SPEAKER_01 0:00

China is far outpacing the U.S. on manufacturing, but we still have an edge when it comes to integration. You're going to hear more about that in a minute from expert Stuart Landisberg, who is working to push robotics and artificial intelligence into the real world with his company, Seneca, a company that specializes in autonomous fire prevention. So more on that, but why are we talking about firefighting here on the AI Investor Podcast? Well, as we have discussed, we believe that robotics are the next great application of AI. But it's an industry that's likely to fall short of expectations in the short run, while I believe exceeding them in the long run. What do we mean? Well, the potential impact on our lives and our portfolios could not be more exciting here. But because robotics require inherently more complex supply chains, there's likely going to be very big bottlenecks, similar to what we have seen with memory for AI and Frontier models. And some of those will be the same bottlenecks between AI models and robots. So you can hear from this episode that Seneca uses NVIDIA chips on their devices as well as Portfolio Pick Ouster, as one example. And then there's also going to be new ones specific to this industry. There's actuators, batteries, antennas, cabling bearings, you name it. I recommend you watch Eric's most recent robotics video for a deep dive into some of these. You can find it on our YouTube channel. So the long story short here is we're really excited to talk to Seneca, which is a company who is deplying AI and robotics in the real world. It can speak to some of the challenges of trying to do that in the US, some of the opportunities of trying to do that in the US. And then we as investors can hear from people on the ground who are actually experiencing these supply chains, experiencing these AI models in the real world, and able to provide guidance about where the opportunities are and how realistic the time frame is. Just me today, Eric is not able to join. Sorry, dear listeners. I know he is the main event, but he should be back for our next episode. In the meantime, I really hope you enjoy this interview with myself and Stuart Landisberg, CEO of Seneca.

SPEAKER_00 2:04

If we don't solve this problem, the American West becomes uninsurable and therefore no mortgages, and therefore, you know, this sort of like life breakdown in the American West. My perception is that China is way ahead of us in most areas. But the area where we have the biggest advantage is that we do a good job of integration of hardware, software, and actual customer needs. The role that AI has played in design has been incredible, right? We were building prototypes extremely rapidly. And so as I was looking for opportunities to use technology to create positive change, uh happened to lose my fire insurance and start chatting with some fire teams.

SPEAKER_01 3:01

Okay, so I'm joined by Stuart Landisberg, founder and CEO of Seneca, a drone-based wildfire prevention, or should I say, elimination company. Stu is the former founder and CEO of Grove Collaborative and has done since with TPG and an international hedge fund, an all-around electric entrepreneur, and someone you should follow if you're not doing so already. Stu, thanks for coming on the AI Investor Podcast. We have discussed the future of robotics and AI quite a bit with our listeners. So we're eager to hear from someone in the field who's trying to meld these fields together, you know, as these technologies are advancing concurrently. But first, I want to talk a little bit about what brought you into this spot today. So you started Grove Collaborative. What made you decide that firefighting was the problem you wanted to spend the next decade solving? And what surprised you most about how wildfires are actually fought?

SPEAKER_00 3:50

Thanks so much for having me on, Austin. I'm grateful to be here. Um love any opportunity to talk about the incredible power that technology has given us to make changes in the world. Um I think the directions that are that are unequivocally good. Um there certainly is a lot of investment going in a lot of directions, but um, you know, speaking specifically about the moment after Grove, for those who don't know, um, Grove is a leading provider of sustainable home and personal care products. Um, sort of a consumer internet company I started in 2012. Still check it out, still still going strong, took that business public in 2022. Um and passed the reins to a very capable CEO in 2023 and took a look at what are the spaces where there's the biggest opportunity for technology to improve the call long-term habitability of the species on our planet, right? What are the what are the really big problems uh that technology has an opportunity to solve? And I live in Northern California, um, spend the better part of my life, not the majority, just the better part, uh, outdoors. And so wildfire is something that's been top of mind for my family and I for many years. You know, I watched firefighters save my uh wife's family home. Um lived through many, many, many uh bad fire seasons out here. And so as I was looking for opportunities to use technology to create positive change, uh I happened to lose my fire insurance and start chatting with some fire chiefs. And I had long thought that our fire agencies have the same technology as our military, right? If you look at the groups of men and women who keep American communities safe. Um but they don't. And there's a huge opportunity to use advanced technology to improve the capabilities of firefighters, to get to fires quicker, to operate more efficiently when they're on fires, and overall to manage the land and protect communities in ways that were never possible before. So we build uh what we call multi-mission firefighting aircraft, um, which are sort of like uh 400 strike teams, you can think of them as small swarms, 450-pound aircraft capable of uh force multiplying firefighters in a whole number of ways.

SPEAKER_01 6:08

Uh incredible and important. And you know, while it's hard to compare to the challenges that fires present in Northern California, uh I was sort of had my eyes opened here similarly. You know, it sounds like maybe you did as well. When I was in Colorado, we found ourselves in a situation where we had a home that needed fire insurance. All the insurers dropped us, we were down to one, and they said, you know what, we're not going to be renewing this next year. We have a good friend who also almost lost their cabin due to wildfires. I mean, it is it is a major issue. And uh glad that you and the team are out there trying to solve it to make our beautiful spaces more habitable. How did you build the team at Seneca? Because when a lot of founders go from one company to the other, it's sort of natural to you know, maybe pull a couple of people from your prior discipline into your new one, but it doesn't feel like uh sustainable e-commerce delivery has a lot of overlap with drones, AI, and firefighting.

SPEAKER_00 6:57

So I mean, we we did take some great folks from the last team. I'm sure, I'm sure, but into the beginning of the time.

SPEAKER_01 7:02

But there's not quite as much, you know, there you gotta go find a whole new set of people as well.

SPEAKER_00 7:06

No, it's not it's not like a totally natural transition. Um and look, I mean, I think about it's it's interesting, right? You know, you start the first company and people aren't like, well, why do you not do you not have all this expertise? And they're like, well, then you're 26. So people don't assume that you need that expertise, right? It's not that different starting this from my gosh, I'll date myself. I suppose I was probably 39 when I started it. Um but uh but I think for me the foundation is the same, which is that there are extremely capable people who want to work on problems like wildfire, which I think are among the most important for our society in the next, you know, certainly in the next 10 to 20 years, probably as soon as the next five to ten years, right? If we don't if we don't solve this problem, the American West becomes uninsurable and therefore uh no mortgages, and therefore, you know, does the just sort of like life breakdown in the American West? I don't know. Right? So this is a a truly interesting and urgent problem. And so when I went to build the team, it was really about how do we how do we find people with the correct skill sets across the various disciplines? Because, you know, robotics is like a massive comment in and of itself, and AI is, I mean, uh both of those terms are overly broad, right? And so if you look about the like at every piece of the stack, you're like, okay, well, you know, if we want this to be internet based, we need people who can build a ground control software, and you need people who can build the back end, and you need people who can build the front end, like what's the thing the customer's interacting with the glass. And then if you want to, you need like design engineers and battery engineers, and you know, just having a decent, decent sense um for what are the right starting points is half the battle. And um I think the other piece, and I probably should have started with this, is having great partners. Um I have a terrific group of co-founders. Um my co-founder runs the hardware size, a gentleman named Nick Foley, who um previously was the sort of CTO of a company called Jump Bikes, which became Lyme after Uber acquired it. Um and uh yeah, so I mean I I don't know. I just think finding great people is it's a very it's a very durable skill across across companies, across disciplines.

SPEAKER_01 9:22

Um yeah, well we'll wishing you and that whole team a lot of luck. I'd love to get to the technical stack. You know, a lot of our listeners are here trying to invest in new technologies and AI and robotics in particular, but I actually want to take a I want to hear a little bit more about Seneca and some of the major partnerships you have, like with Aspen Fire, you've tested with Calfire. So what have you learned from actual firefighters using the system that you never would have discovered in a lab?

SPEAKER_00 9:46

Yeah, it's interesting. My uh my co-founder at Grove has a great saying, which is no product survives first contact with a customer. And I totally believe that's right. Chapeau, Chris, if you're listening. Um but uh and I think that especially in a domain like fire, where it's very dangerous to light a big fire for obvious reasons, the privilege of working with exceptional collaborators is that you get to test your hypotheses frequently. And I think it's very often true that the number of shots on goal is is the single most important variable. And I think that you know uh people have now watched Elon and a number of other good hard tech founders recognize that speed, iteration speed, right? Like speed to attempt, how quickly can you iterate, is really just such a core variable. Um and ultimately iterating in an echo chamber, like or in a vacuum without the without a customer, much, much less effective than iterating with one. So, you know, we've learned an incredible amount from our partners at Aspen Fire, Calfire, San Bernardino, and a number of other departments that we worked with that we haven't announced yet. So um Yeah, I think a a huge part of being able to build the right product, I mean, which probably comes even before building the right team, is just being able to create the type of long-term high trust relationship that allows real feedback to flow. Um and you know, being a high trust, those relationships are a two-way street, right? We've gotta bust our asses to over-deliver to every one of those people so that we're a good use of their time, right? Every one of these people, especially public servants, you know, their time is taxpayer money. And so if somebody spends an hour with me, you know, that's like a hundred bucks of taxpayer money right there. Right. And so I I think that both both sides need to recognize that. And we need to show up and make sure that we bring our A game so that it's it's a great investment of their time. Because I find that our our partners are really take that stewardship of their time and their community's resources seriously.

SPEAKER_01 12:07

This feels like the type of thing that is only recently possible. So we've talked a lot about this on the podcast, and I'd love for you to maybe refute it or tell me your perspective on it. We've discussed that there's like this convergence of inputs and technology right now with AI and drone systems and battery technology that's making more real-world applications possible. So, what has changed in AI, robotics, sensors that makes this possible now that that would not have been possible five years ago? Or am I misreading that?

SPEAKER_00 12:35

No, you got it totally right. This probably would not have been possible three years ago, much less five. Um I think that we are very much still in the steep part of the curve, right? And so if you look at sort of like the rate of change of what's possible, especially in aerial robotics, you know, you sort of were like gradual increase, gradual increase, and then maybe a year and a half or two years ago, the curve inflected, but it's it's got a long way to go, right? And I think if you look at sort of sensing and controls, are we 10 years away from basically sensor inputs in, flight controller commands out of a black box model? Definitely we're less than 10 years. We're definitely more than a year or two, right, before the FAA is okay with that. But I think between here and there, the curve is gonna stay incredibly steep. And so I think we're in this really rewarding phase where you're getting infrastructure, and what we really prioritize making in America, where you're getting, you know, infrastructure, everything from the like send cut sends of the world who can, you know, cut carbon really precisely and get it back to us in 24 hours, to um we have a wonderful partnership with a LIDAR vendor named Ouster, that's American-made LIDAR, right? That's it's not cheap, but you know, good quality, I'll call it that. Good good value, if not, uh, if not cheap. Um but where, you know, 10 years ago, if you wanted an American-made LIDAR that was capable of doing the things we need to do for OA and fire targeting and nighttime operations, I mean, that's gonna be a huge and expensive undertaking, if possible at all. You know, now we just have a single great partnership. And so I think the supply chain and the number of things we have to invent uh is dropping. And for me and Nick and our executive team, I think the the real challenges are what are the pieces that we need to build that are where we can be the best in the world? And everything else, how do we find the partner who is the best in the world? Um, and look multiple years into the future and say, you know, in whatever five years, what are the pieces that we need to build today that are going to get obsoleted? Um, and what are the pieces that we need to build today that in five years are going to be competitive differentiators? Because, you know, it's uh it's both exhilarating and risky to build at the part of the at the part of the curve where the slope is, you know, extremely steep.

SPEAKER_01 15:02

I'm really glad to hear you mention Elster there. I'd love to jump to them for a minute. They're actually a pick in our portfolio because of their advanced LIDAR technology, and we've seen a lot of applications for that across robotics. So, can you just explain to our listeners what made you go with Elster and their advanced LIDAR and help us grok? Like, what other applications could it apply to? I know you know we're talking specifically about your world with Seneca, but like what makes that technology so potent so that people can understand the opportunity here that something like this enables in robotics?

SPEAKER_00 15:30

Sure. So LIDAR is basically a way that machines can perceive the get a spatial model of the world around them is probably the best way to describe that. Um we are very focused on an American supply chain. And we are very focused on allowing the aircraft to understand the space around it at all times, including at night, including if you're flying through smoke, including if you're flying quickly. Um and so for us that led us down the path of sort of, you know, who's the who's the vendor, who can hit those boxes. Um and Alistair's been a good partner for us. I think that the LiDAR technology generally, you know, you can see it on the top of Waymo's, you can see it um in a lot of different places. But it it my perception as I look out across many of, especially the fast-moving robotics, um, because LIDAR can generally do relatively good optical, um you're able, it it gives you the ability to operate autonomously and to reuse a bunch of reuse is probably wrong. Build on top of the learnings from, you know, now many, many years of self-driving. And, you know, we're able to hire a team who comes from a perception background that understands how to work with that, which makes it, you know, we're we're not gonna be we're not gonna build LiDARs in the next 10 years, right? We're gonna um buy what is best in the market, and hopefully that will continue to be LiDAR. But you know, we see it on the front of electronic, uh, excuse me, autonomous ground vehicles for the military, we see it on construction stuff, um, you know, people who prioritize the American supply chain. We do see LIDAR units in there, very often ousters, um, which are quite distinctive to you because they're spot them all over the place. What's the start looking for them?

SPEAKER_01 17:26

Thank you. Uh I I'd love to ask sort of a public markets question. So you've had a career with TPG, you also worked at a hedge fund, you've taken a company public. For investors today who are really excited by the potential of robotics, but maybe they're not technical experts rather. You know, they would they can't really distinguish between the technical capabilities of different providers. What area of robotics would you say, like, hey, I think this area is really most promising right now? Is it like actuators or batteries or the model training themselves? Or you know, and I'm I'm thinking sort of akin to like people who wanted to invest in AI a couple of years ago, if they were pressured and they understood where inference was going, right, they would have been involved in them in investing in memory, right? So if people want to get involved in robotics today, you know, just looking at the public markets, broad sectors, you know, what areas do you think are most exciting or interesting for people to consider?

SPEAKER_00 18:17

Yeah, it's a good question. And thanks for asking my opinion on this. I very much enjoy the public markets. It's interesting. You know, going through this as uh an older entrepreneur, you know, I can very see uh very easy for me to see the evidences of overbuilding that preceded oversupply in many parts of the market, right? Where the demand signal is coming. There's incredibly strong demand signal, obviously, from the end users like us and the consumers in the AI and robotics market. But there's probably even more demand coming from the middle layer that's not yet met by end markets. And you know, we'll see what happens to cost and usage over time. But it feels like a really interesting moment of overbuilding that precedes a moment of oversupply. Um but also it's totally clear that robotics is going to be a bigger part of the future. So to me, the thing that I find most compelling are, I guess, a couple of themes that are durable and not yet robust. Um I think onshoring of components and the sort of continued bifurcation of the Chinese and NATO-based supply chain feels inevitable to me. I mean, is it a good thing? Is it a bad thing for the world? Longer conversation. Uh but I think it's likely to continue. Interesting.

SPEAKER_01 19:40

I I'd love to hear about that because you um, you know, with Seneca, you you've emphasized that the system is built in America. So I want to learn about your experience there. And I'll say my intuition, not as a technical expert, matches yours from the robotics side, that there's a lot of excitement and enthusiasm right now, which is resulting in oversupply. And we've discussed that we expect the robotics investments to, on one hand, be you know, the impact globally to be larger than people appreciate, but it will probably take longer. Supply chains are more complicated. You know, you're dealing with atoms as opposed to bits. But let's talk about your built-in America point because Seneca emphasizes that your systems are built in America. What have you learned about the state of domestic robotics manufacturing? Are we as far behind as we're led to believe, or are we playing catch-up, or is it actually much more advanced and sophisticated than maybe outside of observers would have you think?

SPEAKER_00 20:28

Yeah, it's interesting. I agree with everything you said. I just think a lot of what you do, a lot of the upside is priced into a lot of the robotics companies that I look at. But um my perception is that China is way ahead of us in most areas. But the area where we have the biggest advantage is that we do a good job of integration of hardware, software, and actual customer needs. Um and you know, that type of systems thinking, I think, is actually not, you know, we're not gonna be as good at China as like how do you make a hundred widgets cheaply? We're gonna lose that game for a gosh darn long time, right? We have higher labor standards, we have stricter building codes, costs more to work here, the standard life. I mean, we lose on 10 different dimensions. Uh and there's a lot of great books written about why. And, you know, again, it's probably I don't certainly I'd rather live here than there, right? But I do think we still have a very real advantage, if not in how do you build quickly, in how do you create a system that best serves the user's need, uh, which is to say, how do you truly understand, you know, in our case, not just sort of like what do we think firefighting looks like, but what does firefighting really look like, right? And uh, you know, what are the situations that are the highest value? Not the like most sexy, but the highest value, right? We talked to firefighters about Hauling hoses up hills at night so that people don't have to carry heavy backpacks because people sprain their ankle on a hill in the middle of the night, right? We talk about backhauling garbage out from major incidents so that you don't have to use helicopters for that. And then you can use helicopters in other places, right? These are like real unsexy examples. And so for us, like uh we build a, you know, I don't like to call it a drone, but I'll call it a drone. We build a big drone. Is China better than the US to build in drones? Yes. They've never built a system like ours because the specific use cases for us are so based on truly years, um, you know, people on staff to have fire experience and the time we're doing right alongside the time in the field of customer research. And so I I think that in terms of actually valuable applications, we're much less far behind than we seem. I do think if we don't aggressively invest, we'll we'll be on the other side of it forever, right? I don't know if Europe can catch up, for example.

SPEAKER_01 23:03

So it's it's that true integration where US tech sets itself apart, right? Like that almost like the Apple advantage, right? You know, it's like exactly right. You know, making the product that really is perfect for the user use case and getting really personalized.

SPEAKER_00 23:15

But I'm gonna push back on, I mean, if you've read the book Apple in China, this is a great, great book. But it's like it's I think that the original Apple advantage was that the product was beautifully integrated and all that stuff. And I think that you know, now they can build a product just about as good as the iPhone in China. Right. Right. And so it didn't take them that long to catch up. And so, you know, I think Yeah, the catch up pace is fast, right?

SPEAKER_01 23:40

If you have a 10-year lead, it catch you catch up and now it's a two-year lead pretty quick.

SPEAKER_00 23:45

And the frontier AI models are a similar thing, right? You've got the like chips that are optimized for building models and the correct application layers and all that stuff is getting built in the US first. Um, you know, the pace of catch up there is really high. Um in robotics, you know, the the overall capabilities in China are are leaps and bounds ahead of us. Um I think for many applications, though, humanoids clearly not being one of them, but for many applications, you know, neither side is yet an expert in how you take like a very capable piece of hardware and make it extraordinarily useful in the real world. Right.

SPEAKER_01 24:25

Um, Stu, I know you have to go, so just one more quick question for me, or one and a half questions. Um can you talk about the role that AI plays in Seneca's either deployment, development application? I mean, how are you guys using AI either at the fire extinguishing level or at the development layer? Um just talk to us about that. And then where should our listeners go to learn more about you and the team and follow the story and the awesome products you guys are doing? I mean, I love seeing your videos. They're they're awesome. So I'd love to hear about both of those.

SPEAKER_00 24:52

Yeah. Well, you can Stuart Landisberg on LinkedIn or at Seneca, uh at Seneca Systems on Twitter, Seneca on LinkedIn. Those are probably the two best places to find us. Probably on Instagram too, but less active. Um the sorry, remind me of the first question.

SPEAKER_01 25:07

Talk to me about how you guys are using AI in Seneca systems, either at deployment or development. I mean, what are you using it multiple steps in the process, just for fire extinguishing, just for product design? Just talk about the role AI is playing in the tool and your company.

SPEAKER_00 25:22

So I think it's interesting. Every phase of the company is different. Um the role that AI has played in design has been incredible, right? When we were building prototypes extremely rapidly, we were able to see people with, you know, sometimes college interns be able to produce at a quality and speed similar to people with 10 plus years of experience who weren't using these types of tools, right? So on the production side, it's um it's incredibly valuable in our internal processes. Incredibly valuable. Um and I think, by the way, you know, the integration of some of the frontier models into CAD is not great yet, but you know, it's getting it's getting exciting. Um I think that will change the pace of development. In terms of on device, right, we do, we we have an NVIDIA compute unit on the aircraft itself, and there's a lot of calculation there. I think we are, you know, if we're we're already seeing like the model is not going to be able to put the thing in a field with dirt, splash retardant on it, and then like haul it in and out of a of like a mobile unit 20 times, drive across the country, and be like, cool, these are the unexpected things when a bee flew down the like uh the nozzle. Whatever, right? Like so at a certain point you hear.

SPEAKER_01 26:49

That's oddly specific. That feels like something you've encountered. Or an insane number of other weird, you know, abnormalities.

SPEAKER_00 26:56

Totally, right? Like you, you know, there's a mouse living in every piece of hardware that's ever been in a field somewhere. Um but like we're starting, I think, to see, okay, you know, you know where the ceiling is. There in terms of on-device behavior, I think everybody understands that this is very, very early. I think our job is to build the collection of data, the expertise with the correct sensor payload, and the culture that will allow us to be a beneficiary of improvements rather than somebody who's sort of like downstream stuff with legacy architecture. And so it means you basically overbuild the sensor payload today. Um, you pay a bunch of money to store his data in the meantime, and um and buy into the idea that all of those investments will grow the competitive advantage over time. And so I think you just like sort of like have to, you have to pick a side on it. Um and you know, you're gonna either optimize for today or optimize for five or ten years in the future. And you know, we're very much on the like, you know, I want this company to be companies multiple to be at its peak in 10 years, right? Not today. Um and so we very much do look out uh and try to build a culture of like uh you know, as everything from the LiDAR capabilities to the model capabilities gets better, we are we are well prepared to receive that innovation.

SPEAKER_01 28:28

Wonderful. Um Stuart, I know you've got a hard stop, so I will let you go. Um, listeners, if you want to learn more about Seneca, you can learn about their, I won't call them drones, I'll call them robotic firefighters in the side in the sky. Um, Stu, thank you for the time and thank you for the good work you're doing, keeping uh you know our our West beautiful and habitable and uh you know rooting for you guys from the sidelines.

SPEAKER_00 28:48

Oh appreciate Joss and thanks so much for having me on. Thanks.

SPEAKER_01 28:52

Cheers.