Blake Resnick, Ian Cinnamon, Zach Weinberg, Jen Bucci, Paul Mikesell, Rylan Hamilton, Zach Long, Formula One

35:13So I've been doing that ever since I was pretty small. But really what got me thinking about public safety technology was the October 1 tree. So I grew up in Las Vegas. I knew people that were along the strip when that was happening. And that's what got me thinking, like, maybe there's a home for some of the technology that I love in the hands of first responders to help them save lives during critical incidents. So I went and I reached out to Vegas Metro SWAT and surprisingly, they agreed to have lunch with like 18 year old me at the time. And then interaction, you know, we talked a lot about what happened during October one and I learned things that I didn't realize, but then the conversation broadened to the rest of their jobs, you know, the high risk warrant searches and barricades and hostage rescue missions that they have to face every day.

36:00And, you know, I walked away from that interaction thinking if they just had a way to get eyes and ears in dangerous places, that that would be a life-saving capability for them regularly in the context of active shooter response, but also all of the other missions that I mentioned. And while I also conceptualized the idea of like a citywide nine-way response drone network in that meeting, I was familiar with how technically complex it would be to pull something like that off from my time at DJI, because they were my prior employer before I started Brink. Just the levels of reliability and airspace deep infliction and optics and like many other things that you have to pull off in order to really make that work.

36:41It felt out of scope for me at the time. So I began with the indoor system. which is really like a purpose-built drone for SWAT teams to get eyes and ears in dangerous places. We invented the world's first drone glass breaker. So the drone can actually fly up to a window, shatter it out, make entry. Then it has a LIDAR on board. So as it's clearing rooms, it's drawing a floor plan of that structure and then streaming it back live to first responders. It has thermal imaging capabilities, 4K cameras, and also a two-way audio system. So when the drone finds someone, a crisis negotiator can actually use it like a flying cell phone to try to de-escalate the situation.

37:16And today, over 600 SWAT teams are actively using it, which is about 10 to 15 % of the SWAT teams in the United States. And that success is really what enabled us to take on this larger vision of automating a lot of 9.1 call response. How do the teams test with the product? Do they have traditional sort of, I imagine they can go to ranges and do scenario planning and shoot houses, right? Are they going and using your tech at the same shoot houses and practicing? Or do you build sort of digital systems that allow them to practice as well? What does that look like? Both. Yeah, we have simulators.

37:55So if folks want to fire drones virtually, that's totally possible. But many of them do choose to train with the actual hardware. And I would say like when I was developing the first version of Lemur, I actually wrote along with Vegas Metro SWAT for like six months. So I went on 20, 30 SWAT call outs and I watched them deploy the early versions of our drone. And more than in any other way, that's how the technology was actually developed. Like me paying close attention to what worked and what didn't work. And if something didn't work, I would go back to my mom's house at the time and re-engineer the product to the point where I thought I would do better on the next mission.

38:30And then when I would get a 3 a.m. notification, there's a barricade at this address in Vegas. I'd bring that version of the drone. Can you talk about just how hard it is to deploy hardware in the field in these environments? We've talked about a company called Sonos that I'm sure you've interacted with at different points. We bring this up sometimes because they're in these highly controlled environments, which is just your home, and they connect to your local internet. and you just like want them to be able to like you want to be able to hit play and just like have it play music yet 20 of the time it doesn't play and i i bring that up as an example because it's like it's not like they're not talented hardware hardware and it's like that is the perfect example of just how hard hardware is even in a hyper controlled environment where no lives are on the line yeah and then you go into a situation you know that that you're dealing with with brink all the time where like you need this stuff needs to work because first responders are relying on it you know the nine nine one one operators are relying on it you know so much is kind of riding on it so i'm curious how you think about sort of battle testing products um internally and at what point you feel like they're ready to be deployed in the field i mean the reality is it's just incredibly difficult yeah um i'll tell you i mean in the very early days of brink when i was riding along with the Zwa team, I just faced countless issues.

39:54A large portion of this happened during Las Vegas' summer, so overheating was a theme in the first couple of months. The RF environments where you're trying to operate drones can, in some cases, be incredibly congested, especially in ISM vans. Think about an apartment complex, right? Every apartment has its own Wi-Fi transmission system. And you're trying to share Spectrum with all of that, which is terrible. In some cases, public safety radios will blow out your communications. You have to be aware of that and sort of like integrate with all of the other radios that a SWAT team might already be utilizing.

40:29And then everyone wants pilot assistance on these drones and obstacle avoidance. But many of the techniques that you would normally want to employ in order to give a drone indoor localization are sensitive to environmental factors like visual inertial odometry utilizing an hd image sensor right like maybe that works super well when you're not blowing up shit loads of dust all the time um in the attic right or maybe it works well in like a well-lit room but you you start kicking up insulation in zero light conditions and like good luck making that function so what what you ultimately have to invest in is large numbers of redundant systems right localizing based on lidar data and imu data and vio data and gps data etc barometer data etc and then if you layer enough systems and you're clever enough in your system design and software design and everything else you can get to something that works but i mean this is this is what makes all of this hard right like putting together a quick demo of a product that competes with us you know that's pretty easy yeah yeah yeah that actually actually does better than what we make in the field is extraordinarily difficult i mean i know and it's it's funny because i i imagine that that there's companies that that have or will try to compete with you that fly a dji drone and they're like oh i could make an american version of this that you could fly in a house They make it look so easy.

42:01Totally. I mean, DJI is an extraordinary company. It was really amazing to spend some time within their org. Can you break down some of the design trade-offs from what people might have experienced with a DJI quadcopter? I think about like 20-minute battery life being kind of the upper bound. But I've heard about drones being gas-powered, being single rotor. There's so many different trade-offs, octocopters. There's all sorts of different, what Zipline's doing. We talked to the Zipline folks yesterday. It's a completely different system. So what trade-offs have you made to move your product away from what people might be familiar with, with like a DJI Mavic 3 or something like that?

42:45Totally. Well, the first thing I would say is DJI is actively getting banned throughout the United States. Yeah. So on the order of 10 states have already passed laws preventing their public safety agencies from utilizing Chinese drones. yep dji drones are chinese autel drones are chinese yep dji has something like 90 of the global market autels roughly five percent wow so like 95 of the market um is banned is controlled by companies that can't sell into public safety agencies in 10 states which definitely is uh generating a significant need for more drone makers in the free world right uh there's also potential potentially some federal legislation coming that would have a similar effect on all 50 states.

43:28Since it's... Go for it. Sorry. Yeah. But no, I mean, if we're looking at the actual products, Lemur2 competes most directly with DJI Avada. But we invented a whole bunch of features that make Lemur2 very fit for purpose for a SWAT mission. Things like a glass breaking attachment. So the drone can actually make its own hole into a structure and make entry because it is fairly dangerous for SWAT operators to go up, rake and break a window when there's a person with like a rifle inside. No one wants to do that. So our customers care a lot about that feature. Lemur2 has a two-way audio system. So when our drone finds someone, you know, it can actually be used to communicate with them, which is usually the number one goal besides getting someone out of a structure during a SWAT callout.

44:12Like you want to establish negotiation and a line of communication. If you do that, the risk of something going terribly wrong goes down tremendously. So it's a pretty common tactic. Our drones have thermal imagers. Our indoor drone has thermal imager dji does not offer an indoor system with the thermometer that can see heat which is unbelievably beneficial on basically every mission i could go on and on and on we've invented a million features along these lines how do you uh how do you think about uh supply chains obviously i'm sure it's top of mind this week how are you thinking about it how are you adapting brink i'm sure you've been um you know trying to become independent from china it takes time we are I've been personally sanctioned by China.

44:55I cannot enter China. Brink has now been sanctioned by China twice. So we're not allowed to buy any parts in China. Wow. How are you dealing with that? I feel like we just heard a story about Skydio not being able to get batteries. Isn't it hilariously ironic that we can't pass federal legislation to ban DJI, which owns 90 % of our market, yet they're banning our team? You don't even sell there. Our team? Yeah. Right? I mean... Same old story though. Same old story, right? Like all of their social media platforms can come into our market and compete, but we can't go into theirs. The exact same dynamic is now unfolding in the drone industry.

45:32It's ridiculous. So yeah, supply chain. Like how are you thinking about it? Because it feels like just even if you're fully aligned with Made in America, it's hard to get a reliable source of batteries in America. and we just saw that story with Skydio having their battery manufacturer get pulled kind of out from under them. Well, our drones are NDA compliant, meaning that we don't utilize any electronics from China whatsoever. No processors, no radios, no imaging sensors, like none of that stuff. We had to become NDA compliant a couple of years ago in order to sell to the customers that we sell.

46:09So that was like a great step one. Got it. listen, I mean, you just, you have to set this as like an initial design requirement, like engineering team, we can't use parts from China. Yeah. Can't do it. Yep. And if that's a part of the original brief, then people find solutions. Sure. Yeah. They work with new suppliers, they in-house stuff, like whatever's necessary, they do it and you end up with a product. I think what's excruciating is building something without that initial requirement set And then having to like transition it over. That seems really like really bad. But we just we've not been in a position where we've had to do that because of NDA compliance requirements and other things.

46:50We saw this one coming. So it is very much. Do you think there's the adequate investment in in the domestic drone design and manufacturing today? Or do we need to know? I think it's I think it's off by like a factor of 100. It's actually it's actively terrifying, actually. We have no drone industrial capacity. I think the DOD continues to underinvest in small drones. They're only now beginning to realize that these things are important weapons of war because of the conflict in Ukraine, but they're not writing checks. I mean, the largest military small drone purchase that I'm aware of is SRR, and it's just not very big.

47:30It's like, I think, single digit thousands of systems. Those are the orders that have been placed so far. So when we're buying a couple thousand drones at a time and DJI is producing literally millions a year, it's not a good situation for America. How do you think about focus in that context? Because in my view, if we need more investment, one of the ways to do that would be you launch a DOD-focused side of the business or you start to work with foreign allies and stuff like that. But clearly, you've made a decision in the fullness of time to focus on the local market. But I'm sure you have pressure all the time.

48:10I'm sure you have investors that like send you an article about Ukraine and they're like, Hey, Blake, did you see what's happening in Ukraine? And you have to be like, yes, I saw. But you know, we have our own sort of battles at home. Listen, if military customers aren't buying drones, that's not a decision I can make. Yeah. Fascinating. What about in the supply chain? I could also say, though, just to be clear, like I absolutely adore the public safety market. Like I started this company to serve first responders. I think the life-saving mission that they take on is extraordinary. And it's something that I'm deeply passionate about and excited to support.

48:47So that's what I care about here. But I mean, I'm not blind to what's going on in China or Ukraine. And it's clear that systems very similar to the ones that we build will shape future conflict. um is it is it uh i'm interested in and again maybe getting a little bit out of scope here but i'm interested uh there's this classic idea of like you know u.s uh marine uh training with some type of like ar based rifle and they spend all this time training on this sort of rifle uh are they spending enough time training how to fly fpv drones how to work with drone networks or do you think the training is not even sort of caught up to the realities of the battlefield today?

49:32Well, you know, FPV drones are really hard to fly. I don't know. Have you either of you guys tried to fly like an FPV system before? Yeah, it's really hard. I've just done basic DJI type. They're not, they're not stabilized. Like they don't have GPS. So if you take your hands off the sticks, FPV drones will fly into a wall at like 40 miles an hour. It takes incredible skill to like even keep them in a hover like let alone hits a moving target somewhere um they've been they've been highly successful in ukraine but i don't i don't feel like fpv drones as they're currently designed or manufactured are really like the end state of this technology i think they'll be more automated yeah they'll have stabilization systems you'll be able to select a target and the drone will do its own navigation to intercept that target uh they won't rely necessarily on RF communication systems since the jamming environment is getting so bad in, you know, in many places.

50:28So that's all to say, like, you can say, okay, we should be training our military on how to fly FPVs. But like, that's going to be a huge amount of work. And all of that work might not be very relevant in five years when some of the technologies that it's described or capabilities I described mature to the point that they're commonplace. So I don't know, I think really the right choice would be invest a lot more in purchasing drones, actually buy. Oh, Zach, Zach, we were flipping people over. We have some technical difficulties in the studio today. No worries. We'll be with you in just a minute, Zach.

51:08We got to say goodbye to Blake. No, I guess just to finish my last statement. I think the move here is for the US military to buy drones at significant scale, start incorporating them into inter-armed forces, get people familiar with them, but have drone manufacturers like us or anyone else build something that's actually fit for purpose. So all of those efforts result in a good outcome for all parties involved. That makes sense. Well, this has been fascinating. We'd love to have you back again soon. And thank you for the work that you're doing. And congrats to you and the team on the new milestone.

51:43congratulations thank you there we go massive round cheers buddy cheers talk to you soon bye and it seems like zach's already here so let's just bring him in he's out okay we got to bring him back in uh where we are building the plane as we're flying it and we're also being attacked by state actors so the state actors if you see if you see an air take it up with the state actors because they're the ones that are uh hacking us at every at every turn at every turn uh but we're very excited to have zach weinberg on the show you probably know him from his uh uh repeated appearance on the logan bartlett show formerly cartoon avatars if you're a real logan bartlett head uh zach was almost the co-host he denied it he said i i'm not i'm not really the co-host of that show but we know that so many great episodes came out of that era.

52:35He was debating the crypto folks again and again and again, weeks on end, fighting the fight during the height of Zerp. And now he's moved on to fighting Keith Raboy about tariffs. And we will get into all that. And then I'm sure we'll talk about bio and what he's up to at Curie. So let's bring Zach in if he's here. Okay, he's coming back in. We will send him a message i love his uh x banner it's a george carlin quote never underestimate the power of stupid people in large groups he loves it which uh sometimes is what it feels like on x yeah he's been on a tear this has been a good week for him he's had plenty to comment on and dunk on and uh had a lot of fun chatting i really want to get him to break down his debate with anthony pompliano as well as what he's been talking to Keith Roy and Logan Bartlett about.

53:36But let's bring him back in the studio when we get a chance. And in the meantime, let's tell you about Numeral. Sales tax on autopilot. You can spend less than five minutes per month on sales tax compliance. Go to numeralhq.com. Just do it. Thousands of companies work with Numeral. You should be one of them. And what was the stat? over 75 states now collect tax on? I think it was a couple hundred. Couple hundred. Actually, no, I have it out here. 25 states. 25 ,000 states. Soon, if we have our imperial way, it could be hundreds of states. Yes. The moon's going to be a state and they're going to have sales tax on the moon for sure.

54:17Greenland? Oh, you think you can ship to Greenland without paying sales tax? Not a chance. Yeah, right. We're not letting those penguins free ride on us. they're going to be paying sales tax and you're going to be paying your sales tax on numeral with numeral hq go to numeral hq.com check it out uh in the meantime palmer lucky fired back at folks hating on colossal talking trash about uh how oh colossal didn't create dire wolves technically they just spliced 10 000 year old year old dire wolf dna with gray wolves uh they should spend their time arguing that jurassic park did not in fact depict any dinosaurs and uh palmer shares uh This thing sucks, actually.

54:57And it's literally the coolest thing ever. And I agree. Very excited that the dire wolves are back, that we got hairy, big, gray wolves. Very funny. Who knows where Colossal goes? Maybe it's just more entertaining zoos, but I'm here for it. I'm just rooting for them. Great, great positioning. The de-extinction company. Yep. And they're putting up some big numbers, at least in the funding department. For sure. There he is. Zach, welcome to the show. Can you hear me? Yeah, we can hear you. You kicked me out and then Zoom said I wasn't allowed back in. I'm sorry about that. Well, thank you for joining us.

55:33We're very excited to have you. Can you give us a little breakdown of the last two days you've been on a debate tour? Which debates do you think you've won? Where's your weakest talking point that I can needle you on? We were joking earlier. Hopefully, this is the most enjoyable podcast experience that you have this week because I know some of them got a little contentious. So it's great to have you. Yeah. Yeah. It's been fun. I pray to God I did not lose a debate to a Bitcoin bro. So at that point, I'm just committing suicide and hanging it up. You did go in the lion's den with that one. Yeah.

56:15No, actually, I would say the most interesting conversation I did, we just posted, which which was yesterday with Derek Thompson from the Atlantic who co-wrote the abundance book with Ezra Klein, which I think is a great book. And we mostly agree on 70 % of the things. And, you know, some of the stuff we didn't agree with at the end of, we got to both, which was really nice. So that has been fun. You know, it's funny. Like I, the reason I started to do it is because these topics, tariffs, gifs pie gripping like these are not interesting topics to the vast majority of americans day to day like they don't care about this stuff yeah until right now yep and so i was like all right fuck it let's go have a conversation about things that like i'm interested in and i think are important for america but normally you know the audience was like you know my family yeah so So that's part of why I was doing it.

57:13And still am. I have a few more next week, actually, just to try and make sense of what's going on in this really complicated economic concepts for normal people. That's been my goal. So it's been fun. It's been fun. Are you abundance-pilled at this point? Because I was listening to something about Ezra Klein talking about how he was very inspired by West Wing and the abundance movement feels very aspirational but there's always this question of like when the rubber meets the road can the out can the impact of that actually not just be to me yeah to me at a base level it's an amazing positioning yeah like coming at it it's a huge upgrade like I saw I saw even yesterday there was somebody in San Francisco that was saying we should tax Waymo to fund public transit and somebody was quoting it being like you know basically like the idea is like this is kind of ridiculous like why don't we just make waymos so cheap that that everybody can use it yeah there's an abundance of you know autonomous cars that everyone has access to by the way by the way you do tax waymo already yeah you're like welcome to how this works already guys you know you're a real shitty at building public transport like that's the problem it's not the waymo tax by the waymo loses like a billion dollars a year like they're Yeah.

58:35Anyway. Yeah. I mean, what is your read on abundance? Not just, I mean, obviously give me like the, do you like the framing, but then also is it a meaningful shift from just rhetoric? It will win the election or it will actually have meaningful change if it is implemented. I love the concept because it gets the Democrats in particular, but America talking about growth and talking about how growing the pie and making the pie bigger is better for everybody. And then as a dem or as like a left-leaning person, when you want to talk about how the pie is allocated, which is about taxes and redistribution and social programs and all of that, those are cool conversations to have when the pie is really big.

59:21When the pie is really small and you think of it in this like fixed mentality, then you get into these awful debates of like you assume everybody that there's there's only a dollar to go around but in reality there's way more dollars if you grow and so i love that framing it brings us back to clinton not hillary but like bill back in the day right you know this is the kind of stuff that they were talking about this is why i think the dems you know were a great political party you know 20 years ago yeah um so that part i agree with i think it's great the other thing I think they touched on that I love is that abundance actually matters a lot more in a few specific categories.

1:00:00Like it's not, it's abundance is not created equal. And in particular housing, that housing is kind of like the only thing that really matters in the long run for most people, because it is such a large percentage of everybody's income, either as a homeowner because of the mortgage or obviously as a renter. And like, sure, you can complain about grocery prices, but the reason you care so much about the movement in grocery prices is because housing is eating up 50 % of your income. And so like, go get the big thing and then the little things will matter a lot less. So I really like that positioning as well.

1:00:37I think they nailed the like problem in America is really housing costs and other stuff matter, but not as much. Healthcare is probably the second. And where we disagree, and we had this little fun little disagreement on it, is what is the role of government in creating that abundance? and where do you need government to do stuff such as building roads and providing health insurance through medicare and medicaid through like you know construction of of uh of energy infrastructure and things like that military and where do you need government to get the fuck out of the way because government is the thing holding us up and, you know, rules and regs that prevent the free market and from capitalism from, from building the homes.

1:01:30Right. And there's just a lot of nuance. I think of like where government should and shouldn't be involved involved. And so their book is a little bit more like make government good at its job. And my view on that is like, I don't really think that's ever going to happen. And so So for a variety of reasons, and so there are some places where we actually need to get the government out of the process, and that will help. And it just depends on what you're talking about. Are you talking about, like, you know, bridges and tunnels and highways? Are you talking about homes? Are you talking about semiconductor manufacturing?

1:02:08And we disagree on some of the specifics there. But the goal is noble and good and way better than what we have been talking about for the last 10 years, which is like, I don't even know what we've been talking about. I agree. Yeah, that was a good point. I want to get to tariffs. How did you I mean, like, take me on the roller coaster of emotion that you went through over the past week with regard to Liberation Day and the fallout from it. Yeah, then we can get to the art of the squeal. Yeah, the art of the squeal. It's just been the talk of the timeline today. I have no idea what that is, but interested in learning more.

1:02:41uh you know like because i work day to day and in biotech and in healthcare but specifically in biotech um and in early stage biotech which is before a drug is approved right this is all drug discovery um the tariff stuff doesn't really impact us on a day-to-day basis and so from like a professional side of things it's interesting to watch and be interested in it because you know tax revenue helps fund science. So I'm very interested in growing the economy because it'd be nice for us to have more money to fund more science. That's, that's really my big interest in this, right? Like if America were 10 times richer, then the NIH budget would be 10 times bigger.

1:03:25And you know how much cool shit we would discover as like humanity, if the NIH budget was 500 billion instead of 50 billion, like that's pretty cool. I think it's like 30 billion. And so that's why I'm really interested in this stuff Because I think, you know, as Please. members of society today and our kids, we benefit from a booming US economy in ways that are not always clear and one of them is in science and in drug discovery. So that part, that's part of why I'm interested in this stuff, because I think it's bad for America if we don't grow. But on a day-to-day basis, it hasn't impacted us.

1:04:00Now, there are other things going on in the administration related to healthcare, as I'm sure you've seen with, you know, RFK and the FDA, that we pay a lot more attention to. And there are very nuanced, biotech-specific problems, which, you know, needs a lot of context to share. Yeah. Yeah. With regard to the NIH, why is that$30 billion budget so important? DARPA created the internet, but the free market and venture capital in Silicon Valley is like, well, they kind of took it from there. Why hasn't that happened in biotech? Why do we need the government doing anything? Why can't they just kind of get out of the way?

1:04:40And kid in the lab with his chat GPT, looking it up, formulate something, and they test it. And then the government may be still there to approve it, but it moves a lot faster. I knew this question was coming because I get asked it all the time. And I was trying to come up with a simple way to understand where you need and where you don't need government in biotech. So I'm going to try my best. I want to make the distinction between biology, which is the understanding of how the body works, versus drugs. You can think of it as chemistry, but it's more than chemistry now, which is creating something, a pill, an injection, mRNA, siRNA, whatever.

1:05:29there's a lot of different things you can make to change that biology right one is like i i want to this is a thing that's going wrong in the body i want i want it to do more i want it to do less that is biology and then there's the like i gotta now create something that doesn't naturally happen in in your body or isn't happening to to make an impact on disease and those are two different things. Those are two very, very different things. Biology, you cannot patent. There is no patent for understanding, you know, a process that happens in your cells or happens in, you know, your immune system, right?

1:06:08Like, you got to remember, I think people think we know more about how the human body works than we do. We barely understand how this thing works. I mean, have a little bit of like, we understand, you know, how to like sequence DNA and how DNA creates RNA and how RNA sometimes creates proteins. Like we have some basic level understanding of this stuff, but it is an incredibly complex system that if you don't understand how it works, you can't even start a drug program in the first place, because what are you pointing it to? What are you trying to drug? And that I think is really, really important because there is mostly no business model in finding new biology.

1:06:51Like learning how a certain type of cell works or how cell communication works. Like you can't make money doing that. And it is real. I mean, like, yes, some rare cases, sure. But you really can't. So the solution is to do like a Dow or something like that. I mean, owned. Owned. Admit it. Oh, God. Crypto is the solution. Sorry, sorry. I don't want to trigger you. I know you were triggered. No, it's good. It's good. I'm kidding. And so, like, you know, when people talk about academic discoveries and discoveries in the lab and government-funded discoveries and none of – like, those are all true things.

1:07:29Yeah. But mostly what academics discover, mostly, in this world, in my world, is biology, right? Ooh, this is how this thing works. This is how this process works, right? Like, and then private groups go and go, ah, okay, now that I understand, you know, this thing is causing a type of cancer or this thing is causing cell signaling in a weird way, or this thing is causing Alzheimer's or dementia or some immune condition. Now I'm going to try and make a drug to go and stop that thing from happening or modulate it down. And those drugs you can patent, right? And that's what creates an incentive for private business to go and take a massive amount of risk because most drugs fail.

1:08:13And it costs hundreds of millions of dollars to figure out if they fail. But because you can patent those things and then because that patent life is long enough and the prices on the other side of this are high enough that you can make money doing it. You know, it's not easy to make money. Let me be clear. So you need government to discover biology. without it, you won't, it won't happen. And here's like a simple, there's a few examples of this that I'm going to try and make it tangible, but like, we do not know what causes Alzheimer's. We have no idea. I mean, we have some, we have some things that show up in the brains of Alzheimer's patients that seem wrong.

1:08:50Right. And this is the whole like tau and, you know, some of the, the best way to describe it. There are things in the brain of Alzheimer's and dementia patients that shouldn't be there. And so, okay, we're going to try and like remove those things from the brain, but we don't actually know what's causing those things to, to, to, to be there in the first place. And is that actually, is that causing Alzheimer's? Is it driving the disease? Is it, is it, or is it actually a passenger? There's something else driving it over here and that's causing this, which, what are these things do we try and drug, right?

1:09:23Like, and you know, this is like tau and beta amyloid and all these things that we think are kind of involved, but we don't actually know what causes Alzheimer's. We're not even close to understanding this. And so if you want good drugs for dementia and Alzheimer's in the future, we need to know what the hell is happening now. And without government funding that kind of biology research, we will never figure this out. And because there is no business really in doing that kind of lab work. It's open-ended science. It's, you know, this is where you need philanthropy and money that isn't chasing profit.

1:10:03I'm the first person will tell you capitalism is the greatest thing that ever happened to humanity. But there are some places where it doesn't work. And that's where there isn't, you know, there isn't profit to be made. And biology is really, really, really key. Can you I wanted to get your thoughts on tariffs specifically in the context of the pharmaceutical industry. Trump had come out this week and basically said that pharma is next. But if you actually look at the U.S., China, there isn't a huge pharmaceutical trade imbalance between the U.S. and China. It seems like we buy a lot from them.

1:10:41They buy a lot from us. Should that even be a part of the conversation or is it just a total distraction? It's really, really stupid. And it's stupid for, it's even dumber than that, which is that the pharmaceutical product that you're manufacturing at the end of this is the approved drug. By the time you're at that point, all the hard stuff has kind of really been done, right? Which is figuring out what drug to make in the first place, how to design, how to develop it. All the risk, all the failures are in what happens before this, before the approval. And that's where all the money is made, right?

1:11:21All the money, all the stuff that is like strategic for the United States to be good at is in drug discovery and drug development. It is not drug manufacturing that matters, right? Like that is not the hard part. It's not easy, but like that's not the thing that we need to be doing in America. These are also like giant facilities. There's not a ton of people. It's really more of like quality control processes. It's actually better for us if we can build them elsewhere because it allows drugs to be slightly cheaper for Americans, right? Like I don't view drug manufacturing as strategic to the United States.

1:11:59Like someone's got to do it, but that's not the part that matters. The part that matters is drug discovery and drug development, which is really what our biotech and pharmaceutical companies, that's mostly what they're good at. But just because it's made in Canada or Europe, I don't think really matters as long as the quality control is there. So tariffing the last mile manufacturing is irrelevant. It just doesn't make any sense. Why do we need these things here? It's just going to make things more expensive. So it gets the problem wrong? Yeah, that makes sense. Can you talk about AI in the context of the early stage investing that you're doing, drug development work that the portfolio is doing?

1:12:45I did a command F on your site, by the way. You don't mention AI at all. Yet every time Sam Altman needs to raise another$10,$20 billion, the idea that LLMs are going to cure all disease seems to come up. So I'm curious to get your take on the opportunity generally. Because like our takeaway from this week is like with all the potential of AI, it's sort of upsetting in many ways that we've had to kind of like, you know, you know, take this like detour down, you know, trade war lane. And, you know. Yeah. I mean, at the same time, like when AlphaFold dropped, it was like the protein folding problem is solved.

1:13:26Biotech markets didn't really move. And I'm always wondered why there's such a disconnect there between like the hype around AI and bio versus what the market is saying. But maybe it's taking place at the early stage, but would love just the current rant on AI and biotech. Okay. One of the hardest things to explain in a short period of time, the TLDR is like nice, cool tool in the toolbox, useful. We'll make a few steps of this process, potentially more efficient. not a panacea by any means. Why? Let's see if I can try to explain this. Making a drug is like a 40 step process that requires you to actually make early versions of like the physical molecule itself, right?

1:14:21Because you don't, just because the computer says do X, Y, and Z doesn't mean it's right, right? Like it's not, it's not software predictions where it makes you a picture and then you go, ah, that picture. You have to go and make the thing that the computer is suggesting. If we're talking about in this case, using AI for the drug making part of it, not the biology part. We can talk about biology in a second or differently. So you got to go like, think like make test iterate cycle and it's atoms. It's not bits. It's at, these are physical atoms. So you do go like, you know, and then you take that early version of the drug and you stick it in like a Petri dish and you see what happens.

1:15:00And then you have to, then you have to tweak it and you've got to go make more of them and you've got to run this little loop. And then if that works to where you want, now you've got to go and stick it in a mouse, which mouse does the mouse have the right model? Does the mouse predict the human, which version is it a human? Like there's so many things. And then, okay, you get through the mouse dog or monkey right because because mice aren't people mice aren't people right and now their dogs are monkeys but they're closer to people all right and so now you got to go do all that stuff and so like and then by the way even after all of this even after all this we've gotten through the the petri dish and the mouse and the dog and the monkey and then you go to the people you still don't know if the drug works because you have to go like humans are weird right biology is weird we don't know how it works you know you stick it in if it does everything right in the monkey that doesn't mean it's going to work in the human.

1:15:48And so unfortunately, like there's just the data you would need to make all of these predictions. And then it's not just the predictions. You need the feedback loop, right? Like why is scale AI a$20 billion business? Because you need a bunch of people clicking like not the right picture, not the right picture, not that you need that loop for training. And you don't have that loop in biotech. You do not have that. You have to, the date, the size of the data we're talking about here is like one, one billionth of what you can do in digital land, because in the digital world, your training data set is the history of every written word and picture ever in the history of humanity.

1:16:27Like think of the size of the training data in biology. You don't have that. One other little nuance is your training, even where you do have training data and there's different types of training data, it depends what you're trying to predict. It's protein structure is one thing. A lot of that training data comes in a Petri dish, right? Like we're not actually, when we think about like how a cell works and what's going on inside of it, we're not actually measuring and watching that cell in a real biological system, meaning like it's touching other cells and doing stuff. You're doing it in a Petri dish, which is a piece of plastic.

1:17:03and like the way that cell works on a piece of plastic is not the same way that it tends to work could we tell you exactly what that difference is no and so you have some really big challenges in even just the quality of the training data if you want to use this to make high quality predictions and so it's going to come off as i'm like super negative about all this and i'm not I'm not. I think it's awesome that we have slightly better tools to make predictions about, you know, how to design a small molecule. That's really what this is particularly good at. But this idea that like a computer is designing a drug is so far from reality that that's why you don't see the people who know.

1:17:53Yeah. No. And that's why it's not moving. It's great. It's cool. I'm glad we have it. It's it's it's not going to change the game. And part of this is like not to finish my rant here. If you don't understand the biology, what is the computer trying to predict? right like all the way back to the thing we just you know we talked about in the beginning if we don't even know what causes a particular disease what is the computer going to make for you right like and so so much of the problem of making drugs for the future and there's so many challenges in all of this and it's great that america is good at this but like it's actually on the, it's in the biology side of things where, you know, you know, we, we, we need more science.

1:18:42We need more scientists. We need more smart scientists. Um, so the idea, I think it's a, I think it's a good, a good, uh, it's, it's not necessarily negative, but it's a good wake up call to the sort of EAC, you know, super AGI pill types that say just survive the next two years. And then we're all going to live forever, you know, which is, uh, I did want to ask, uh, Specifically, how you thought about the potential of even just base LLMs within healthcare. There was a study recently around how, you know, basically LLM-based therapy was actually showing to have tremendous potential. And that's cool because the idea that everybody can get access to a, you know, relatively high-quality therapy effectively for free could have tremendous benefits.

1:19:31But I'm curious if you think AI can be deflationary within health care broadly outside of, you know, drug discovery and development, things like that.

1:19:45Probably not at a giant number, but there are some things that I think these elements are particularly good at. I'm not going to talk about the boring part of health care, which is like insurance, reimbursement, prior author, like paperwork. Yeah. I actually think LLM should be awesome on the paperwork side. And so things should get better. You should see things should be faster, more efficient. That'll be great. No more. You ever make an appointment with a doctor and like some, you know, front desk person is trying to call you 18 times. You're like, Hey, have you ever heard of like a text message that I can confirm my appointment?

1:20:18Like that stuff should be solved. Hopefully. Um, I think the other area that is really useful is in a differential diagnosis, which is basically like the job of a doctor to try and figure out what's wrong with you. And part of that is understanding the science. And then part of that is like the literature that you have to, you know, based on the symptoms that this person is talking about and telling you like, what might it be? And that's a probability weighted type analysis that, you know, the docs do and they try and rule stuff out. And a lot of that just comes from experience and education. And that should be replicable by models over time.

1:20:59You know, I think that's getting better. Do you have any specifically one thing that I think is interesting is like way it's seemingly like people are much more excited to get their biomarkers done. And there's services now that allow you to test a really broad range of biomarkers for seemingly less money than going. Yeah. Yes. companies like Function and Superpower and others. Are you excited about just having a lot more data? And obviously, it's like a snapshot in time, but that and data by itself is not valuable. But I imagine that's it right there. You just hit it right there, which is like data by itself, just because you can measure it doesn't make it important.

1:21:43And And just because it's going up or down doesn't mean it matters. And sometimes the data can be confusing and actually point you in weird directions that don't make sense because this biomarker that we think is important, we don't actually have great data to know whether it is. And, you know, certain things can be elevated in your blood because of like the coffee you had in the morning. And it doesn't mean, so I think it's great that people are interested in their own health and interested in understanding what's going on in their body. I think the chances of like wide blood-based biomarker screening, like basically what function does, actually mattering for the vast majority of people is probably not true.

1:22:34It's kind of like healthy neurotic people staying more healthy and neurotic. Uh, and things that really matter are much more for like people of chronic disease and cancer. Like that's really where health, like tech people have this really weird view of healthcare because tech people tend to be healthy. And like, that's not healthcare. That's not, I mean, it's healthcare, but that's not the healthcare that we, that really matters, which is mostly like when you're older and sicker, uh, and chronic disease is just, you know, it was a symptom of aging. And so I'm much more interested in how do you think about health care as it relates to disease rather than, you know, these like biomarkers that don't necessarily tell you tons of things.

1:23:19The flip of it is, look, like every once in a while it is going to catch something that matters, right? It's not never. It's just not frequent and often. There's also a risk in health care. I know it's jump, but like, there's always a risk in healthcare of finding something you think is going wrong, trying to figure that out. And while you are trying to figure it out, harming yourself. And a simple example of this is a biopsy, right? Like you go and get a scan and you got to remember, it's a scan it's not a detail so like you see something in a scan could be an x-ray could be mri whatever and it looks like oh we have this like little nodule here and so is it cancerous is it not we don't know so you have to go get a biopsy which is like a giant needle right that you got to like stick into the whatever that thing is and pull out some cells and take a sample and then go test it and like biopsies are not danger-free and you can hurt yourself especially if it's in your like GI area and especially if you're older right and the doc can do it incorrectly and you can have symptoms and complications doing more in health care is not always better and i think that's really hard for people to understand is like more is not always better um because treat diagnosis and treatment is not risk-free except when it comes to the NIH budget maxing over here.

1:24:58Doing more healthcare services is not always better. But thank you for joining. This is a fantastic conversation. I really appreciate you. Yeah, let's do this again soon. We got to do this again soon. We could go for a lot. I probably have 40 other questions. Yeah, this is fantastic. Thanks for joining. Official biotech correspondent. Thank you. There you go. There you go. Works for me. We'll talk to you soon. Bye. Next up, we got Ian Cinnamon from Apex coming in the studio. Apex is a very cool company. If you haven't looked at it, apexspace.com uh you can actually track uh the aries sn1 the orbit of little uh nominative determinism ian cinnamon okay cinnamon is a spice dune oh okay i was wondering where you're going i like that i like that um yeah you you can track its orbit right now uh the last update 21 hours ago uh manufacturers apex uh the mass 200 kilograms wet 200 kilograms dry it launched on a falcon 9 uh launched back uh over one year ago so it's been up there for a year ago air launched from the air force western test range we're excited to talk to ian about apex and everything in space uh we are going to do a full space day on the show soon um yeah was talking to Ian about putting that together and we'll certainly bring him back for that.

1:26:17But we have him in the studio now. So welcome to the stream, Ian. Great to see you. How are you doing? Boom. Hey guys, great. Great to be here. I heard you talking all about Aries and it's been on orbit over a year now and we have a lot more going up soon. Congratulations. I mean, I feel like there's a very clear binary when you're starting a space company and you get your first thing into actual space. Delian's talked about this with Varda. It's like the number of companies that raised a lot of money and never got a single thing to space. It was too high, but it's coming down and you're part of the movement.

1:26:46So can you break us down for us just the history of the company, the goal, the mission, how things are going and kind of what's keeping you up at night now? Yeah, no, absolutely. So for a little bit of context, the company is very new, right? We're only about two and a half years old at this point. I started the company with my co-founder Max Banassi, who spent his career scaling production at SpaceX, September of 2022. And it took us less than one year from starting the design of Ares, a satellite that's on orbit, from the clean sheet design, blank piece of paper, to it actually working in orbit.

1:27:20That's insane. Which, to your point, it's not actually just about can you put the piece of metal into space? Because frankly, it's not too hard to put a piece of metal into orbit, right? You pay SpaceX, you throw a hunk of metal in there. It has to work. And that, I think, is the key thing that, like, Delian and I actually went to college together. So we've known each other forever. And I think one of the things that we bond over is both of our companies not only got something to orbit, but the damn thing actually works. And that is something that is rare, frankly, in this industry. And like there's other great companies out there that get to orbit.

1:27:51It lasts a few months. It lasts a few days. But having it actually work after a full year and keep working is a little bit unheard of in this industry. So I'm just really proud of the team that we put together that's let us actually do this. Yeah. So talk about the actual product, what you actually put in space. It's on your website, it says Space Cross Platform. I'm seeing satellite bus. Break those terms down for us. Yeah, so in the industry, a satellite itself is not just this piece of map, right? It's effectively comprised of two components. One is called the bus and the other is the payload.

1:28:23So the satellite bus itself is all of the hardware, electronics, the avionics, the control systems that actually let a payload function in space. So let's say you want to go take photos of the Earth, right? You can't just stick a camera and a lens up into orbit. You need to give it power. It needs to be able to move around. It needs to talk to the earth. The satellite bus does all of that. In the U.S., we call it a satellite bus. The rest of the world calls it a platform. Frankly, platform, I have to say, you know, I'm a big fan of the U.S. and America and everything that we're doing, obviously.

1:28:54But platform might be a better name than a bus. But I just can't convince the U.S. government to change the name from bus to platform for whatever reason. They're stuck in calling it a bus. Oh, well. Well, we'll start calling it a platform. So, yeah, I'd love to know about some of the satellites that you're excited about going on the bus or the platform, as we'll call it. What are you excited about? We've heard a lot about Internet communication. We've heard the story of Starlink, obviously. We've heard Planet Labs and Albedo working on stuff to take better images. Are there any? And then obviously Varda is doing crazy stuff up in space now.

1:29:32What's exciting to you? and how are you designing the business to take advantage of those trends? So when we call it a platform, we mean that because we really want to be a platform provider. And whether you want to do something like take images of the earth or you want to beam power down to earth or you want to be able to shoot down missiles that are coming to attack the US, we want to be the platform that works for everything. And I will say, I think we've done a pretty good job. If you look at the customers that we have so far, we have people who are using our platform from on one side of the spectrum.

1:30:01They're doing hyperspectral imaging, which means you have a camera that could see in hundreds of colors that you and me as humans cannot describe. And from space, they could look down at Earth and say, that's farmland with potatoes that are seven days old. Right? Like that's nuts to be able to see from space. Then we have other customers. Aetherflux is one example, founded by a good friend of mine, Baiju. He started Robinhood and now he's going and he started Aetherflux, which is collecting power from space and beaming it down to Earth, all using our platform, which is phenomenal. Then we have other customers who are using it for defense missions, right?

1:30:36Andruil's a great customer of ours. We continue to do more and more with them. And then there's others that I can't talk about yet at this time. But the theme that honestly excites me more than anything, it's not just connectivity or taking photos of Earth, but something that has become, frankly, something we've always loved, but has only been talked about recently in the public domain, is this idea of how do we develop the technology to protect the US and our allies from the scariest threats that are beginning to emerge? And to me, that's everything from hypersonic glide vehicles, to missiles, to everything else.

1:31:08And being able to use satellites in orbit to actually protect from this is phenomenal. And that's, you know, you read the recent executive order titled Golden Dome, right? It's asking for space-based capabilities that protect the US and our allies from these very scary threats. And that is what, frankly, for me and my company, we were most inspired to support. Do you think this speed at which you guys have not just become a company and developed products, do you think that can be the new normal just based on the sort of talent density that SpaceX had and other companies of that generation training people like yourself on how space works?

1:31:48One question I have is people in Silicon Valley have for a long time had this sense of looking at an industry they know nothing about and saying, I'm going to go do it. I'm going to fix it. I got this, guys. And they can raise a lot of money for it. And that's honestly, in many ways, played out very well in a bunch of different industries. But space seems like an order of magnitude harder, potentially more. And I'm curious if this is an industry where you can get people coming, you know, true outsiders coming in and, you know, actually solving problems or you're better off, you know, putting your time in at an Apex or a SpaceX before kind of making that leap.

1:32:30It's a great question. I mean, I'll be the first to admit I don't come from a space background. I come from a software background. I built an AI computer vision company and I sold that to Palantir. Like that was my background. Now, at Palantir, I got to work on software related to satellites, but I wasn't manufacturing hardware. So I'm an outsider. I'd argue that you look at other successful companies out in the market, right? Like we mentioned Dellian and Varda, right? Like Dellian does not necessarily come from an aerospace background, right? Like he was working on venture investing and other things before.

1:32:59I think the secret, though, is you need to pair that outsider point of view with a brilliant technologist who can bring it together. So in my case, my co-founder, Max, spent his career at SpaceX. In Delian's case, he had Will Brewey, who came from SpaceX as well. And that, like, the combination of that together, I think, is what truly is special. If you look at companies who are founded, who frankly come too far from the direction of purely kind of the space industry, it's difficult, right? Because you end up falling into the same traps that you're used to. There's phenomenal companies out there, you know, True Anomaly, Impulse, all these others, that frankly, like, they haven't had success in orbit.

1:33:36and their satellites were unfortunately not able to succeed. And that's bad for everybody, right? You need a little bit more of that, I think, kind of outside perspective to come in and combine in order to succeed. It's a tough industry. I will say, to answer your question about speed, I look at where we are today, and all I can say is I certainly hope that the market is not going to go at the speed that we're going at, because I think we're going way too slow, right? I think we should be moving much faster. I think you should look at what Apex has done and say, why did it take a year to go from a blank piece of paper to a satellite in orbit?

1:34:07Why can't you do that in a week or a day? That to me is where I think we need our mindset to be. Not, oh my God, it went from five years to one year. That's still slow. And I think we all need to embody that. Can you talk a little bit about, I mean, your craft is that it looks like something around between 450 and 500 kilometers up right now. Can you talk about what's happening in VLEO and then what's happening above you and where the biggest opportunity is going forward? Oh, man, there's a lot of opportunity out there. And I would say it's all orbits are fascinating from my perspective. So if you think about it, right, we're here on Earth, we're at, you know, zero kilometers of orbit, right?

1:34:48We're standing here on the ground, the higher you go, the different different capabilities are unlocked. So there's capabilities that frankly are best served by being airborne, right? So that is flying at, you know, what an airplane or a high altitude vehicle or even a high altitude balloon would be at. Then you move into very low earth orbit, VLEO, you're moving a lot faster, right? Because you're circling the earth, you're kind of looking at these different things, you're much closer to the earth. So you're able to get much higher resolution imagery. So there's interesting trade-offs there. Then you move into where we are, which is low earth orbit.

1:35:20That's an orbit that frankly, I don't think there was as much commercial demand for it until, call it about 10 years ago, when this wave of proliferation started and you realized you could put a lot more vehicles up. Then you go higher altitude than that and you go to medium Earth orbit. That's where GPS satellites are flying, things like that. That's also a very interesting orbit because you're further away from Earth, so you're more protected from things like missiles that want to shoot down a satellite. But at the same time, you're not so far away that it takes too long for radio waves to communicate between Earth and the satellite.

1:35:53And then you have a geo, right, which is where traditional satellites tend to operate. And geo is interesting because as Earth is rotating, you have the satellite rotating at the same rate. So you're always over the same point over Earth. And that's companies like astronomers, I think have done tremendous work over there. And I'm very impressed with the work they've done. And then from there, frankly, you go to deep space, this lunar, you have companies like Astro Forge launching asteroid mining missions, setting kind of world records for the furthest to commercial satellite is gone. there's more and more capability with each of these orbits.

1:36:26Can you talk about how the aerospace industry is reacting to the tariffs broadly? Obviously, you're doing a lot in defense, so I'm sure you are not sourcing from China. But I'm sure generally there's a lot of space companies that are kind of caught off guard here. I'm curious what your read on it is. Space is a global industry, right? And I'm a big believer that we are in what I call the new space race right now. And it is the U.S. and our allies against China and Russia. And we have to win. That being said, as a proud U.S. company, right, most of our parts come from the U.S. So we looked at our tariff impact and it's pretty much nothing, right, as it should be because we're a good U.S.

1:37:11defense company. We do sell a lot of our satellites to allied nations overseas. So as the tariffs come in, that doesn't affect us, but reciprocal tariffs from other countries may affect us as that happens. The good news is, though, most things in aerospace and defense purposes tend to be exempt from these tariffs because at the end of the day, these are the kinds of trades that we want the U.S. and our allies to be engaging in, where we're sharing the cutting edge technology that allows us to stand up as one unified unit against those near-peer adversaries. So I think the more commercial oriented areas of the space industry are probably more affected.

1:37:48That being said, we are able to produce in the US and we're not having any issues with it. So what I would just say is, you know, we want to bring more of that knowledge and that manufacturing onshore and continue to see that rise as a trend. Awesome. Well, thanks so much for joining us. We'll definitely have you back. We're going to put together a whole space day and go way, way deeper. But I'm so glad you could come on the show today. I really appreciate you taking the time. Thanks, guys. It's great to be here. We'll talk soon. This was great. We'll talk soon. Congrats on all the progress. And he mentioned Anderil's a customer.

1:38:19We have a person from Anderil coming on the show, Jen, who has been leading the design charge over there, creating some incredible hero films. I don't even know how you describe them. Motion graphics, anime, all of the above. But I wanted to talk to her about the evolution of Anderil's design language. They still do a lot of these like military inspired films but now they're in an entirely new territory with the animation and I'm sure we'll have a lot to talk to her about the AI trends, the different motion graphics trends, all the different things. So Jen, welcome to the stream. Hey, thanks for having me.

1:38:59Thanks for joining. I'm gonna go out on a limb and say you're probably the most copied designer in tech right now And I'll give credit to your team too. It just feels like the entire industry is just racing to like do what you guys did maybe two years ago, but you're consistently staying out on the edge. So it's great to have you. Got to keep them on their toes. Yeah, you do. Just keep innovating. And by the time you get copied, they'll just be copying the old thing. I do want to talk about that evolution of the design language. It feels like the cartoon anime style probably came from Palmer. He's a big Sword Art Online fan.

1:39:36Was that the genesis of this? And how do you think about the evolution of kind of the, I don't even know what you describe it as, like the brand language, the video themes, whatever you call it. The brand world. The brand world. Yeah, yeah. I mean, so I actually pitched this about two years ago doing an anime and like obviously very easy style to Palmer, right? But it just wasn't the right time. We didn't have the team that we have now. Sure. And it obviously pairs really well with Palmer's identity, right? But also this anime style for Andrel just made a lot of sense. So if you look at anime as a medium, historically, it's leaned really heavily into military themes, right?

1:40:14The battles and the really complicated dynamics using really futuristic machines, like the mecha and the cyberpunk of it all, right? And really advanced technology to defeat adversaries. And man, the future is here. Andrel is building it. So of course, we lean into that. and like you know and like the way things that we do things um it's fundamentally different the way we build how we operate but part of the way that we do things differently is really the way we approach our communication and our product identity um the way defense looks and feels and we just want to cut through uh the noise right the sea of sameness like it's uh really authentic to to who we are uh how has does uh recruiting changed over the last couple years on the design side specifically there was not i i remember you know i imagine now you you guys have hundreds of applications for every role it's sort of normalized you know working in defense but um certainly only feels like it half you know five years ago it felt like there was this pretty frequent tendency if you're if you're a startup doing anything defense related you would get consistently turned down by designers that said like yeah like it's just not not in my wheelhouse um but i think that's like changed i'm curious what your read on it is yeah i mean i can speak from my own experience as well right i think there comes a point in your career where um you want to utilize your skill set and uh you know towards a greater mission i think what we're finding now with um you know all the conflicts around the world a lot of the designers who are applying are coming to us saying like i I really want to support and use my creativity to be able to do this.

1:41:54And what an amazing opportunity that somewhere like Andrel is providing that is welcoming these people. And I want to see that, right? Like the end game for me is seeing all the different disciplines in design coming here and creating all of this work is such a rewarding and unexpected thing for me to witness. And so it's been really incredible. how did you process the studio ghibli moment specifically in the context you guys came out with the film was it i mean was it monday or something like that i think it was monday um but you i'm sure you guys have been developing it way before studio ghibli uh and it feels like you guys were maybe the last company to get to launch something like this before there will be an inevitable wave of others being able to kind of create versions of it but for like you know one one hundredth of the cost but i'm curious how you reacted to that uh yeah it's so interesting i mean like the ghibli like look that it's it's pretty good i will have to say right it's a really it's really good i'm really impressed and i think there's just like so many tools out there uh right now um you know we launched our first one barracuda uh last year so we were sort of like on this wave and we also started noticing the trends in the industry too like mcdonald's did one the chargers did one and we're like yeah we must be kind of onto something here um it wasn't until like just recently this boom and obviously like the ghibli filter on chat had a lot to do with it as well um but i think for us it was really about again it made a lot of sense and uh you know pushing our uh you know skill set in-house to our toolbox of all the different uh programs that we're using and you know one could argue that all the rendering tools uh that we've using that we've been using for decades now are that was AI right and so I think here like we want our autonomy to enable humans to do what humans do best palmer says all the time and robots to do what they do best and I think we've got the best of the best in our in-house focusing on making the art and we're utilizing the tools to do the can you walk us through the actual process of making one of these films does it start with storyboarding and then you move into blender cinema 4d I'd love to to know kind of like what the team looks like from the motion design perspective to take from an idea to a finished film.

1:44:13Totally, we certainly use cinema, you know, blender, redshift for texturing, like, you know, Rhino, Keyshot, I mean, the list goes on. Like everything starts with a mission set and as we call it, the concept of operation, Conop for short. And our storyboarding is really all about distilling like an incredibly complicated, as you can imagine, technical and operational scenario into a like one to two minute narrative, right? That is digestible and is visually provocative for really any audience. And we do some keyframes to dial in that texture and the lighting, get that right. Then we move into pre-vis pretty quickly.

1:44:54And that really helps us block in all of that kind of technical authenticity early. We find a rhythm, then we go into production and all like the final paint overs and stuff. And all the designers that are involved in all aspects of the project, they're working together in this process. So we've got art, we've got motion, editing, graphics, music, and sound, right? And we're also vetting these storyboards with our customers along the way, making sure that we have their feedback. And so like this entire process, I think for the Maritime short, the most recent one, we went from sketch to final video in like five weeks.

1:45:29Wow, that's very nice. It's crazy. I mean, products at Andrel move quickly into production. Like we're just here trying to keep up. Where has AI been particularly impactful in your guys' process today? I'm sure Anderil doesn't do a lot of vibe coding on the engineering side just because you need a bit more reliability. You want to kind of make sure that things work consistently. But I'm curious where you're seeing an impact today and where you hope to see kind of more of an impact in the future. Yeah, I know we're an AI company, but we don't use any AI in our art and certainly nothing generative.

1:46:09Right. And like I said, I think a lot of the tools that we use, right, could could be in that AI kind of bucket like we've been using it for a long time. But I think where I see this being really useful, though, and where I want to invest in is really in that previous stage. So and it's not just for animation. I mean, for blocking in practical photography and videos as well, like figuring out the composition and the lighting setups, the camera movements. It's just so critical for us to have a watertight shot list whenever we're going into, whether it's animation or a video. A lot of the times, like especially on the video side, we have one take to get the shot, right?

1:46:47And so like kind of building this ahead of time and all of that planning is so important for us. And that's kind of where I'd like us to spend a little bit more time and investment in. Yeah, it's interesting to think about the workflow now where even if you're planning to do a one-day shoot and it's going to be really expensive there's going to be a big crew there's going to be an andrel's case like actual hardware involved you still you might generate you might generate all the core scenes so you can like visualize it but then still still actually go through it so the shot list just becomes like basically a preview yeah essentially uh with with ai kind of commoditizing or democratizing certain art styles.

1:47:27Do you think that there's maybe alpha or even just like a creative urge to find specific art styles that are incredibly difficult for AI to work with? We were talking about Houdini earlier, and I don't know if you followed like what Man Versus Machine did with Nike and the Flyknit. It's like this incredible weaving of threads to create the shoe, and it's basically impossible for generative AI to crack that at this point. But it's very different than like the Studio Ghibli style, which is kind of becoming easier and more accessible. When AI commoditizes or democratizes a theme, does that push you to go deeper in that area because it's more accessible or find a different niche to kind of go and explore?

1:48:09I think it's a different niche. I think, again, like our subject matter is really tied very intentionally to the mission set. You know, it's hard to prompt things like, you know, we need these machines to talk to each other and come up with a way to do that, come up with a way to visualize that. Right. It could be any number of things. And I think so really kind of helping guide that by, you know, using I think just the people who are creating this art is probably where we're going to stay and kind of where we're going to lean into even harder. Is being so technology first in the Anderle organization broadly, is that an acceleration for the motion design team?

1:48:53Because you can just ask an engineer, hey, can you pass me a CAD file and I'll export as OBJ and just import it into Cinema 4D and like, boom, I'm good to go? Or because of ITAR, I imagine that you might have to rebuild these models from scratch. Is there anything where you're like, hey, actually we can't use that model over here because it's like, we'd be leaking if it got hacked or something. Yeah, no, we're definitely not, you know, picking up like the engineering CAD file and dropping it to this, right? Like even for previous, like we definitely, you know, a lot of the designers on our team, like they're, they're very, you know, technically fluent.

1:49:27So a lot of the CAD work that's done even to get it there is like about scrubbing it. And it's really like creating, again, turning it into almost a whole new OBJ that's water tight. So you can like texture it and throw lighting on it, right? So there's like a lot of work that goes into even setting up the files. And obviously also like, you know, make sure that it's not like exactly the thing that, you know, we're building. Yeah. Have you had any explorations using virtual reality in combination with motion design? You know, there's a lot of move towards, oh, we're going to do a lot of motion design in Unreal Engine.

1:50:02I don't know if you've been tracking all of that stuff with the real time motion graphics, but you can imagine, you know, there's even these demos where you walk around, you sculpt something then you export that obj and then you're playing with that in cinema or houdini uh i imagine that palmer makes it easy to you know uh expense a vr headset if you need one but uh what have you have you played with any of that and are you optimistic about that going i mean like uh i think like that that like iron man future is certainly something that i want um there's so many times when like you know we're building something even my past experience like if i could just like just reach in there and just like pull that you know like i just want to like do it right yeah and you know i i think uh if there were was uh you know a forum to do so like it may be in some of our exploratory phases like right now we've got um designers we have a whole different pipeline just to kind of do experiments right we have an australian designer his name is joel he's doing some crazy experiments with blender right now with textures just like you know pushing that to its limit and like James who comes from working on movies like Avatar, like, you know, he's trying to like recreate like, you know, worlds and like do world building and like unity and stuff.

1:51:10Right. So, yeah, we're like constantly trying to figure out ways to also level up our technical ability to be able to do better discovery, right, better experiments so that we can like produce better work. Yeah. Well, since it only takes you five weeks to produce these, you're on the clock. We look forward to checking in with you in four weeks when the next one drops. But I think everyone, I think everyone in the audience is very excited when these drops. So thanks so much for coming on the show and breaking it down for us. That was really fascinating. Yeah. And thank you for the work you do for the country.

1:51:42Yeah. Thank you. We'll talk to you soon. Thanks so much. Bye. Fantastic. Very cool. Yeah. Anderil has the blessing and the curse of whatever they do, They do it so well that they get immediately copied. And imitation is not flattery. It is a shame. And what would your mother say if you're imitating people? Well, we have our next guest. Come on into the studio, the Temple of Technology. We're ready to talk to you. How you doing? What's going on? I'm doing great. Welcome to the stream. Can you start with just giving a little bit of a breakdown on who you are, what you do, what's happening this week?

1:52:19Okay. My name is Rylan Hamilton. I'm the co-founder and CEO of Blue Water Autonomy. We design and build autonomous ships, ones that can operate across the open ocean. So think thousands of nautical miles of range and being out there for months at a time, like no people. And we just came out of stealth today. So like super yachts for billionaires who want to just like hang out without any crew for months. Is that what we're talking about? Eventually. There's like some defense angle here. I like that. That sounds a little bit easier than what I'm doing now. No, but due to regulation stuff, it's the Navy.

1:52:52They have an urgent need. We have a shipbuilding issue within the United States. Yeah. What's super cool about what we're doing is because there are no people on board. So when you remove the bridge, the galley, the birthing, the showers and all that stuff, you can make the ship a lot smaller. So our big Navy shipyards today are at capacity. And we're trying to build destroyers and carriers and submarines as quickly as possible. But our size allows us to build it in all these mid-tier shipyards all over the country. And so we could build them today. So how small are we talking? Like 25 feet, 10 feet?

1:53:22So think like 150 feet. So like if you get too small, you can't go across like the ocean. You don't have enough fuel. You're going to get whacked in different kind of sea states. But if you're too big, you're too expensive. So there's kind of this sweet spot in the middle. So what does it take to cross the ocean? Is this like you just diesel fuel and you're good? Or is there some other alternative? That's not a bad guess. We're not refueling everything. But yeah, we're pretty practical. like we're taking commercial off the shelf green kind of um components sure but we're redesigning the ship from the keel up because uh our learnings have said that if you just take like a commercial design that's meant for people and put sensors on it that's okay for a week or two but like eventually it's going to break down like you have one engine room and something leaks and if you've no one on board or no like you just you you can't you know you're gonna be dead in the water in the middle of the ocean that's the last thing that rdb wants yeah yeah it's interesting um when when we first talked rylan i think we were talking about this you there's this youtuber and i'm blanking on his name now but he just like nerds out about all the autonomous vehicles being used in ukraine and it's crazy stuff like there's people just applying you know putting like a sensor on a jet ski and it's just like out you know going around it's cool it's really wild i'm curious so i talked a little bit about your backstory earlier on the show you started your career in the navy then you went into robotics you're at amazon sold a company to shopify how much of both of those chapters of your career shaped what you're doing now it's kind of like you're the perfect guy to build this business but you had to spend you know however long in the navy and then you know do the robotics thing and and then actually you know do it at scale um but i'm i'm curious you know how um uh the kind of learnings you took from each of those chapters yeah i think this is kind of my dream job except we're a startup and we're hustling every day working seven days a week so it's not easy but yeah last 20 years like my first four years as you said surface warfare officer in the navy i was an engineering officer i was actually an electrical officer so on our ship we had fires floods like you name it we got hit by lightning the ship would lose power uh but i learned a lot so over the you know four years after i went to graduate school and then i actually got lucky and i joined this company called kiva systems that got bought while I was there by Amazon, became Amazon Robotics.

1:55:37I was employee 200, became the youngest person on the leadership team. But I was there within like a couple thousand robots were in the field. We scaled it with Amazon to 15 ,000. And now like my guess is there are over a million robots in Amazon warehouses based off that company and that technology. And so I left and then I started my own robotics company, again, collaborative autonomous mobile robots. So ones that didn't need fencing or stickers or anything like that. And I kind of feel like right now we're building kind of one of the world's largest mobile robots. It's just a ship that has to cruise across kind of the oceans.

1:56:09And so we're taking all these lessons where you take the hardware primitives and you smartly design it into something that's reliable and works all the time. And you got to write the software on top. And so like that's the talent that we have on our team right now. um how when how uh how much or how little has the average navy vessel advanced since you were on them back in the day i imagine it's not as much advancement as one would like so we let me just be clear we build the best warships in the world um they are amazing you look at a guided missile destroyer it's been through a lot of iterations it's it's fucking awesome however it takes a lot to make a design change on a warship because there's so many different components.

1:56:50You have to make sure the supply chain's there. So it was a couple of weeks ago, I was actually walking through a big Navy shipyard and I was looking and like you look at the gauge and there's like the concept of IoT is just not there, but they work and it's been like hard tested. And so we have this opportunity when we redesign everything to kind of put and make everything smart and have sensors over everything. And so right now, just in our core design, it'll get better over time. we have 550 kind of data feeds into our edge compute, just to know what's going on in the ship. So it's like, we need to know everything, but that's just not where you can get on a modern Navy ship.

1:57:22That makes sense. Can you talk about the Jones Act? Was it good or bad? Maybe you don't have to give that aggressive of a take, but in this era where deregulation is popular, I'm curious if you think that should be good or bad? further. Obviously, we're just not building nearly as many ships as our adversaries. And you guys are hopefully changing that. But I'm curious your thoughts. So I'm not an expert on the Jones Act, but I'm definitely aware of it. But I think what's critical is we need to be able to build ships in the United States. Like sea power has been the bedrock for our prosperity and security for the last 100 years.

1:58:05And that needs to be, you know, we need to have sea power for the next 100 years. So we can't outsource shipbuilding, we need to have that as core ability within the United States, and we need to get a lot better and increase our capacity. I think because of the Jones Act, we still build ships here in the United States, and that's important. Maybe there could be some tweaks to it, but we can't outsource shipping to other countries. It's like too critical for our nation. Paul Merlucky was on the Sean Ryan podcast talking about how China has vastly, vastly more shipbuilding capacity. And a big part of that is that civilian ships in China must be built to military specifications.

1:58:42They need to have roll on, roll off capabilities for troop carriers. You could argue that maybe American civilian ships should be built to military specs, but that kind of spits in the face of the free market. What if I don't want tank treads on my pleasure yacht? But where do you stand on that as kind of just from an American standpoint? So I haven't walked to a Chinese shipyard, and I don't think I ever will. I don't think you're welcome. But I know that we build ships a lot better in terms of warships than any other country. And they're built. And if you look, there've been incidents where we've been hit by mines or ship missiles or like any or like anti-ship missiles and they withstand and they don't sink.

1:59:24And so like a lot of credit to our Navy in terms of how we design things. But if we just try to compete with China head on head in terms of like who can build more destroyers, they're going to win. So we need to out innovate them. And I think there are a lot of there's not a silver bullet but if you look at the things that andrew is doing underwater it's awesome right and if you look at our nuclear submarines we're generation ahead on submarines versus chinese and now what we're trying to do is saying on the surface let's complement our manned fleet with an unmanned fleet so we can kind of be a force multiplier so we don't have to put a ddg in every area that we can kind of be the front line of the picket fence for some of these destroyers and put it in harm's way before you put you know the destroyer out there and so it's not like um I think the point is, let's innovate and lean into autonomous systems, but let's not pretend like we're just going to change our shipbuilding capacity overnight and compete with the Chinese.

2:00:12How do you think hypersonics change the nature of naval warfare? It definitely changes it, because if you think about our exquisite platforms, whether it's aircraft carriers or destroyers, we're not going to put them as close as we would to an area of conflict because of these anti-ship missiles and hypersonics with extended range. Now, one thing you have to appreciate about our military, it's a system of systems. So there's a lot of, and I'm not an expert, there's a lot of cool stuff that's happening in space and others. So it's really hard to hit an aircraft carrier that's going 30 plus knots that's a couple thousand nautical miles away.

2:00:44And we have some countermeasures to it. So, but it definitely means that we're going to sort of keep these ships farther and farther away from areas where there could be either hypersonics or anti-ship missiles. That makes sense. What do you think about, you mentioned early on about there seemingly is domestic, our existing domestic capacity to make the size ships that Blue Water Autonomy is making. Do you think that the actual manufacturing layer still needs to scale significantly or is the capacity there and it's just, you know, needs to be allocated more efficiently or needs to get more efficient?

2:01:22Yeah. So the capacity is there to build them, but we can do a lot better than just like take the status quo. And so these shipyards, they build ferries, they build tugs, they build barges, they build crew transfer vessels, but their business is kind of lumpy. And so it doesn't make sense for them to invest in a lot of automation. But when the Navy can give a strong demand signal and say, we're going to build a lot of these, that's when you can really double down in terms of building like these shipyards of the future and put a lot of investment into it. The other thing is since we've been designing this kind of ship from the keel up, everything's like, you know, in CAD, it's all there.

2:01:54It's all digital. It's all already tied to a product lifecycle management system. And so we get the benefits of starting like today with modern tools to build this, which allows us to kind of like outsource this to a lot of different like some of the parts of the ship could actually be built in like a central facility in the United States. You don't need to put everything in the shipyard. But when you build a ship today, so let's say it's like a steel hole ship, you build the hole, you may be building components, you stitch together, and then you send all the electronics and everything and then you commission it on the ship.

2:02:22Well, that's the hardest place is actually turn on a system in the middle of a hole in a shipyard. And it just takes a long time. And so I think there's a much better way that we can kind of build these ships at scale. How do you think about what Anderl's doing on the lattice side? I could imagine if you're producing a ton of autonomous ships, you'd want those to be somehow teaming with dive XL submarines and drones overhead and all sorts of stuff. And I'm sure the modern warfighter would want to have all of that on a map, essentially, or be able to integrate all of that. How are you thinking about integration with other systems that might be competitive in the financial markets, but close, close allies in the geopolitical sense?

2:03:05Well, I haven't talked to Andrew yet, but we just came out of South. And maybe because this podcast will reach out to me, someone from the Lattice team. But I think it's really important that we talk to systems like Lattice because we're not going to be the control system for all these like UXVs that are out there. That's not our bread and butter. Our bread and butter is building autonomous ships that are reliable and that just work. And so we want to make sure that we have the integration layer to kind of connect all these different systems. At my last company, we built an automation solution where we were in over 120 warehouses all throughout the world.

2:03:34and we talked to like many different types of warehouse management systems so we needed to build in that integration layer and not just be beholden to one w mess and we have some of those same people on the team and that's what we'll need to build like we need to be agnostic to kind of the layers that we speak to above us um random question not related to uh the current business but i'm really interested to get your take on it um uh we've talked a lot about the various companies doing humanoid robotics uh one of the the pitches for humanoid robots is that hey we have all these you know um fulfillment centers throughout the world humans are working them humans are great at a lot of stuff but they're also you know some sometimes they wake up and they're sick or they don't have you know enough energy or whatever um you were working in robotics in fulfillment for a very long time at the you know incredible scale uh do you think that uh humanoid robotics are overhyped in the context of fulfillment or really have potential there that you know maybe people aren't seeing so humanoid robots are very divisive in my kind of like robot mafia i i kind of think like when i started my last robotics company is 2015 everyone's saying in five years we're gonna have autonomous vehicles and it ended up being 10 years and i kind of feel like we're at that moment where it's super exciting to see all the progress we're making i just don't think it's been five years.

2:04:57I think it's closer to 10 years. And I don't think humanoid robots are going to do everything, but I think they're super exciting. One thing actually, like a little tidbit that I learned recently is that everyone's like, why don't you just put it on a big mobile base? It's a lot easier. Well, if you think about like a warehouse and some of the activities that you do, when you have two little legs, you can actually get into like areas that are a lot tighter. But if you have a really small mobile base, they're going to tip over if they're trying to carry something heavy. So I think there's a lot of experimentation that's going to happen.

2:05:23um also like things about humanoids is when they fall humans we brace ourselves with our hands well then those hands need to be robust if you put a lot of sensors on the hand and they fall all the time well then it's not going to work and you have to repair them so i think there are a lot of things that we're going to have they're going to have to iterate on over time but there's a ton of investment that's happening in that space and it's super exciting for robotics overall uh last question last question in the context of humanoids is there ever a point in the future where you could imagine a humanoid on one of your ships or is there just going to be better form better better form the ship directly that's the future it's gonna be a sailboat operated by humanoids yeah not not that not that silly but but i'm curious if you could imagine a future or there will just be cut the jib or they'll just be better better form factors um i don't see it right away uh i don't think it makes sense to have a humanoid actually turn the wheel when you turn the wheel just just saying that's what i want i want to be drinking don't parry on in the back having the sails go up and down by the humanoids pull pull pull pull the ropes here's the thing um i actually think as we think about building our ship we're like the taxi for a bunch of other uxps so it's not humanoid you can put on these but you can think about what andro just launched they just launched these really cool uv's or whatever they're calling it um or if you think about drones or other autonomous aircraft like we could become the modern aircraft carrier to carry these kind of drones or these smaller things that don't have the range but we can carry them out in the same way that SpaceX is carrying all these cool satellites up into space.

2:06:51We want to be that vehicle that makes the cost of getting these things out there cheaper. And so we can carry a portfolio of these kind of drones on our ship. That's fantastic. I mean, congratulations on the launch. Look forward to following the company and thanks for coming on the show. Thanks for having me. This is fantastic. We'll talk to you soon. Bye. Next up, we have Zach from Conductor AI coming on the show in a minute. That was really cool. Well, if I had one more chance, I was going to ask him about to take me on a tour of the various seas across the world. I've heard that the Taiwan Strait is particularly unique in terms of the wave formations and the tides.

2:07:32And so that there's certain times. It's almost like Taiwan doesn't want to be invaded. It's like they perfectly designed the sea. Topology or topography built different for sure. Yeah. Anyway, do we have Zach in the studio yet? Not yet. We can go to some timeline. We can go to some ads. We can tell you about Bezel. Go to Bezel. Go to getbezel.com. Download the Bezel app. Your Bezel concierge is available now to source any watch on the planet. Seriously, any watch. You name it, they'll find it for you. They're listing a bunch of stuff. They got the day date. Nate Bladorn posted earlier. He says, you might never be as tall as John Coogan, but there's always a way to feel like you're up in the clouds.

2:08:15head on over to Bezel and choose a pilot's watch from their fully authenticated luxury selection. Tell them the technology brothers sent you. I love that. Thank you. Thank you for posting. Great post, Nate. Fantastic. And we have Zach here in the studio. Let's bring him on in. Welcome to the show, Zach. Good to be here. Hey, guys. Thanks for having me. The man with no X account. Oh, yeah. You're the second guest, I think, in show history to not be on X. What's going on? I mean, are you not into technology and business? What's truly tragic is I'm in totally a Twitter, like I've actually blocked Twitter on my phone because I use it so often.

2:08:50I just never post. I only read. So clearly, clearly I need to become more involved in the scene and actually post something. Yeah. Yeah. Well, if you create one tonight after the show closes, we'll tag you in the clips. We'll try and get some audience. Or you can never post, never use the app, just come on the show. We'll clip it up and then we'll posted for you and you could get all the benefits with none of the time invested other than, you know, like he's so elusive. He doesn't even count yet. I'm seeing him all over X and these viral clips. I love it. Anyway, can you do a brief introduction on you, the company, kind of what you're building?

2:09:22Sure. Thanks so much. I'm Zach. We're building Conductor AI, a company designed to automate paperwork in the government. Yeah, that's what we're doing. What's the first go to market? Who's the first customer? What's the first piece of paperwork that you're automating? Sure. So we are working with the Air Force, the Office of the Secretary of Defense, and a few other defense customers around problems like security classification, foreign military sales, ITAR controls, and things that are these very important and very sensitive document review workflows that if you've spent a lot of time working with the Department of Defense, you will unfortunately encounter that we are trying to hopefully make it easier to deal with.

2:10:07Just how much paperwork does the government do? I feel like it's a lot, but I think it's probably more than I'm even imagining. Yeah. So the government, there's internal estimates. They spend basically like$18 billion a year on classifying information. And it's this very complicated apparatus that has grown over the years to ensure that we keep secrets safe. But one of the unfortunate byproducts of having a complicated and expensive classification apparatus that can make it very, very hard to actually get information to the people that need it. And so there are many, many separate chains in the government around how you can share and approve and release information.

2:10:51And conductors trying to make it easier to solve that problem and then other problems where you have these very complex approval processes and navigate through them. How much of this is just going from paperwork to a modern CRUD app using standard software as a service versus AI, agents, LLMs, all the on-trend tech? So the people in the government that are doing this, they actually often have some form of system here for this. What we're actually seeing is the people that are sort of doing these, that are sort of these adjudicators that have these very important jobs of saying, yes, you can say this, no, you can't, they're sort of in charge of these complex approval processes.

2:11:39The challenge for them is there's not just like one page of rules or like one document, oh, this is the rules to follow. There's like 20 ,000 pages of policy that they are supposed to have read and follow and apply on each of these cases. And obviously, That's incredibly hard. And so what we're trying to do is make it easier to uh more quickly read and apply those those very complicated policy guidance to those four PowerPoint slides that you're trying to. Yeah, when you're working with the government like this is are you are you following like the standard defense tech sales playbook like SBIR program of record or is it just a completely different sales motion for you?

2:12:22The so I worked at Palantir for seven years before this. So we have some familiarity there. SIBRs are great. They're not all worth the same. Our very first contract that we got was a SIBR. Unfortunately, we haven't won another one since then. So since then, it's all been getting OTAs and subawards off of existing contract vehicles. But yeah, the go-to-market motion in the government is... I mean, frankly, it's fortunate. I worked at Palantir, so I saw saw how this worked from a best-in-class perspective. But if I was a 23-year-old just starting out, I'd have no idea how you'd figure out how to navigate this world.

2:12:59It is not friendly to outsiders. Is it weirdly frustrating when you see just how much the government spends on tons of random things, and yet you're selling them a product where you're like, this is gonna save you a bunch of time, and through that, save you money if you just sort of adopted it? Because it's the or or is it almost like a positive feature that it's hard for the government to start? Even even if they know they want something, it can be hard to, you know, get onboarded. And yeah, yeah, it's more the latter there. It's like the thing that I found pretty interesting is that sort of these challenges you have of how hard it is to break in, how hard it is to get these first contracts are in some ways the defensibility from a company perspective.

2:13:46Right. It's that like, it does take a super long time to get your foot in the door with these organizations and then build trust. But to a large extent, because of how much process there is, once you're in, there is a bit of a word of mouth. So I think that's like, I don't have a Twitter, but there's people in the government and various like means that people communicate in the government. And that's, that's sort of where if you're in the defense tech space, you want, you want to be more present, if you will. How much of your company is built on like this thesis that there would at some point be a shift towards like government efficiency.

2:14:25And then you maybe hit the jackpot with Doge becoming like the biggest meme in government maybe ever. But it seems like some of the, like, even though Doge has been somewhat controversial and partisan, you know, people have gone back to, oh, like Joe Biden was quoted talking about efficiency and Obama was quoted thinking about this process and Clinton was saying this again. And so government efficiency could be something that becomes bipartisan, could stick around for a very long time. And you suspected that this was going to be an ongoing trend? Was that part of the thesis behind starting the company?

2:14:58Yeah. I mean, frankly, part of the thesis was just I lived it and I'd seen all of these processes. It was very clear to me early on and as AI became very popularized that there were so many of these complicated paperwork approvals processes that I'd experienced and been frustrated by. I do believe that, I mean, I think that making the government more efficient is a good thing. I think that Doge is certainly, you know, is now definitely a partisan effort. But I think it's inarguable that we can make the U.S. government operate more effectively. And I personally believe that that's a really good thing to do is we want the government to be more effective.

2:15:42What I want to make sure, when these processes, though, are not fake, they're very real. I don't personally believe that you should be allowed to build any form of weapon and then sell it externally with no form of review process. These are important, complicated procedures to review. What constitutes a weapon? What doesn't? And like, these are actual questions. And those processes shouldn't just be discarded or written of as like, oh, that's unnecessary government bureaucracy. But we do want to make it possible to navigate those complicated chains without it being a three-year cycle. How do you think about, so you're creating workflows, but then, and so you guys are a software company, but you inevitably, I'm sure, run into these sort of like offline workflows.

2:16:37and is that a is that i mean i'm guessing right in sense of like hey like this was approved but we need to take it to this other building to get this other person to sign off on it and it's like i have is that is that the right yeah or can you fully digitize it you do see you definitely do see that like i can tell you that my colleague eric like some workflows are on classified networks and you can only access classified networks in certain spaces. So we are literally flying to places to go to upload forms to say, yes, we can be deployed into this network so that our software can be deployed to perform this type of document review.

2:17:19So you definitely see this challenge around there's only certain access points and you kind of, that just is the reality of it. it's gonna take a while and there's a lot of process to be automated. I mean, you mentioned like 20 ,000 pages of rules. That seems like you might need a large context window for some of these problems if you're throwing an LLM at it. What are you tracking in the progress of the foundation models? I imagine you're not using DeepSeek. Maybe you're taking a look at Llama. what are you excited about in terms of foundation model breakthroughs? Is it just IQ or is it, you know, RAG or any of these kind of buzzwords that we're hearing about?

2:18:06Yeah. So one of the challenges that we deal with is the state of the art, like the GPT 4.5, Gemini 2.5, they're really, really good at a lot of this. But in the networks that we're operating and for a lot of our government customers, they can't yet access those. So for us, there's this somewhat, it's like really an engineering tension. It's like, we can see, oh my God, this is going to work so well in two years when these models get accredited to be available in your network. But for now it's, what is the state of the art model that we can actually deploy and run in these more resource constrained environments?

2:18:43One of the ways that we're working around this problem now is, and I'm sure other, other, you know, companies at the, at the workflow layer are doing something similar is we're trying to disambiguate the the user from the we're trying to basically make it so that when like a user submits a request they don't expect an answer instantly and so if you think about how you interact with like if you were submitting i don't know have you ever renovated your apartment or anything like that yeah yeah so you probably had to get like a permit from the department of building something you submitted your your form into the department of buildings and then someone got back to you maybe three months later or something like that, that we want to compare against that type of cycle time.

2:19:25So we want to have a workflow where you submit a document and then you get an email response 10 minutes later. Oh my God, that's amazing. That's so fast. But if you're in like a UI layer of software and you upload a document and it takes 10 minutes to process, that's actually kind of a bad user experience. So we want to try and drive it to like the human review side. I was talking to Jordi about this a long time ago. like in the permitting process, I wish that there was just a button like check permit compliance in the CAD software, right? And it would just be real time. Like you can't save the file if you're in violation.

2:19:59And then you just know that, yeah, when you send it off, it's going to be approved. And that should be deterministic. And this is so much of what we actually see is that in these complicated cases, so much of the stuff that kicks you back is not some edge case question of, oh is this permitted or not it's like you just don't understand what it is you're submitting like and i apologize i have to be like a little vague when i talk about like what these things are but it's like um you just filled this out wrong and like i can tell you instantly hey you feel that you're gonna get instantly rejected say this instead and it's gonna rather than taking four weeks and requiring like nine separate signatures you're gonna be good to go and the material information is almost identical and that's that's the type of thing we're trying to push back to the user very quickly so that they understand oh this is a really hard question but this is something that you can just get a yes on if that makes sense any last questions very cool i'm excited uh we we recently went off uh our paper process so once you're starting to uh we used to print out every everything every topic that we would discuss we'd print out the tweet and read it very very complicated we're going digital yeah we're going digital too so we're glad that you're helping the government do the same.

2:21:15It's been much more efficient. It has been more efficient for us. Um, well, thank you for coming on and the work that you're doing. Yeah. Hopefully hope to see you on X. Oh yeah. I'll make sure I create an account. There we go. We'll talk to you soon. Thanks. Uh, next up we have Paul from carbon robotics coming in. Uh, carbon robotics, very cool company has built a, uh, self-driving tractor, the laser weeder g2 the world's most precise weed control and the auto tractor the most dependable tractor autonomy solution very excited to talk to the auto tractor great names honestly the laser weeder g2 combines computer vision ai deep learning technology robotics and lasers to identify crops versus weeds and it shoots the weeds with lasers it's so intelligible for uh what is probably a very complicated product.

2:22:06And it seems a lot more healthy than what, you know, just dumping pesticides. You're going to love this. I'm going to love this, guys. This is amazing. I'm going to love this. Yeah. Eliminate hand labor, cut costs by 80%, reduce labor costs, and address labor shortages. Eliminate the need for herbicides and manual and mechanical cultivation. And you want to know something fun, John? What? The S &P 500 closed up 5.7%. 5 %? On the week. On the week. That's great. Well, if you're doing well, let's bring Paul into the studio. Paul, welcome to the stream. Hello. How are you doing? How's everything?

2:22:43We're doing great. Thanks for joining us. What's going on? Would you kick it off for us with just a little introduction to carbon robotics and yourself? Man, you know what? I love this job. I love our company so much. We do some cool stuff. Our company, we build AI robotics. That's the core of our company. but we focus on agriculture and agriculture is such a great place to be so we make the thing called the laser weeder maybe you've seen videos of the laser weeder i hope you have and the laser weeder uses an ai system we find weeds live in farm fields and we kill them with lasers so that's that's great why is that good because you don't have to spray chemicals all over the place that's what i was just that was just what i was saying before you joined it's just it's a Beautiful future.

2:23:30It's a beautiful future. The thing is, like, you know, for years, I mean, we had, you know, humans have been doing agriculture for thousands of years. Maybe tens of thousands of years. And we invented this technique for growing all of our food. And, you know, we came up with things like, you know, chemical fertilizer and nitrogen-based fertilizers and herbicide sprays and things like that. That was just like dousing crops. But that's how we built our food supply. but now we're at the point where we can use ai systems to be very specific and targeted in only killing the weeds that are in the field and not having to just dump these chemicals over the place and then you know of course we have all these immigration problems and things like that and so if we can just do all this work through automation it just helps everybody yeah so little backstory i found out through experimentation that i was super sensitive to food created with glyphosates in the process, which is traditionally, from what I know, I'm not an expert, but any sort of mass-produced flour in the United States, for me, was basically a no-go.

2:24:42So I was super excited about this. I guess some of the stuff I'm super interested in is agriculture is a huge sector of our economy. It's critical, but it's also, I imagine, in some ways been, I'm interested to hear about the adoption of your products and selling these, because if you go to a farmer and they're using, let's say, these massive amount of pesticides and they're really, really cheap, and then you say, hey, I can replace this, but it's this device, you can do financing and stuff like that but talk to talk maybe about kind of like the go-to-market and what you're seeing from an adoption standpoint and and just the like receptivity to to the technology because i imagine when you go in it's you're not like you're leading probably with the value that you're offering not like hey look at this new ai system they're like well of course Yeah.

2:25:38You know, OK, so glyphosate is the main chemical ingredient in Roundup. And a lot of people have problems with with Roundup. They think some of the extra sensitivity for things like celiac disease can be kicked off by by glyphosate. There's some evidence for that stuff. But I think everybody has this experience with somebody who had some kind of food intolerance, intolerance or allergies. and then wound up going to Europe or whatever. You hear this all the time, right? I went to Italy and I discovered I can eat pasta. Why can't I eat pasta in the United States? Well, a lot of this is because, you know, I mean, they do these studies where if you were to take 100 people and you do a urine analysis on those 100 people, at least 80, 85 % of those people have glyphosate in their system at any given time.

2:26:31and so um and when we know that glyphosate is a carcinogen and like many carcinogens the question is just how much is enough to to cause a response not is it bad for you it's like what dosage and how sensitive are you we had to be clear like this this is the thing that we're using to just kill all life out of farmland right like we're using it you know to basically and it's somewhat targeted i guess this is like where gmo crops come in where gmo crops can be resistant to glyphosate, but it just kills everything else. Is that correct? That's exactly it. Yeah. So in your vegetables, onions, carrots, broccolis, cauliflowers, celery, any of those veg, lettuces, right?

2:27:14We don't have Roundup-ready vegetables. We don't have Roundup-resistant vegetables. So we have Roundup-ready versions of corn and soy and wheat. and in those fields they can just completely douse the field and the crops are resistant against the herbicide but in everything else that you eat the rounda but you know the herbicides actually really damage the crops as well and so it's also just it's not good for the soil there's runoff issues you know all of these things and but you touched on an interesting thing okay so So, you know, I have a tech background and I come from places like Seattle and San Francisco, and our farmers are in places like Salinas and Iowa and Othello.

2:28:08right and so I'm coming into these areas talking about the things that we've developed and you're right we can't talk we don't say hey do you want AI or you know what is your thoughts on the latest trends in neural nets or you know right and and and so we really are value selling and we're talking about the ways in which we can we can give them a solution to how they can keep their fields clean of weeds, not spray all over the place, and also not have to rely so much on labor, because labor is really the big hot button issue for these farmers, even more than the chemicals, because we just don't have enough labor in the United States to do all of our agricultural work.

2:28:54And in fact, almost all of it is migrant labor. And so with your labor constrained, the last thing you want to do is send people out in the field to pick that's just the least value I've had because the chemicals don't you wind up sending folks out in the field anyway and then especially if you're producing organic crops, your access to herbicides, I mean there's only a couple things you can use and so you use people a lot, you use labor a ton and so we talk to them about the way is the way we did it but that's not the selling point. And so the cool thing is that the farmers are way more receptive, inventive, looking for new solutions, willing to try things than I ever could have imagined.

2:29:48So as a customer base, they're kind of the funnest, coolest, most easy to work with, easy to get to know people. Especially, you know, I came from IT, man. If I'm going to sell to, you know, a storage CIO or a farmer, right? I'll take the farmer all day long, right? Yeah. Um, and it's like the guy who's going to write me, you know, our machines are expensive, going to write me a multimillion dollar check is like the fifth generation farmer whose, you know, name is on the building and, you know, his great, great grandfather started the farm. And so there's like history and pride and, um, wanting to, when really wanting to grow good food and wanting to make help make people people healthy wanting to provide a good product and doing everything they can to do that and so they need help they need automation they need solutions that can do stuff that isn't going to cost them a bunch of money that or at least when it when it's expensive like ours that like you can show the roi right and so here's my thing i want to make sure that all my machines pay back within three years so whatever money you outlay that should that should be saving enough money every year that by the time that three-year horizon hits, you're basically, it's paid for itself.

2:31:03Everything from then on is just brief. What is the market structure of the farming industry right now? You've heard about these mega farms kind of buying each other, rolling up, but it sounds like at the same time, you're still running into folks who, it's a fifth generation farm. Who are you selling to and what is the overall trend in the industry? Is it consolidating still? There's been consolidation. I don't know if I would say it's consolidating a lot because most of the farms that we've been selling to since the beginning of this company are still their own independent entities. There's been some private equity acquisitions.

2:31:41There's been more PE activity than there has been consolidation. I'll say that. So it's been a farm that's running very well, that's been going for a long time, that is highly profitable, that a PE firm comes in and buys and operates, but I have not seen, you know, eight farms that suddenly get gobbled up and turned into one giant entity. So I think, I think, I think the consolidation, part of the reason you get consolidation too, is that you get economies of scale with the consolidation and some of that, and what that affords you is the ability to buy more modern equipment and be at the cutting edge of technology to get the most, you know, optimum solution.

2:32:21So we kind of pride ourselves that laser weeder, our main product, is actually available to growers of all sizes. So we kind of help them be competitive, right? We help them be competitive and not have to get consolidated. So, you know, I feel pretty good about that. But I do not feel like there's been a massive amount of consolidation soon since we've been selling how how do you think about uh on what timeline is robotics really deflationary within food and agriculture right because the idea here is like on a long enough time horizon if you don't need human labor which i imagine is the core kind of cost for a lot of products you could and especially if you can make it you know a lot cheaper to even do organic food you can imagine a world in the future where organic food is priced similarly with you know whatever perfectly gmo you know grown um crops um i think what's really what's happening is it's i kind of look at it the other way around the herbicides are actually getting less effective we're getting we're getting herbicide resistance in the weed population there are weeds coming through the country right now that we don't have chemicals that can kill.

2:33:37And so that's going to keep happening, just like we have antibiotic resistant bacteria when you have ores of antibiotics. Same process is happening. So I look at if you just roll the clock forward, what do you think is going to happen in the end, right? Where are we going to be, pick some time horizon in 100 years? Well, most of the chemicals aren't going to work anymore. We're going to have severely labor constrained on ability to grow food. We're going to really care about the health of the soil because we have a limited amount of that and so how do you what do you need to do well you need to plan for a world in which you're not so heavily dependent on chemicals because they are they are getting less useful every year you need to be able to protect your topsoil which means not digging stuff up so much right one of the common techniques one of the ways in which you can help kill weeds with cultivation where you want blades through the ground, but that's bad for the topsoil, right?

2:34:31So it's like, okay, so where we have to get to is a point where we're not damaging our topsoil, we're not spraying stuff all over the place. So we're actually, I think, helping to keep things stable. The price of food is, you know, it's priced into the CPI, right? So it's one of the core components. And so what happens is, it like many of the main constituents of the economy kind of tracks with the economy right it doesn't really separate in the way that some of the more consumer goods well um so i don't know if i call it deflationary so much as being able to sustain the profitability of the farmers that we so depend on to give us healthy food how how have your customers reacted to the tariffs uh i'm sure you've had conversations or just heard kind of industry news because obviously i'm you know as much as we consume the food that we produce in the united states we also export it around the world it's true okay so the united states for the last uh three or four years has been in an agricultural trade deficit um this is relatively new for us we it's not been the case before that.

2:35:52And this is concerning. Part of the problem is that we are losing acreage to Mexico and south of the border every year because of the labor challenges and the price of things like herbicides and just sort of how much there's a lot of government regulation as well. And that makes a challenging environment for these farmers. So in that sense, maybe the tariffs kind of help. But there's a lot of negatives here, right? That it makes it, it does reduce the places in which they can sell, right? Even though we're at an agricultural trade deficit, we still are selling a lot of stuff overseas. And a lot of this equipment, you know, components and parts all come from overseas.

2:36:42Everything electronic comes from Taiwan, right? We're not, whatever you want to think about the Chips Act, right? We're not up and running, right? I can't buy GPUs that are made in, you know, Iowa or whatever, right? Right. So, I mean, realistically, you know, right, all of that stuff still needs to come inbound. And then we do a fair amount of business overseas as well. We sell a lot into Europe and a lot into Australia. So for us, that's challenging as well so this is a long way of saying we don't really know where the impact nets out at the end of the day and even the farmers i've spoken to they kind of say the same thing paul paul i don't know if this is good or bad like i can't it's hard to run the numbers right this is a multi-variate equation and and it's changing right it's like it's on it's off it's right um so i haven't heard a consistent position on this from anybody and i think it's one of these things where we nobody knows what their position should be until they understand the impacts and the impacts are like very hard to calculate right now yeah yeah it's interesting to think about seasonality in the context of the tariffs too if you're producing or you're maybe off season and the tariff hits it's like well i don't really know what that's going to do to demand until i actually pick the crop and go to sell it or yeah can you give us a little bit of a farming for dummies farming 101 on what a tractor actually is used for on a farm and then take me through the auto tractor and what that product means for your business yeah i mean growing food is setting up the land so you'll run there's a this term is generally called tillage where they'll run large blades through the ground and they're trying to tear it up turn over that soil loosen things up, allow access to the nutrients and the microbiomes in that soil, get it ready for planting.

2:38:34Then you put your seeds in the ground. So you've done tillage and then you've done planting and the tractor is pulling the tillage bar through the field and then it's pulling the planter. And then you spend your time doing fertilization, irrigation and keeping weeds at bay. And so there's a couple of different ways to do irrigation, but some of them involve laying pipes in the field and there's a tractor for that. And then when you're doing weed control, you may be running cultivators through the field that I mentioned before or a spray rig if you're using a lot of herbicides or a laser weeder and you're using tractor to pull through for that.

2:39:13And then usually when you're running harvest, you may have a harvest machine that you're pulling tractor through or some of those are self-driving as well. But there's a bunch of activities and the tractor is kind of the main engine for all of these. The tractor is the thing that provides you locomotion. And from that locomotion, you get power. So for example, tractor has a thing called a PTO, stands for power takeoff. And it's just a spinning shaft with a gear on it. And you can plug into that and then you get power. So we use that for a generator, for example, we get three phase 240 volt power off the PTO spinning generator.

2:39:50So the tractor is kind of the main, it's the main engine. It's the starting point for everything moving on the farm. And like pulling weeds, we found that one of the major challenges with getting enough done was finding tractor drivers. How do you get enough people to drive the tractor? Tractor driving is fun and it's interesting and kind of cool, right? But try doing it for 24 hours a day. When it's time to farm, it's time to farm. and for tools like LaserWeeder, in order to help that ROI really pay for itself, like you want to run 24-7. So who are you going to get to do the midnight to 8 a.m.

2:40:33shift, right? Like these things are really hard. So we started working on tractor autonomy, and the original LaserWeeder was actually autonomous, the very first one that we built many, many years ago. And so we revitalized the autonomy stack. But here's the thing. Here's the key that I feel like people didn't realize or whatever. When we were thinking about this, the big reason why tractor autonomy hasn't taken off in the world, farming tractor autonomy, is because of all the exception cases. So if I'm going to put my tractor out in the field at midnight, my autonomous tractor at midnight, I want to come back at 8 a.m.

2:41:14and see it did its job. What everybody discovered from the, call them like the existing autonomy solutions, was that in reality what happens is you put that tractor out at midnight, you could buy because it saw a pig in the middle of a field or a deer or whatever, like these things happen. And so they couldn't reliably get any of the autonomy solutions to do the job. And so what we figured was, well, what if we could remotely log in and handle the exception cases? And the problem was, you know, we investigated control methods and video feed methods. What if we put up big radio towers so we could broadcast?

2:41:54Because these are rural areas, right? You don't have good, you certainly don't have 5G, and maybe you don't have LTE. Maybe you have no way to talk to the machines at all. But we wanted to do remote observation, command and control for the exception cases. So the thing that came up that really changed the game for us was the satellite based Internet solutions like Starlink. Right. I think that is going to be the biggest game changer for agriculture. You know, other than LaserWeeder, that might be one of the biggest game changers for agriculture. And I think that not a lot of people realize it yet.

2:42:32That's very cool. That's very cool. uh what have uh the tech the tech bros uh messed up with farming i'm thinking of like you know high profile failures like bowery which was important unfortunate was that vertical farming that was vertical farming that you know was a cool attempt at you know kind of futurizing farming and then another one is like we've we've looked into like larry ellison's project in hawaii that seemingly like the probably lost more money than any farm in history. Just just on a cash in cash out standpoint. And is this, you know, clearly your solution is much more like, hey, let's lean into how farming is already done today and make it more efficient, make it better versus just reinvent.

2:43:18Like, is the idea that we just need to completely reinvent farming wrong? I think you got it. Yeah, I think that that's wrong. I think that and Bowery wasn't the only high profile vertical farming failure. There were a couple of them. We're talking about tens of billions of dollars at this point. OK, so here's the here's the fact. If I take food, grown food, vegetables, and I throw it on the ground, it will make more food. OK, so this is how things happen. So then if you say, right, so the whole conceit about vertical farming was I can put these things in these urban environments and I can grow them near to where people can eat them or whatever.

2:43:59But the economics just don't pan. It just doesn't pan out. And farmers have experience and knowledge and specialized understanding of soil health and how things grow. And there's a lot of nature involved in food growing, growing well and growing efficiency. And so this is an example of where you try to take everything apart and then closely manage all the inputs and keep track of everything required to make that thing grow optimally. You wind up doing a worse job than you just lean into the natural environment about the ways that things evolved to happen this way on this planet. And so that's so what you said was absolutely right.

2:44:42We're going to reinvent farming as the tech bros of the urban environments. It doesn't happen. And so we went out and said, how do we help farmers with their job they're trying to do instead of how do we reinvent what we think they should be doing? Yeah, it feels like one of those things like tech bros now trying to reinvent private equity with roll-ups. right? You know, private equity, super efficient industry, you know, trying to reinvent farming. We've been doing it for a very long time. Yeah. Yeah. There's this misnomer that, oh, farmers don't know what they're doing. It's like, no, they take their business really seriously.

2:45:20It's a hard business. It's super competitive, super smart people in there. It's not some, yeah. Yeah. Like the country bumpkin thesis is like wildly incorrect. I have a, I have a portfolio company that, that sells into the ag world. And, um, I've been amazed at the ingenuity at the farm level where they're just, it's like, oh, this product doesn't exist. I'm going to make it myself. And I'm going to, I have a follow-up on that. I mean, you've raised money from bond capital, uh, storied, uh, capital allocator, Mary Meeker known for investing in consumer technology during the.com boom. Um, what, what do you think, uh, venture capitalists get wrong about a company like yours where you are a tech company, but you're a couple steps away from, oh, it's a website, it's a SaaS product, it's your robotics company, and you're selling into an industry that they might not be familiar with.

2:46:11What's that process been like? Yeah, I mean, yeah. So Bond Capital, Mary Meeker and Mood Raghani, I mean, really quality people. So the challenge for us is that, part of it is that, that we're selling into a market that they're not familiar with. The other challenge is, of course, that it's hardware, right? Like hardware requires capital, it requires inventory. Some of the R &D is money you're spending on parts to try stuff that you're just ultimately going to throw away. It's a little bit different, right? But I really lean on this idea about where is AI useful for you in society and in the world, right?

2:46:53ChatGPT is cool and rad. Cursor is cool and rad, right on. But how are you going to make your life better? How can you make yourself more healthy? How do we prevent people from having the kind of food intolerances that you were mentioning? What are we trying to do really with this technology? And if we can actually make the world a better place, not in the Silicon Valley show style, but we can actually do some real things. Plus, farming is a big enough TAM that you can make a good business out of this. I think carbon robotics will get to the point where we're going to have a really nice IPO. know.

2:47:30And so that's kind of, you can plot a course to get there, but you have to be able to tell this story. And I think that maybe we had a little bit of an advantage because we're kind of half, we're like half tech, half agriculture. And so we knew how to talk in both worlds. And when we're talking to the folks in Silicon Valley and Bond and all those folks, you know, we're definitely leaning on some of our Silicon Valley tech bro roots, but to know how to tell the story, I will say. But you know what's interesting about our company? This is controversial, sort of. Look at our investors. It's Bond and Sozo and Ignition and Fuse and Voyager and Anthos.

2:48:13But who hasn't invested in us? Who's not on our cap table? AgTech VCs. We have no AgTech investment. it. So what, so why is that? I think part of it is that you had to sort of step out of what you knew about agriculture to get what we were doing. And we, you know, a lot of people, like a lot of people who knew, who were experts really like love to tell us how stupid we are to try to do lasers. Like what a dumb idea. Why would you do that? That's just stupid. You just put chemicals all over the place. What are you trying to do? Like literally. So it took, it took the people who who are used to doing bold, innovative new things and turning industries upside down to see what we were doing.

2:48:55Would you guys ever do consumer hardware? I got, my landscaper told me yesterday, hey, we need to add another day of the week just for weeding. And I was just like, honestly, very annoyed. I was like, this isn't part of landscaping weeding, but having a little robot in my backyard that's just sniping weeds as they come up, you know, it sounds like I would definitely, I'd definitely pay, you know, I'd probably pay just as much for a robot to do it as a human, just because I'd rather have a robot in my backyard all the time than... You know, I think we have a ways to go to get there. I mean, like, the point at which you'd see, you know, laser weeders, like Home Depot, I think we have a ways to go.

2:49:35And look, I'm a B2B kind of guy, right? Like, consumer is not really where I've been successful in my career. I was at Uber for quite some time, but I was working on the infrastructure, right? So I feel like, you know, and the deep learning and all that stuff. So I feel like I would love to get there at some point, but we're going to get there based on our trajectory and, you know, where we get in success on the industry side. In terms of going deeper in farms, you know, we've seen with Anduril, they build a lot of hardware, but then they're also working on a software platform called Lattice to kind of allow the autonomous system to team with each other.

2:50:13Is there an existing farm management platform that you want to integrate with? Or is that something you could see growing out of your technology? Yeah, this is a good question. And it's actually somewhat controversial right now. So we have a thing called the Carbon Operations Center. And the Carbon Operations Center shows you everything that LaserWeeder is doing and everything LaserWeeder saw. How many weeds, what type, how big they are, how many crops we saw, how big they are, what the spacing looks like. You know, all of this kind of stuff. John Deere has their John Deere command center, which you can log into as your tractor.

2:50:46Every time I have this question for folks about what you want us to integrate with, nobody has a favorite. There's been no clear leader in this space. But the reason why it's controversial is because maybe this is kind of a theme here based on some of the other stuff we're talking about. There have been a number of companies, startup companies out of Silicon Valley, who basically said, this is great. wait, I have my brand new computer science degree from Stanford, and now I'm going to go teach farmers how to farm, right, with my data platform. And they all fail just over and over and eating it, just not making it happen.

2:51:22Because it turns out that you can go collect data and show things, but unless you actually can make it actionable, there's no insight there. It's just telling people what's happening in their fields. And so having that be your business model, your primary business model, I think is a failure and it's been a failure. So maybe, you know, something like the Carbon Operations Center or maybe even the John Deere Command Center where you're kind of there to help show what's happening on some existing machines and then you can pull more in after that. Maybe that's the solution there. But there's not been a great platform for that and i would love to see there be one it's but it's different also it's going to be it's going to be mo it's going to definitely be mobile first because these farmers they're in their fields all the time and so that means phone right it doesn't mean desktop um and and you know and those kinds of things are um there's a lot of it specific to what farmers are trying to get phone and star links two major drivers this change yeah it's totally true it's completely true yeah having access to it yeah go ahead i was gonna say last question because i know we're at we're at time but uh after all the time you've spent in in agriculture uh what keeps you up at night because i feel like for americans we take food security for granted we just sort of accept it that you're going to go to the grocery store and maybe the price is up or down for eggs one day but you know there's generally going to be eggs what what scares you um i'm worried about us in the united states and many of these even many of the economies in western europe being able to continue to be self-sustaining and that means agriculture that means manufacturing ability to make things and grow things that keep your country moving along.

2:53:20In a world, we cannot compete globally dollar for dollar for labor prices. We can't. And without a drastic reduction in our standard of living, right? That's never going to happen. So if you watch all of our manufacturing leave, all of our agriculture leave, we're in a very, very vulnerable spot. And this to me should be a national emergency right the slope is the slope is already clear the the most deleterious effects haven't happened yet and by the time they do it's going to be too late so we need more ai automation or just say automation in this country to be able to help us sustain those very important things for our country growing food building products right making things that we use um You know, this is like, whatever you think about Elon Musk, this is a good quote.

2:54:13He's like, if nobody's making things, you won't have any things. Right? So it's a weird, right? So that's what I'm worried about it. That makes a ton of sense. Well, this was a great conversation. Thanks so much. Yeah, I really enjoyed it. I'm really excited about what you're doing, too. Thank you very much. I appreciate that. Huge impact. So keep up the great work. I mean, it's already been a massive success, but I'm sure you have a long future ahead of you. and you're officially our agricultural correspondent yeah we'd love to have you back and talk looking forward to the next one agriculture as it happens let's do the next one in a farm we can go get you guys out of the studio right sounds great boots boots on you you can go we can go see some laser weeders all right yeah yeah my first job was uh my uh my mom would give me a nickel for every weed of a certain you know stature that i would get out of the yard so i feel like uh i'm glad i'm happy that you're doing this, but I'm built for it.

2:55:08Yeah. I'm sure you can tell the farmers exactly how to do it better. Yeah, yeah, yeah. Yeah, yeah. We'll give them some great advice. We'll tell them what's really going on. Here's what you got to do. Yeah, don't worry. Well, we'll solve all this crisis. Anyway, thanks so much, Paul. Cheers. Have a great weekend. We'll talk to you soon. Yeah, you as well. Thanks. Bye. Bye. Yeah. If you're looking to go spend some time near a farm, maybe you should book a wander. Find your happy place. Find your happy place. Book a wander with inspiring views, hotel-grade amenities, these dreamy beds, top tier cleaning and 24 seven concierge services.

2:55:38Join 430 ,624 wanders and counting. That's a lot of people. One time, like the guest is not going to have checked out of the Zoom and they're just going to hear us do this deranged ad read and they're going to be like, why are they singing? I don't have full context on the show. You got to be singing more. It's pretty good. We should. Well, should we sing this TechCrunch article? FinTech founder charged with fraud After AI shopping app found to be powered by humans This was relevant to the conversation yesterday We were talking about how figure AI is claiming to End-to-end robotics system End-to-end AI closed system Yes, no plus plus in there Has not been fully achieved by any other company in their category.

2:56:26which makes it very impressive if they're doing it, a potentially massive breakthrough, but also a potential risk vector if they're sort of promising investors that they've achieved this, which they already have. Yeah, and this is going to be an ongoing theme, I think, with AI companies, is the claims around the level of automation. We saw this with this TechCrunch article, the CEO of Nate, an AI shopping app that promised a universal checkout experience, which makes a lot of sense. You go to the AI app, hey, buy me a bottle of Dom Perignon. It goes and finds it. It buys for you. Should be automatable with agents and operator and all the amazing LLM magic that has been in the boom.

2:57:09Nate was founded in 2018. They raised over$50 million from investors like Cotu and Forerunner. They raised a$38 million Series A in 2021, slightly before the AI boom. But they were still pitching this idea of like you can buy from any e-commerce site with a single click thanks to AI. but in reality, Nate relied heavily on hundreds of human contractors in a call center in the Philippines to manually complete those purchases. Now, no problem with that if you're upfront about it with your investors. You don't even have to be that upfront about it with the customer, I think. I don't think the customer cares that much.

2:57:43The customer doesn't care. I just want the output, which is one click checkout on any e-commerce site. That's what I'm being promised. I don't really care how you do it. You could be using LLMs. You could be using deep learning. You could be using some special algorithm. Code it by hand. Who cares? I remember this company. Yes. I remember their billboards specifically. Yes. And there was a lot of hype around it. I didn't fully get it. Yep. But I also think anytime these shopping apps launch, they tend to just launch these massive incentives so they can really run up users where they're basically giving money away for user acquisition.

2:58:14And that creates very high burn. And that makes it harder to spend that money on R &D to actually solve the AI problem that requires all the training. And there's no problem with hiring a bunch of folks in the Philippines or anywhere to do a bunch of jobs, get a bunch of training data, and then create the final algorithms. This is the story of scale AI and open AI. You know, they get a bunch of contractors. They answer a bunch of questions. They bake that all into the training data. And you get a magical experience at the end of the chain. But you've got to actually deliver that magical experience, especially if that's what you're promising to investors.

2:58:43And so this, now the DOJ's Southern District of New York is alleging that Nate relied heavily on hundreds of human contractors. And they made the claim. Well, and they found it was effectively 0 % was actually happening via AI. And the reason that this should have been a red flag is that what Nate was doing was basically computer use. Yep. Or OpenAI operator, very similar, or stuff that BrowserBase is doing, helping companies achieve now. Totally. Which even now still requires a human in the loop. Totally. It's not really reliable. When you watch OpenAI Operator do this, which is one of the best agent teams in the world, it's still like watching your grandma use a computer.

2:59:26It's sort of slow. It gets hung up in places. It will get better. But Nate was promising all this in 2021, which I imagine that if there were AI founders at the time, people that were thinking about this agentic experiences, I imagine they would of looked at this and said like, it's a cool idea and we're going to get there. But yeah. And so the company ran out of money, sold the assets, leaving its investors with near total losses. And that is the basis for the lawsuit. The investors probably wouldn't care about the involvement of humans if it was delivering shareholder value and the business is growing really nicely.

3:00:04Amazon went through something similar with their one. The bigger issue is not in the loop. check out if nate had been saying that they were using humans for everything totally and they were gonna over time use it for less and less and less it's a little bit less sexy of a story but i think it can still totally probably viable there's probably a good chance they wouldn't have raised the last round without that ai narrative yeah it's the issue totally yeah so so if the investors feel duped they're going to sue and it's going to get ugly but uh good luck to everyone involved hopefully it resolves smoothly and everyone can move on and build something cool in the future uh anyway speaking of ai brad gerstner has a great thesis i think on um yeah we talked about this yesterday basically everybody said llms have no moats there's no switching costs uh introducing memory yep effectively large context windows around all the yep sort of inputs and conversations that you have with the model So OpenAI launched Infinite Memory.

3:01:05Let's get those switching costs up. Yeah. Switching costs far too low in the foundation model space, but not anymore. Not for long. Thanks to Infinite Memory launched by OpenAI earlier this week. Somebody was pushing back and saying that you could just export your data from chat GPT. No one's going to do that. You can export your stuff from Excel too. Yeah, you can export your stuff from... People were saying this about Google Circles. Wasn't that the competitor to Facebook that they launched? Oh, well, you could just export all your data, import all your posts there. No one wound up doing it.

3:01:39If you get in the flow of just training ChatGPT again and again and again, it knows what you like. It knows how you like data to be formatted. It knows about what you know, what you don't know, what you've asked previously. It has all that extra context. It's going to be harder and harder to go back and forth. And I think people really have been going back and forth for, oh, there's a new Claude model. It's hot on Twitter or on X. Let's go jump over there for a little bit. let's jump back to the latest and greatest, kind of the model wars will maybe be winding down as the switching costs increase.

3:02:09So interesting story. This is also interesting. Base 16Z says he's pulling something out from Netflix earnings report or some type of report they did. They said engagement remains a key catalyst for Netflix's growth during its last earnings call. Netflix revealed that subscribers spend an average of two hours daily on the platform. How is that possible? I don't know how this is possible, but that's a good answer for why we put out three plus hours a day of content. Because if all those people that were basically doom watching Netflix. Cancel your Netflix subscription. Leave us five stars on Apple Podcasts and Spotify.

3:02:52But that is a lot of Netflix. But yeah, I mean, I guess if it's replaced your TV, you have it on the background. a lot of what Netflix does best is this like laundry TV. Have you heard this term? The idea that put it on while you're doing laundry. It's meant to be a lot of the reality TV shows. They kind of recap what happens at every like 15 minute interval. And so you can kind of tune out, tune in. It's not, it's not all content. That's like a game of Thrones, super layered. You gotta be following along. You gotta be reading the fan theories. A lot of it's just, yeah, you throw it on. It's, it's, it's reality TV and you just kind of let it sit there in the background.

3:03:25Anyway, this was a cool story from Steven Zhang. He built the app that someone was talking about. Preston went viral with a post saying, I want a finance app with three numbers. What I spent today, what I spent this week, and what I spent this month. Preston's post went super viral. Yep, and Steven built the app. And he shows a little time lapse of him vibe coding it probably on his couch. Was he worried about his feet? Yes. He's worried about people. He blurred out his feet in the video, which is hilarious because you could have just put on shoes or something. Maybe he realized after the fact.

3:03:59He might have realized after the fact that he didn't want to wind up on feet. Never put your feet on the internet. Never put your feet on the internet. I like this post from Joe, a friend of the show, sort of a long lost brother. This is a great one. I just threw this in because if you're in New York, you should go check this out. There is a store across from the UN in Manhattan that's called the UN of beers. which is just an amazing you know example of you can just name your store or your business whatever you want we saw the advanced manufacturing company of america it's a cool name uh there's bank of america i don't know if you've ever heard of that that's a cool name imagine coming up with that name you know yeah you're like looking around they weren't looking for domains at the of america beer of america the united nations of america but the un of beers is a fantastic name That's great.

3:04:50Well, after you go testify in front of the UN, you might need a beer. And so you head across the street to the UN of beers. This post from Pavel, I enjoyed. He said, quote, maybe if I hire someone really good at math, my product will get better. Why did this become the default Silicon Valley narrative, Lowell? And Pavel, I think you know why it became a meme. There's been someone who's been recruiting IOI gold medalists into tech companies for years. His name is Eric Lyman. Yes. And you once worked at that firm. So this should not be a surprise. Everyone is cargoculting ramp. No, it's just, I think, I think he's more basically more saying, I think that if you blend people that are really good across different domains, they're probably going to get a great output.

3:05:35Duan says people just farm PhD hires thinking 27 year olds that haven't written a single line of production code in their lives can change the trajectory of their company yeah um but uh it takes more than that mathletes are low-key goaded so low-key um i wanted to cover one final poly market before we sign off for the weekend and that is the 2025 masters winner uh right now they got scotty at 25 rory at 21 bryson at 20 and then it kind of drops off pretty hard from there that's pretty even balance kind of anyone's game anyone's game um but good luck to everyone watching masters tj was posting something earlier that that the masters was the closest american like institution or event that feels like japanese culture it's like just very like calm and pleasant yeah people are well dressed yeah there was a guy uh basically wearing a full blazer playing in a blazer that's cool uh yeah i don't know if i actually played but i saw him warming up in it yeah it's one of those things it's one of the sports that you don't need to take the athletic attire to the absolute extreme and so you can express yourself a little bit more and um and then the other one we got ufc 314 this weekend going down in miami and uh we uh who's playing uh there's a team space there's two teams yeah this is great the new york sluggers this is relevant relevant for me because um uh chandler is uh fighting this weekend and he is a Rora athlete.

3:07:11That's fantastic. He's been posting about his, his water consumption this week. So root for Chandler. We want to see the win. Uh, we want to see, um, you know, his continued success. So that's a good place to sign. What's his signature move? Does he do like a karate chop? Is that, is that what it is? Karate chop? Cause you can do anything in, in, in UFC, right? This is the, this is John's best bit that he has no idea how, how UFC works. Good luck, Chandler. I hope you make it all 25 rounds. Yeah, all 25 rounds. All 25 rounds. I hope he makes it. And yeah, good luck to you. Godspeed. Have a great weekend, everybody.

3:07:45It was a crazy week in the markets. We made it through. I think everybody made it through. One last size gong. We ended up up 5%, folks. Down 12 % if you count last week. But still, up 5 % this weekend. That's what we're celebrating today. Have a great weekend, folks. Have a great weekend. Bye.