yeah, seriously like maybe there is I think with all this abundance and all this materials We're making like we're definitely gonna have like structures on around orbiting earth that we live in and space and like on the moon and I don't know if it'll be like our lifetime because it might take like 20 30 more years I don't know maybe that's still our lifetime, but it depends on that lung condition You've got depends on my popcorn lung and honestly yeah, oh other host of ish health issues, you know I don't even have I'm no tracking. I don't actually I don't where watches either because my wrists are too effeminate I like to say Just get a smaller one like yeah, that's a little nice. Yeah, all right. Yeah, we're small wrist bros Small wrist bros. Yeah, we are. Yeah, that's good. I can't rock a big one. It looks very uncomfortable big watches on Visually uncomfortable for others. Yeah, big watches on guys like us look weird. No, they just don't work. Yeah, I try I would honestly I used to have what I tried to for a but I was lose that shit I just always lose that like sunglasses and watches. I don't deserve to have were you a watch guy? No, I know you know I'm talking about I know people really into it. I know people that is like that's their career is that buying cell watches I know a guy who sold the startup and that's like what he does now just like I mean there wasn't yeah Just buy a collection cells just for fun. There wasn't an reason back to startup that I knew a founder that did this in New York It's like a luxury goods but like mostly watches It was like This is so where my brother startup is like I forget the name of it He was trying to look it up, but there's I think there's a couple of them So it's kind of hard to find and I don't think it did well But there was like there was a couple that we're doing this kind of thing So you have no data though. That's the point like you don't track your health at all I did the I'm pre-new though. Have you heard of pre-new? Oh of course skin right I did that right when I started ACS and I had a clean bill of health And then I was like now I'm gonna start chain smoking because I was clear at that point. Wow So I figured I was smart. So you're like a self-study. It's like an end of line How bad is this we're about to find out? I was perfectly clean and then I'll like I'm gonna start ripping this evap Do you double down? Do you do Pernovu and the Grail test? I'm afraid to go back to no I haven't but I should I'm afraid to go back to Pernovu now after Three years of smoking. I'm like no So this is how far it's gone and then we got then I won't take him around like the health I like this but like So I did the Grail test recently. It's my second time doing it now what they do They sent somebody out here. Oh, yeah, it's a lot of gallery. Sorry. Yeah, Grail is like that thing like the company that makes gallery So they send somebody out and they look for like a hundred different types of cancer in your blood and it's like you know less than a grand But this is how convenient they've made it. I'm lung cancer, can I I don't know all of them like I should do this So which is this cost this will here's the crazy part so it's not it's less than a thousand dollars Which is a lot of money to people but Here's what they've done recently. It's just my second time doing it you go listen I'm really busy. They know like we have to schedule it They send someone to your house and I'm like I'm not my house in that the office pretty early every day. They had someone in my house at 5 30 in the morning That's awesome isn't that amazing they walk right in your living room They pop droop blood and they were out five minutes like I check for throat cancer That's a bit threatened long around like all of your vices coming in at once. I'm gonna check this I might get this test now But you know you think about that right? They so to make that cost come down by an order of magnitude just one order of magnitude Anybody can afford to get cancer screens every single year yeah, well, they do this in China right don't they do this now? Do they I think they just started doing this yeah, I never got a cancer test in China Well, then I'm saying the Asian I've heard that the Asian markets are really into self-diagnostic like they really is that it makes total sense Which I think the America's becoming more into it now too like because as as the products when it's that Convenient yeah when they will show up at your house at 5 30 in the morning and they had two people there By the way, it was like a team and we should do a perk for ACS with the where if you're over 35 we pay like once a year for Cancer screening and then it's I don't think under 35 should get this perk though because it's a little too young Oh my gosh at our company we actually give people we give people a certain amount of like thousands of dollars a month instead of Telling you like you have to use our bad care. Yeah, we give you the opportunity to go Here's exactly and it's enough to get like the Cadillac plan. Yeah For you and your family or you could be very Alternative and how you approach it and then that way we're like libertarian and as much as it's like here's the capital that you would have gotten from us Now we don't have anybody that's the managed the benefits and everything else and now you have enough to buy whatever you think you need like go Go hard. Yeah, you know, we've got one person who like use parts of it parts of it for you know health care but they're the part for Dedicated private yogi that they go and see a couple times a week. It's like all right You know, yeah, yeah, it's your choice right like your choice. That's good perk. Yeah Like it but it's just a different way to think about it and the advantage for startups and you know I see you here at Benro and your office You all are doing some quirky things for sure Which is you can kind of go from first principles? There's no bureaucracy. Yeah, there's no like corporate overlord. What's the right way to do something? Just like you know you've worked at a bunch of places. What's the right way to do it? Yeah, so that's how we think about it. I like to think that I'm the corporate overlord, but I'm a benevolent Right, yeah, yeah, probably listen to people well, I've mostly just Like I just sit in this room all the High-casts I hide meetings come in here. No windows cameras. There's no windows. There's people. There's like a whole team behind Yeah, so where do you see the as we wrap? Where do you see the um like drone market going in the US then like obviously There's a bunch of drone companies. It's exploding. Yeah, it's exploding, but like we're working with a number of how does one of them break out and win? Well, manufacturing it comes down to volume. Everyone's got the same series of constraints and they're different than the constraints that you have It takes months if not a year and a half to stand up real manufacturing capacity. Yeah, it's also a lot of capital upfront Some of these drone companies, you know, they're newer as I've been through some of their shot like they're newer companies. Yeah And so it's how quickly can you get actual volume now? There's this Capital war you think it's a little bit of capital But actually I don't think it's capital because all the capital in the world can only make manufacturing move at a certain speed. Yeah, I know that so I feel that yeah, right? Yeah, you rolled class problem, but like so you know our view which is kind of selfish is you know whoever learns how to produce and manufacture quickly The other challenge that you have as a drone manufacturer is it isn't going to be one design. It doesn't gonna be one type Yeah, it's gonna be changing constantly So if you're doing like for example injection molding which is what you're probably using for parts of the bullfrog You might guess I don't I don't want to mostly CNC and cast all right, but for your polymers. Oh Like that this crying assembly. Yeah, you've got a couple of polymer parts. That's three printed actually is it? Yeah, that's so that's smart because like that means every time you have a better idea You can change it if you were injection molding which is what a lot of drone companies are doing hard to change the line on that Yeah, well, I mean it's a literal metal mold. Yeah, it's the hardened that steel Yeah, then you have to like get time on someone's machine Gotta load the thing in I mean it it's months. Yeah, so you just can't adapt I've watched the induction injection mold factory go off for making this fresco machines. It's cool really cool Yeah, it's fun. Yeah going one of those in a little it just I mean parts just come out so fast. It's press right like yeah the cycle times are really fast It's the whole thing is fascinating, but I do think it's it's who can first like yeah Obviously like the best approach in the best tack and then volumes of really big problems that we have. And what you don't want to do is commit your capital to like, you know, one million or two million units only to realize there's a slightly better way to do it. And so I genuinely believe in this idea of like what we're doing, what Path Robotics is doing, and a couple other companies where it's like truly the prim--the original premise of, you know, it's like a factory. It's probably a data center. Like you give it a workload and then it changes and it makes it in the full things autonomous and whether it's like metal strip path or it's polymers through us. So a year from now you're going to have how many factories up and running. We're doing--we have two different ways that we're doing this. So one's going to be a factory because there's some companies that just don't want to actually make their own parts. Yeah. There's like some auto companies that we're working with that would prefer to be in someone else's dark factory. But then if you're like, "I-TAR regulated for example," and you really would prefer to have everything within the form. You're going to help set us up. Yeah, you just--by the way, we don't--the hardware you can choose, we actually work with it all because it's just an AI agent. And then the robotic instructions, yeah, but otherwise just pick what you want. But it gets better and cheaper every year. And then materials are getting unbelievably inexpensive. Yeah. Yeah. So like the whole world's changing right now. And I think it's going to be really abundant. And there's going to be ecosystems. So if you need something like this made and you're like, "All right, I need glass. For my optics, I've got a bunch of chips that I need. I need-- Sorry, if you can't see us on camera. And you that, then I need connectivity. So I need like cellular modules. Then I'm going to need metal and it's going to have to be like heavy. I'm black and like heavy, heavy. And then I'm in polymers and you want them all then finally assembled. There will be an ecosystem. There's one being stood up right now in Ohio that we're a part of that will just take you and all the way through it. And that's--you know, we're going to do that. And then it'll also be fully decentralized too. So we support both. So we'll have one factory for sure. I would say, Dennis, we appreciate you trying to build in next American Shenzhen in Ohio. Great state. And yeah, we've got to have you back on in the year to check progress. And we appreciate you for what you're doing to actually try to stop the asymmetric warfare that's occurring. I have family members who serve. And so the idea that these are getting out there, it actually makes a real difference. They're going out there. And it's crazy to appreciate it. It's crazy times right now. Thanks for coming on. Thank you for having me on. Really? Yeah.