From Drive-Thru to City-Wide: Scaling an Innovative Robotic Delivery System

[00:00:00] Audrow Nash: Hey, everyone, Audrow here. Imagine getting your food delivered in just 30 seconds. Sounds crazy, right? Well, that's exactly what today's guest is working on. I am talking with Canon Reeves, the CTO and Co-founder of Pipedream Labs. They're building an underground delivery network using pipes and autonomous robots. It's like a subway system, but for your burrito if you are into robotics, urban planning, or just curious about the future of delivery, you're gonna love this episode.

[00:00:30] Canon walks us through their journey from fast food pickups to potentially revolutionizing citywide logistics. Let's dive in.

[00:00:37] Introduction to Canon Reeves and Pipedream Labs

[00:00:37] Audrow Nash: Would you introduce yourself?

[00:00:39] Canon Reeves: Absolutely. My name is Canon Reeves. I am the CTO and co founder of Pipedream Labs.

[00:00:44] Audrow Nash: Awesome. Would you tell me about some of the details of Pipedream Labs?

[00:00:47] Canon Reeves: Yeah. we're Pipedream. We are out of Austin, Texas. We've got a team of about 10 or 12 people, growing right now. yeah, we do underground delivery. So there's a lot to unpack within that. I don't know if you like get the whole [00:01:00] pitch right now.

[00:01:01] Audrow Nash: Sure.

[00:01:02] Canon Reeves: Sweet. Sweet. Yeah.

[00:01:05] The Vision for Autonomous Delivery

[00:01:05] Canon Reeves: we believe that autonomous delivery is coming at some point for our cities.

[00:01:09] it doesn't make sense for a person to be driving your burrito. and we looked at all of the sort of modalities around to do that, whether it's a car or a drone or a sidewalk robot. and we just felt like there wasn't a solution that was fully adequate, to work today. And we looked at how, do we get our utilities down, our internet, our water.

[00:01:28] or sewage, it's all underground, and cities have really been built for this. And so it led us to thinking, is it possible to have a thing pipe in the same way you have a water pipe? if not to your home, at least to your area. we started going down this, rabbit hole in like 2020, 2021.

[00:01:46] and have been at it ever since.

[00:01:48] Audrow Nash: Hell yeah.

[00:01:49] How Thin Pipes Work

[00:01:49] Audrow Nash: And so thin pipes, tell me more about that. how does, so what does the whole system look like and how are these, like what are you doing with these [00:02:00] pipes to get people food?

[00:02:01] Canon Reeves: yeah. So we lay all of these pipes ourselves, and they are basically the same material as your standard storm drain pipes. So the stuff that is like already miles and miles of this in a city, we do have to add it in. But basically, we put a small rail into these pipes, and we turn them in from dumb pipes into smart pipes, with a little that's like a train.

[00:02:23] So there's really not much to the underground part of the system. We try to keep it as simple as possible. To be a utility, you really have to keep it simple, Make it something you can always maintain and continue to upgrade.

[00:02:33] Audrow Nash: Yeah, for sure. So with the tracks in it, I'm thinking like little train tracks for these to go on. is it very similar where you just have, you have your point of contact and I think I saw on videos that they're rubber wheels, which is probably how you grip them. but then you maybe have a little inner wheel on the inside of the track that keeps it on, so you can just like accelerate and go, is this how it is?

[00:02:53] Canon Reeves: yeah, it's like a roller coaster. So there's several sets of wheels that keep you defined in all degrees.

[00:02:59] Audrow Nash: [00:03:00] Oh, so you have little like, clasps wrap, wrapping around the tracks in a sense, so that it doesn't fly off if it's going real fast and there's a, bump or something.

[00:03:08] Canon Reeves: Yeah, and then you've got some that kind of keep you right side to side. And so the robot, from the robot's perspective, it's just go forward and stop or go backwards, but the mechanical system really keeps it aligned.

[00:03:16] Audrow Nash: Okay.

[00:03:17] Challenges and Solutions in Last Mile Delivery

[00:03:17] Audrow Nash: And so how did you, backing up a little bit, so how did you decide food delivery and how did you decide that this was the approach? You mentioned looking at drones and looking at several other technologies, but I'm interested, was it food delivery that was the main thing that you wanted to do and that drove finding the solution?

[00:03:38] Or did you find the solution and go to food delivery or how, did we arrive at this?

[00:03:43] Canon Reeves: So we really, we're much, we think about last mile delivery as a whole, not just food delivery. So food delivery is a kind of a good common one because it's one of those few categories that is express So within the next 30 minutes of you placing that order, you really want it and you're used to getting it. but we see it [00:04:00] in grocery e commerce, anything, cross country. So Not the pipes, but just, packages, general parcel. so yeah, anything last mile, the reason we focused on that is that's where a good 40 to 50 percent of the costs of a delivery are in that last mile.

[00:04:14] Audrow Nash: You say 40%,

[00:04:15] Canon Reeves: yeah, 40 to 50,

[00:04:18] Audrow Nash: but ballpark, it's it's a, substantial part of the whole shipping cost is the last mile cost. 'cause I get, when you're shipping things over from overseas in a giant shipping container, that's you, have so many goods. So the cost per good is very low. And then when you're doing it in warehouses, are really efficient, the routing, the big trucks to send them around.

[00:04:41] And then, so you have to keep getting smaller and smaller. And then you get to these little trucks with an individual driver driving around the little truck. and so then that cost is, you're saying like a huge percentage of the total cost of the whole shipping. Probably because it doesn't have good economies of [00:05:00] scale for this kind of thing.

[00:05:01] Canon Reeves: yeah.

[00:05:02] And there's some categories like parcel and Amazon, they're able to batch and go once a day and get enough things on the route. That

[00:05:08] Audrow Nash: volume. Yeah.

[00:05:10] Canon Reeves: Exactly. And you've got to get, you got to have that volume, but when you look at food delivery or any kind of like on demand express delivery, you don't have that luxury of being able to fill up an entire van.

[00:05:18] You've got just your item. And so that's where the economics really break down.

[00:05:21] Audrow Nash: I see. So when you have the van, even though I was saying it's less efficient, which is probably true, it probably is still a good bit of the costs, they still get some economies of scale by having so many to deliver in the neighborhoods that they go through and everything that it's worth it. whereas it's very hard if you're doing like one DoorDash or dry carrying one sandwich, this kind of thing, that's where it's very expensive.

[00:05:43] I would imagine that gets really expensive very easily. Okay.

[00:05:47] Canon Reeves: Absolutely.

[00:05:48] Audrow Nash: Okay. So you're thinking more it's, more parcel delivery, but you are looking for something that has urgency with it and doesn't have a good solution, and that's how you arrived [00:06:00] at food delivery.

[00:06:01] Canon Reeves: Yeah. And we don't, I don't think anybody here really thinks of ourselves just as food. we've got some beachheads in food as a category, but it's, yeah, we're really going for the whole pie.

[00:06:10] Audrow Nash: Hell yeah. Awesome. Yeah. On the website I saw mostly food. but I suppose it makes sense. Yeah. It's a beachhead in a

[00:06:18] Canon Reeves: And yeah, to give light on that too, we, one of the things we realized as we started to go down this path is there's two kinds of nodes in a network like this. There are supply side nodes and then demand side nodes, which is like your homes, office parks, hospitals.

[00:06:32] Audrow Nash: Yep.

[00:06:33] Canon Reeves: On the supply side, you've got a few different categories.

[00:06:35] You've got your warehouses, your restaurants, your grocery stores. and we started to look for ways to provide value in those supply side nodes ahead of building the full network. Because, the company is called PipeDream. We know this is going to be crazy. It's going to be a lot of work to go install all these pipes.

[00:06:49] we wanted to see if there's a way for us to add value and start working with these supply side partners, start generating revenue, and start adding nodes to the network, ahead of the full network.

[00:06:58] Innovative Instant Pickup System

[00:06:58] Canon Reeves: What we landed on is [00:07:00] this product that we call Instant Pickup. Instant Pickup And it's basically like a bank tube, but for food.

[00:07:04] So if you could imagine placing your fast food order on your phone, pulling it up to an automated kiosk and getting your food in 30 seconds and going. If we can make that experience to get that fast food so much faster, get the customers to download the app so they're going to be ordering more. It's going to be a more convenient experience.

[00:07:20] we can start to install these. These are the pick up locations across the country, way ahead of these networks, and start to generate revenue and start to work with these partners. And the key is that it's all the same underground technology. You're just making it a little bit different of a portal.

[00:07:35] and so that's where a lot of our kind of food delivery stuff goes. That's where most of our early partners are, are in the quick serve restaurant space.

[00:07:41] Audrow Nash: Very cool. Yeah, it seems to me like, the more that I've been doing this, like interviewing different startups and things, it seems like the real challenge is the sequencing of how you. Generate revenue without investing too much. Because if [00:08:00] you say you were to like, just try to raise billions of dollars to lay tons of pipe down all at once, that would be maybe like, maybe if you guys were the best at raising VC money and things like this, like maybe, but then it's like you might be a little off, it might not be the best market fit.

[00:08:17] It might be, I don't know. Who knows? Like the idea could be slightly different or it just could take years before you're profitable. So you can't, it's like you're at a disadvantage for that, but you can grow through finding these good ways of going through it, that it makes it so that the investment is probably more on your terms.

[00:08:36] You get to develop the things piece wise, which is just way better. so it seems like you're taking that approach and so you're starting with food and you're starting with restaurants, like where to, just to make it so I understand that. Is it like. We'll have a Wendy's and Wendy's drive through gets crowded.

[00:08:57] So what we'll do is we'll make it so that the like surface area of [00:09:00] where people can go and pick up food is bigger. So we distribute it. and you have these robots just send it out to them, wherever the people are. It'll say, it'll be like an airport pickup, but for food kind of thing. Is that right or

[00:09:12] Canon Reeves: Yeah, exactly. and it really turns, fast food model on its head a little bit. Most of these locations and restaurants are built around their drive thru. It's where a lot of their revenue comes through and the drive thru is all about convenience. Customers want to get there, get the food and get out quickly. With our system, we can now do a disjointed drive thru and anywhere in their parking lot, it doesn't have to be just their kind of area. It could be an entire strip mall anywhere in there. I think it plays this portal. they can now create an order point where customers can more quickly get their things.

[00:09:41] a big part of that is because we have automated storage, connected to the store. So when you place your mobile order, it waits until you get within a certain distance, say, you're five

[00:09:50] Audrow Nash: Ooh, that's cool.

[00:09:51] Canon Reeves: they'll start making the order, they'll put it into this automated system. And so when you pull up, the food is still fresh.

[00:09:57] But it's just 30 seconds to get it over to you and then you can get [00:10:00] underway so we can create a faster experience and drive loyalty to these restaurants and do it in a way that's like every restaurant could be different where they put this portal, how many they put, it's just totally up to them because it's all just, routed underground.

[00:10:14] Audrow Nash: That's really cool. And then you have a very convincing way of selling it to them, which is more efficient. Like you, you get to focus more on the kitchen, you get to focus more on producing orders. It's really cool if they have the app and the app has the location of the person so that you can make it so that when they're close, so it's warm right when they get there.

[00:10:31] That's really cool. You, mentioned you have partners with this?

[00:10:35] Canon Reeves: We do. Yeah. I can't name any names. but we have a few different brands that we are working with and, working on a pilot right now. So

[00:10:42] Audrow Nash: Gotcha. Okay. Really cool. And then so that strike, so with that approach. Why not just use bank tubes that are air pressure driven? because it seems like the scale of the [00:11:00] networks that you need are smaller, and that seems to be something that's very well thought out. Maybe robots are more scalable and easier to add new like robots to the system.

[00:11:10] but why, not just go with one of these, like more convention, make it a bank tube set up for these fast food

[00:11:17] Canon Reeves: yeah, a lot of

[00:11:18] Audrow Nash: why haven't they done that themselves, I suppose too?

[00:11:21] Canon Reeves: Yeah. Yeah.

[00:11:21] Audrow Nash: it seems like it makes sense.

[00:11:23] Canon Reeves: Yeah. Yeah. It's, you would think so. I think it's really, when you think about the bank tubes, how they are, they're routed at multiple, directions. If you imagine like your burger and your shake being inside of that, it's like down, like twisting.

[00:11:35] It's it may come out like a salad and

[00:11:37] Audrow Nash: ha

[00:11:37] Canon Reeves: spilled everywhere. our system, we've got A pretty large tote that holds your order and we maintain the orientation, control the acceleration at all points. so your, food is going to be delivered in a good quality. and that also allows us to do some things like automating the storage.

[00:11:52] So when you get there, you get it faster.

[00:11:54] Audrow Nash: That's really cool. Okay. I really like that. So yeah, you can control for the accelerations, you have a lot more [00:12:00] control than just sending a tube through a big vacuum for this kind of thing. And that makes it so it's a lot better for food delivery 'cause you don't get like this giant mixed thing of everything you ordered.

[00:12:10] Canon Reeves: yeah,

[00:12:11] Audrow Nash: sense. That's super cool. And then how does, so how does that allow you, I guess the robot is similar, the piping is similar and then having it move towards last mile delivery.

[00:12:26] Scaling the Network and Future Plans

[00:12:26] Audrow Nash: tell me about that. 'cause that seems like quite a, there, I'm sure there's additional problems there. tell, me how that's going and how you're negotiating those problems.

[00:12:37] Canon Reeves: yeah, I'll go with a really simplified model of it and then expand it out. If you imagine, this fast food restaurant that is five minutes away from you and, you're on your lunch break and you're like, I just want to get this and get out quickly. I really care about being a fast experience. probably going to go there instead of driving to another part of town to go to a different restaurant, even if [00:13:00] they're both like the same speed, or if there's a restaurant, you want more. If it's too far away, you don't want to go there. With our system, we envision a world where we, we build this sort of expanded drive through in the parking lot.

[00:13:12] But if you imagine expanding that out, maybe just a couple miles downtown, and you go from an area that's like really commercial area with a lot of restaurants, and you move it to more residential area. So now if you're in that residential area, and you're thinking to yourself, Where do I want to go eat?

[00:13:26] What's going to be the fastest? If that's what you're looking for right now, this portal is almost like a drive through for that, location that's further away. So you can order and get your thing here and go pick up, and it's just going to be a much faster experience because now maybe you don't have to cross a major road.

[00:13:40] Now you don't have to go sit in traffic. You're just going to the edge of your neighborhood. if you imagine And then you add a couple other things to that system. If you add grocery items, if you add some convenience store items, if you add a few more restaurants, once you get the customer used to using that, when they are thinking about what they want to [00:14:00] buy next for lunch, they're going to go to that app that kind of has all of those items.

[00:14:03] And they're more likely to buy from that because it's closer and more convenient. You still have to drive and go pick it up, but it's the same thing as if there was a drive thru for that restaurant, just all the way over, so anyways, that's how we, it's a very simplistic model of one of the ways that this network scales beyond just the parking lots of these

[00:14:20] Audrow Nash: Yeah,

[00:14:22] Canon Reeves: and it, you have to approach it. yeah,

[00:14:24] Audrow Nash: it's like having more locations, but you also can have. I don't know, a Wendy's and, taco Bell and a local restaurant and you can have them all routing into the same network 'cause they're often like clustered it seems for these kinds of things.

[00:14:38] And then you can have them all go to the same end point and at your end point you can even distribute. so that it's not like one drive through, but like 15 drive-throughs.

[00:14:48] Canon Reeves: totally.

[00:14:49] Audrow Nash: Yeah.

[00:14:50] Canon Reeves: you can just find places where you can find places where there's like abundant underutilized parking areas like strip malls, and now you can add things to them. And so you can change [00:15:00] the value of real estate. Like instead of opening up a new food location at that area, you just have to plug into this network at any point in that

[00:15:07] Audrow Nash: super cool.

[00:15:08] Canon Reeves: And now you get all these outputs.

[00:15:10] Audrow Nash: And so that's like intermediate growth that's making it so you drive to a location, but that location can be outside of where there's major traffic. It can be close to population centers. It could be both. It could be other, different constraints that make it so how you cho or that dictate how you choose.

[00:15:25] And then from there, I suspect you can just keep growing the network, so that people, it gets closer and closer to people's homes. and cl like it could be in their neighborhood, but then it could be in their, to their house if it was like, you guys have expanded enough for this kind of thing.

[00:15:43] Canon Reeves: Yeah, and I think it's interesting because you start to see network effects on both sides. So the more demand side nodes we add, the more customers we make this convenient for, the more value this provides to the businesses on the other side. But the more businesses you add to the other side, the more options those customers have.

[00:15:58] And so the more they're purchasing through that [00:16:00] system, And so the goal is that you get this sort of like organic, not organic, it's a lot of planning, but you get the sort of network that expands over time. and as cities like continue to build and add on, whether they're building apartment complexes or neighborhoods, you start to really plug into those from when they're being constructed.

[00:16:17] and that's when you get that in home experience and you also get delivery. So if I. Put this portal near you and let's say you don't want to go pick it up. let's say you're a parent, you're watching your kids, like you don't want to go get in the car. You can still have a delivery driver make that final delivery, but

[00:16:31] Audrow Nash: But they just pick it up. Yeah.

[00:16:33] Canon Reeves: it's cheaper because you're just really, close.

[00:16:35] Audrow Nash: Huh. It'd be funny if there was like a gig Uber in your neighborhood or like a gig DoorDash or something in your neighborhood where you can pick up the package for you and your neighbor and you get a dollar or something like that if you drop it at their thing. 'cause you're already going.

[00:16:53] That'd be a, that'd be super cool.

[00:16:55] Canon Reeves: it'd be really cool. And you could mix demand, like you could have your Amazon package in that, you could have somebody's [00:17:00] food order, you could have somebody's parcel, and because it's all coming from one point, you can really batch it there.

[00:17:05] Audrow Nash: Yeah. Now I see a few things that are really interesting and I'd like to get your thoughts on. the first one to me is the size of the tunnel for this kind of thing. 'cause looking at the videos, it looks like they're like a one foot diameter pipe. Maybe they're a bit bigger, 18 inches or something like this.

[00:17:25] Is it true or am I,

[00:17:26] Canon Reeves: So the initial system was 18 inches. We actually started at 12, then went to

[00:17:32] Audrow Nash: ha, I've seen videos and I guess that's what I've seen.

[00:17:34] Canon Reeves: yeah. but we actually ended up moving recently, to a 24 inch system. So we started working with some grocery partners. Yeah. And it's, it's, it's more dirt you have to move. It's more pipe material.

[00:17:46] So there are, there's an added perfect cost, but when you're digging up a trench like this, it's not that much bigger to add six inches. so that's where we went to.

[00:17:55] Audrow Nash: And it probably moves it and eventually, you might have a three foot one or something like this. [00:18:00] but okay, so you're increasing the size. I'm just wondering it seems like I, just talked with a company, qbi, hey, and they are shipping you a factory and shipping containers, like those 40 foot ones.

[00:18:13] And so the thing that is interesting to me is it's like the easiest way for them to move what they're shipping is to just conform to these super efficient transport methods already. And I wonder if you guys like had a incredibly distributed network and you chose everything to be a 12 inch pipe, it'd be like, okay, you get your, you get a pizza, but it's cut in thirds and then they're all stacked or something like this because it's real efficient to send or something like this, but you're making it bigger and that makes it so more goods are be, are able to be sent more easily, which is great.

[00:18:49] Canon Reeves: Yeah, you're, that's the exact reasoning why we started with the smaller sizes. our take was that as delivery becomes near free and almost instantaneous, you are no longer [00:19:00] batching and getting a whole shopping cart's worth of things. You're getting just your individual items.

[00:19:03] And so we thought, yeah. And that's going to be cheaper to lay. And so we'd started at that approach. But what we realized is that for this to work, it's a huge coordination problem between like supply side partners and cities and landlords and construction companies, and just, and all these people.

[00:19:18] And if you are,

[00:19:19] Audrow Nash: in some

[00:19:20] Canon Reeves: Tons of stakeholders. Yeah. And so if you're not able to conform it and address existing problems as like an existing customer behaviors, it makes those conversations much harder. And so we opted to size up the pipe that way you can fit, your full thing of toilet paper, your full thing of paper towels, anything that you'd really need at your house coming from a grocery store, which opens up just such a large market for us.

[00:19:40] Audrow Nash: F for sure. Yeah. And there's not really anything like it. And I suppose that a big reason there's not anything like it is, 'cause it's really hard to do with this, with all of the, because you, guys have to work with all the stakeholders and you have to like, you have to find a good pipe solution and you need some [00:20:00] Robotics knowledge to make all this work.

[00:20:01] So it seems like a super hard problem. And if, you tried to do like the boring company thing where it's super deep and stuff, now it's super expensive to do 'cause those are enormous tunnels. But if you just lay it like a utility that's very clever for this kind of thing.

[00:20:15] Technical Challenges and Solutions

[00:20:15] Audrow Nash: Tell me a bit more about the piping.

[00:20:19] and so you said it's sewer piping, which is probably a big advantage that it's not gonna let water in, I would imagine. so that makes it easier for your systems. go ahead. what, I guess tell me more about the piping.

[00:20:33] Canon Reeves: Absolutely. Yeah. So we use, storm drain, corrugated storm drain pipe, and yeah, it's, in previous

[00:20:41] Audrow Nash: it means like cardboard, where it's like you have the back and forth going between the two edges. So it's like bouncing between, so you get a little bit of insulation, or what's the point of corrugating it? I suppose

[00:20:53] Canon Reeves: More

[00:20:53] Audrow Nash: flex.

[00:20:54] Canon Reeves: you want to minimize material and maximize strength because you're putting these under roads and things. And so there's a lot of [00:21:00] weight kind of compressing these pipes. And so by doing,

[00:21:03] Audrow Nash: with all the triangles from the corrugated.

[00:21:05] Canon Reeves: There's like an inner layer of this kind of separate one, like you said, that has these ridges to it, and so that gives you your strength. Yeah, that's it.

[00:21:13] Audrow Nash: And so laying these, you're just, is it hard to do in cities? So I'm assuming that, I think we talked, before and there's a city where you guys are building the city with these, Like you're building the city as you're laying these pipes in the case, right? Or is that something, or am I imagining

[00:21:37] Canon Reeves: It's something that we've talked to a few partners about, kind of new build mastermind communities installing. We've actually done a retrofit in Peachtree, Corvus, Georgia, just outside of Atlanta, about a mile.

[00:21:48] Audrow Nash: Hell yeah. So tell me about, so the retrofit, I just digging in like New York City, I feel like they're like completely full, underground. there's no [00:22:00] space for this kind of thing, but maybe other cities it's a little bit different. Tell me about retrofitting.

[00:22:04] Canon Reeves: Yeah, we really thought that it was going to be a nightmare to do that and permitting and,

[00:22:10] Audrow Nash: ha. That would be my guess, too. Yeah.

[00:22:12] Canon Reeves: Yeah, I'll touch on permitting and then construction because they're very interconnected.

[00:22:17] Audrow Nash: Uhhuh,

[00:22:18] Canon Reeves: by using existing utilities Methods and materials.

[00:22:23] Permitting has actually been pretty straightforward for us. Cities are really comfortable with these kinds of things because there's not a lot of opportunity to cause harm to citizens. It's really out of the way. The construction, you don't want to shut down roads and sometimes it's annoying to have construction in your city, but people are really used to that.

[00:22:38] Audrow Nash: Oh, it's all the

[00:22:39] Canon Reeves: not Yeah, it's not if you wanted to deploy a self driving car, like there's a very high standard there because you have the ability to inflict harm on citizens. We just don't have that. And so the permitting has been pretty straightforward for us, especially on the smaller builds that we do.

[00:22:54] We're not doing anything that's hanging above people. We're not, we're just adding stuff underground. So people don't really care as much [00:23:00] from the

[00:23:00] Audrow Nash: is it very expensive to add it underground? Like I imagine so I'm imagining, I don't know where it'd be. Maybe it's on the side of roads or something. at where it's underground, but it's not covered by asphalt, or maybe it is covered by asphalt and you have to rip up the road to do it. Is it expensive?

[00:23:17] Is it difficult? I, don't know much about this.

[00:23:20] Canon Reeves: Yeah, we do it in the city right of way. So just to the side of the sidewalks. And so you're really just tearing up grass and where you can, you've just got a backhoe and you're just trenching. so there's two kinds of ways to install the pipe. Trenchless entrenched. we prefer trenching.

[00:23:34] it's a little bit cheaper. but basically you open up the earth, you move the dirt, there's other utilities. You work around those and you go do scans before you deploy so you know what's there and where. but yes, it's pretty much like dig a big hole. Put some stuff in it for backfill, put your pipe in, put some more backfill in and cover it with dirt. There's your pipe. It's in the ground.

[00:23:54] Audrow Nash: Okay. Hell yeah. And is it, I suppose that if you were to do it in New York [00:24:00] or San Francisco, it'd be more expensive than doing it in Georgia or I don't know, Texas, like San Antonio or Austin. tell me a bit about that. 'cause I guess all the regulations and things like this, or maybe even existing, systems that you might have to work around.

[00:24:16] Canon Reeves: Yeah, absolutely. when you're in those more crowded places like New York and SF, like you mentioned, I think those would be some of the last locations we go to. Yeah.

[00:24:26] Audrow Nash: being like a hub of innovation and whatever, it's so hard to get anything done. So things arrive there last, like my, I'm in San Antonio now, close to you guys. and we have super fast internet and I did not when I was living in San Francisco, which was really funny. It kept going down all the time.

[00:24:42] 'cause flaky systems, but like the tech hub kind of thing doesn't get the tech for a long time, which is so funny. Okay. okay, so those are some of the last places that you're likely to go. it's cheaper and easier to do at other [00:25:00] places that have maybe less regulation or less existing systems. and then you just, you lay down the type pipe.

[00:25:08] How does the, one thing we haven't talked about yet is the end points. so where people pick things up, what does that look like?

[00:25:16] Canon Reeves: Yeah. So we call them portals and I would argue it's about half to two thirds of our development efforts go actually to the interface points. the underlying Transportation is pretty simple. for every customer that we work with, we create, every customer segment, rather, we create portals that are specific to that kind of need.

[00:25:36] Audrow Nash: What is a customer segment for this? You mean like a use case for this? Where it's like these are drive-through picker uppers is this kind of thing. Is this a customer segment?

[00:25:44] Canon Reeves: more so a quick serve restaurant, grocery, in home, warehouse, the, like on the supply side, different segments.

[00:25:52] Audrow Nash: I see.

[00:25:53] Canon Reeves: and really where the differences come is like the volume of totes and the way the customers interact with them. So [00:26:00] a fast food store, they're really only sending one tote over and the customers are it's like a pretty immediate thing.

[00:26:05] whereas with grocery, they're sending several totes per order and the customers are used to getting out of their car and loading their stuff. So it just changes based on how the customer is going to interact with the system. and then it also changes based on how the workers are going to put things into the system. for our QuickServe restaurant system, we have a very small, we really try to get as small as footprint possible, because the kitchens are so tight back there, that's the input, so a worker can put in food into one tote at a time and send that tote, we'll go store it, and then when you output, you have a portal that presents it, it lifts the tote up and presents it out to you.

[00:26:41] And so when you think about somebody handing you something through a drive thru window, there's somebody kind of leaning over and handing and you're leaning over and grabbing. We move totes around. that's the base thing in our system. And so we just lifted a tote up and made you like reach over and be a really bad experience.

[00:26:55] And so we put a lot of effort into, okay, we're going to lift that tote up and we're going to extend it out to you. [00:27:00] We're going to make it easy to grab. So we have to put a lot of effort into. Making these systems like safe and easy to work with and like a pleasant kind of pick up experience

[00:27:09] Audrow Nash: Yeah. Very ergonomic

[00:27:11] Canon Reeves: of time.

[00:27:12] Audrow Nash: kinda what they'd expect. That's such a funny thing. 'cause it's not, it's not something you would necessarily expect when, or it's not something I would expect when starting out on a problem that a large amount of, I guess, it makes sense when thinking about it for a little bit that it's okay, you need to make it so users really like using it.

[00:27:30] 'cause if it's super inconvenient, like you, you basically wanna be way better than the alternatives, because the alternative is driving for them. And if they have to do a super painful maneuver to reach it. Deal with some like clampy jaws around it that are frustrating or something like this.

[00:27:48] and then they'll just go drive the five minutes to the restaurant rather than taking your pipes. Interesting.

[00:27:54] Canon Reeves: Yeah. And it's, and you think about the number of different customers and [00:28:00] types of people that this has to serve and like use cases for how far they can grab it. And, it quickly becomes Driven, like that drives up and ends up driving a lot of things. And that's why we call them portals. we want this to feel magical.

[00:28:12] It is like a really cool thing to get your stuff, delivered underground. It's invisible. Yeah. And so we want to like, we want to make sure that we're building that, that DNA and that muscle as a company of building great user experiences, because at the end of the day, it's gotta be a human putting it into the system.

[00:28:27] And it's gotta be a human taking it out at some point. So you gotta make it great on both sides.

[00:28:32] Audrow Nash: How do you balance that with, being a startup with limited resources and, like how do you make sure it's not premature optimization in a sense? Because you could fixate completely on this and you could bankrupt the company making the perfect user experience, or you could solve a painful enough pain point, which it sounds like you're doing, that customers deal with the crappy jaws that clamp around it and make it like they have to fight for the food a [00:29:00] little bit or whatever.

[00:29:01] How do you balance that?

[00:29:03] Canon Reeves: Yeah, we, so for one, we try to hire just killer engineers who of thinking about things almost as a product owner, as much as they do an engineer. And Typically for these portals, we only have one engineer designing the entire thing. And so they've got to be, both really fast and really good as an engineer, but they also have to really care about what it's like to use this thing.

[00:29:22] and the second thing we do is we really push, we push for 70 to 80 percent customer experience in the hardware, and then we plan to improve the rest in software. And so you want to get all the main axes that you're going to end up having, and you want to have all of the screens and sort of the accessories planned in and designed in, but you don't necessarily need to get those to attend from an experience level.

[00:29:44] When you ship, so you just

[00:29:46] Audrow Nash: that's clever.

[00:29:46] Canon Reeves: motion there and then fix the rest of the software. So we'll tune that kind of motion control when it comes out to you. So it feels nice and we'll tune the opening of the door. But for now, just make sure there is a door that opens and make sure there is a thing that comes out.

[00:29:57] Audrow Nash: That's so clever. Yeah, you can shove it [00:30:00] into the software, all that tuning and the fine behavior. I really like that. Building the capabilities, making it so you have access to the things you think you're likely to need, and then you can really refine from there. And it's very low cost and pushing out the changes is very easy and you don't have to swap big parts of every single portal that you have and this kind of thing, which I would imagine gets very expensive very quickly for this kind of thing.

[00:30:24] Oh, okay. I like that a lot. I feel like that's such a clever thing. and. J just to push it all into software. And I know we've seen that a lot with a lot of companies are doing that with a lot of products, but it's, I don't know, something about that's really cool to me. Hell yeah. So you have these different portals, these portals have different interfaces depending on which customer segment.

[00:30:48] How does it, with the robots, I am imagine, I think we talked earlier and it was that you have like loops that the robots go [00:31:00] on to get to a spot. So you don't have them like trying to go opposite ways down the same track. tell me about the, like routing them and how do you switch tracks and all the details of getting them from A to B or C or D or E or whatever, all your end points.

[00:31:18] Canon Reeves: Yeah. Yeah. The cool thing is. is that we've made the system to be just very modular and can be arranged in so many different configurations based off like what is optimal. Like when you think about it we're making a product that can work really a machine that can work at the size of a parking lot or a full city and every parking lot is going to be different and so you have to be able to adapt and fit into these things and so the only way to do that in my mind is to reduce the base component of that system.

[00:31:49] Down to the most simple things

[00:31:50] Audrow Nash: Love it. Hell

[00:31:51] Canon Reeves: once you do that, you can really mix and match these things and you almost get it, you get it to a point where even the customers themselves are able to mix and match these components to get what they want.[00:32:00]

[00:32:00] Audrow Nash: Huh.

[00:32:01] Canon Reeves: but yeah, the base components for the transportation, like a locomotion system, you've got pipe and rail and robots go straight. But from there, all you really need are like turns and then junctions. and junctions are just, small switches where the rails, what changes rather than the robot. So the robots have no concept of turning. They're just going,

[00:32:22] Audrow Nash: only.

[00:32:23] Canon Reeves: yeah, Forward, backwards, stop. that's all they got. that's some differential when they're going around turns, but, yeah, you basically make the track that can adapt.

[00:32:31] and you do this all within standard utility basins. So we have these like big plastic basins that, We drop in anywhere you want to make a turn or you want to make a junction, and there's a manhole cover. And so these are powered components, these junctions, but they're really easily accessed, and serviced by going into that manhole.

[00:32:50] So

[00:32:51] Audrow Nash: Is it is power. I don't know anything about being in a manhole and, the. How is it hard to get power to them? [00:33:00] Do you, I could imagine, you don't need that much power 'cause you're just switching tracks, but, so I'm imagining like you could even have batteries that you swap occasionally as a very simple, not very scalable solution for this kind of thing.

[00:33:14] But is it, can you like tap into other utilities that are down there? Or you break a wire that exists that's already passing power if it's underground and then you, I dunno. Power it that way and then, or I, don't know. How, do you get utilities to those junctions?

[00:33:31] Canon Reeves: yeah, when you're going citywide, there's a couple of different options. I'm thinking about as if building a new building and then tapping in, You can also, on our shorter installs, we just run Conduit with power and data in it. So once you open that trench up, that's most of the cost is just opening that big trench and then filling it back in.

[00:33:48] So if you wanted to throw in a little two inch thing of pipe with just power in it, it's not a big deal.

[00:33:53] Audrow Nash: that's cool. I. Okay. Yeah, that makes sense. So you just basically have long wires that go the whole length of [00:34:00] it.

[00:34:00] Canon Reeves: And

[00:34:00] Audrow Nash: don't need, go ahead.

[00:34:02] Canon Reeves: I'm sure that'll change as we go to the city wide, but most of our installs right now, we're just really like parking lot size. So it's easy.

[00:34:08] Audrow Nash: Yeah, I'm, wondering if you had to go for like miles, if there would be transmission losses or something that you'd see on the electrical side, because like they have boosters every now and then for electrical ones. I don't know how frequently they are.

[00:34:20] Canon Reeves: Yeah, I think we'd have to do something a little more complicated and probably tie into different points. I'm definitely not an electrical engineer, but,

[00:34:25] Audrow Nash: I am, but I don't know anything about this,

[00:34:27] Canon Reeves: it's, but yeah.

[00:34:31] Audrow Nash: Okay. So that's a solution for now, basically. yeah, I could, and I could see you could try, to place your junctions very intelligently so that they're close to. locations that you have power and things like this. So that is say it's like all of, there's a junction at each, restaurant.

[00:34:51] You connect all of them in a loop or something and so it can decide to go to the restaurant or not this kind of thing. And you just have the restaurant power it because you're adding that infrastructure anyway. [00:35:00] Okay. And so you have that little pipe. So you have the big pipe that has the rail that sends the robot down.

[00:35:07] You have another pipe for your shorter ones where you just have another pipe inside of the other pipe that, has power so you can have junctions for it. That's really cool. So those are powered and you can get switching. How do you, we talked about this before, but not on this interview and I don't remember the full answer.

[00:35:27] How do you handle fail failure cases? Your robot burns up and a wheel falls out, wheel fail falls off in the middle of the track. how, do you handle this kind of thing?

[00:35:37] Canon Reeves: Yeah,

[00:35:38] Audrow Nash: And also comms with the robot too would be interesting to talk about too.

[00:35:41] Canon Reeves: yeah, those are the fun ones. you just have to assume that it's going to break. it just, it is going to break at

[00:35:47] Audrow Nash: Especially if you have enough of them, you'll be, your success will be that you have robots breaking in bad locations because you're running them so frequently for this kind of thing.

[00:35:56] Canon Reeves: Yeah. We think about two kinds of failures. so one [00:36:00] is like an intact failure where you just, lost comms, like a reset, some kind of power issue. But it's not destroyed. The other is, two robots collide for whatever reason, and there's, it's not intact, and it can't be pushed out.

[00:36:12] For intact failures, where it can be pushed out, our robots are designed where they just literally passively push each other, all through motors and electrodes. So they're very simple mechanical systems. And so if one dies, you just have another that pushes or pulls that other one out. and then you just take it to an access point and at that access point, you can go recover it.

[00:36:31] And we've had versions of systems where you put in like planned failure areas to where there's a junction that just pushes the robot off. So you just want to get it off that main track.

[00:36:41] Audrow Nash: so you can get it later at your own leisure in a

[00:36:45] Canon Reeves: Yeah, so you don't have to wait on the human operator. and then outside of that, it feels like a really catastrophic failure, like something insane happens, which, those will happen.

[00:36:52] the pipes aren't big enough to where you can go inside of them. We have a little cart that you can ride on, so you fix So yeah, as crazy as that [00:37:00] sounds, it ends up just going, you just end up going in there.

[00:37:02] Audrow Nash: It sounds like the most claustrophobic thing I could possibly imagine. That sounds great. Just like sitting on a little, what I, don't know, like one of those little, I don't like a luge kind of thing and you're just zipping under the ground. That'd be crazy. Okay.

[00:37:19] Canon Reeves: yeah, we,

[00:37:20] Audrow Nash: Uhhuh.

[00:37:21] Canon Reeves: We have an 18 inch system in Atlanta, and so that's a lot tighter, but we've had to go in there and fix things on that, and it sounds really scary, but when you get in there, it's just such a fixed environment,

[00:37:30] Audrow Nash: But you're like, all

[00:37:31] Canon Reeves: get so used to it.

[00:37:33] Audrow Nash: that's wild. Okay. Hell yeah. Yeah. I don't, I'm like, I don't think of myself as claustrophobic, but the idea of doing that is terrifying. being like under pipe for, say, say it's like a big one, where you're in a city or something, and it's like a half mile between points.

[00:37:49] Like you're a half mile enclosed by pipe. oh my God. Underground, no seals. That would be nuts.

[00:37:57] Canon Reeves: Usually you put access points [00:38:00] pretty

[00:38:00] Audrow Nash: more frequently,

[00:38:01] Canon Reeves: Yeah, you really want

[00:38:02] Audrow Nash: would you put 'em?

[00:38:03] Canon Reeves: I would say every couple hundred feet. It's really like low points.

[00:38:06] Audrow Nash: okay. So you want have a half mile stretch with nothing.

[00:38:10] Canon Reeves: no, really 500 feet is a pretty long stretch.

[00:38:14] Audrow Nash: then that's less terrifying, I would say. 'cause being that exposed would be nuts. And you don't, do you ever, do you ever get, I know they're sewer pipes, so they're probably quite good. I. but do you ever get like water in 'em or, I don't know, other undesirable what?

[00:38:29] Canon Reeves: Yeah, never, through the pipes themselves. So the pipes that we use have been rated for a hundred years plus. And when you're, when they seal, like we fuse seal them. So it's all it's all, it's as if it's

[00:38:40] Audrow Nash: chance or ba basically no chance. Like you need a car to crash into them

[00:38:44] Canon Reeves: Yeah. Yeah. Or it's really roots. Like when a root grows into it or something, which we haven't ran into yet, just because it's Our systems haven't been in for that long, but it'll happen.

[00:38:52] Audrow Nash: huh? I guess that makes sense. Yeah. They'll burst the seal. Okay. So if you get water in your pipes, how does that work?

[00:38:59] Canon Reeves: [00:39:00] Yeah, you're gonna have to plan, just like robot failures, you have to plan that water will get in.

[00:39:04] Audrow Nash: Yep.

[00:39:04] Canon Reeves: you put vaults at low points and you put automatic pumps in them. And so if there's a, there's like an air bladder, it's, if that gets lifted, we'll send some signals and say, oh, there's water in the system, but also we'll just start pumping it out and injecting it into the ground.

[00:39:15] So

[00:39:16] Audrow Nash: I like it. Okay. Hell yeah. That's a good failure mode for this, that you're bas I like that you're already planning for this kind of thing and that you have pumps that can remove the water and things like this. And then you have a sensor that lets you know about it, so then you can, take corrective action, but you don't have to, put down your system or anything

[00:39:36] Canon Reeves: absolutely. And, it sounds like we're smart and come up with these things, but we're just stealing ideas from utility companies. that's how they all work. You can go in Home Depot to even buy these exact same pumps, because these are just such common things, which is just, it's really nice.

[00:39:50] So we've got a civil engineer on the team and it's so cool for her to just be able to pull things from other solutions and then apply them to us.

[00:39:57] Audrow Nash: So cool. It's like software reuse, but

[00:39:59] Canon Reeves: [00:40:00] Yeah.

[00:40:00] Audrow Nash: hardware stuff, which makes sense that they have scalable good systems. And also like sewer pipes. I'm sure every sewer is using these. maybe not this exact one, but like pretty similar. There's a lot of lessons learned from all of humanity, having sewage for this kind of thing.

[00:40:16] So yeah, it's really cool when you can tap into existing systems with this. I feel like that was very clever to do that. Hell yeah.

[00:40:22] Canon Reeves: Yeah. It's really nice to have multiple vendors that you can use. There's multiple kind of economies of scale. And it's, an interesting market to borrow from because there aren't the same, I would say it's very commoditized, but not in the same way as maybe like raw metal or raw, plastic.

[00:40:42] there aren't as many pressures in construction world to get priced down. You're really working. with you sit with like city clients and so there we definitely see a lot of room for us to take what is existing and really optimize cost and optimize process especially when we're doing kind of one one [00:41:00] style of pipe all the time so we see there's a great starting point with what's there but we see a lot of room for us to improve

[00:41:06] Audrow Nash: Oh, I'm sure. Yeah, it's an old industry and so it's been doing things for a while. And then I also think that a large cost, like talking to the QB people that I mentioned earlier, they, 70% of the cost of building a home, if I remember correctly, is labor in most places. And other places it's like bonkers.

[00:41:28] It's, even higher for that. You go to California, for example, it's probably even go to San Francisco. It's probably much higher than that. It's 90% but so if most of the cost is labor, which I think is a fair assumption, then why would you care about your materials being priced high?

[00:41:46] Because most of the cost is labor anyways for this kind of thing. It's like the 20 20% went up to 25% for the thing. It's eh, no one feels that really.

[00:41:56] Canon Reeves: Yeah. Ours is interesting because when you do such long [00:42:00] systems, we talk about costs on the per foot basis. And so you're really like, you're really starting to pull out, And a lot of that material selection starts to actually matter a lot. it's the, honestly, the cost is pretty well spread between the pipe, the rail, the trench, and then you have like repair costs.

[00:42:15] So if you go through a sidewalk, you've got to repair that sidewalk. But then, like you said, you've got labor spread throughout all of that. And these jobs are so big, even for a modest city network, there's so much money there, that the economics work out pretty well for us to, bring a little bit of those operations in house and you

[00:42:33] Audrow Nash: I would

[00:42:33] Canon Reeves: save some pretty good amounts of money.

[00:42:35] Audrow Nash: I would think so. Yeah. 'cause you guys, you're interested in scale because your system gets way more valuable if you get larger economies of scale for this. And so then that means you're gonna be doing a lot of the same type of process. You're gonna be laying a lot of pipe and then, Yeah, so I, it would make complete sense to me to optimize around that. and I bet you that you guys [00:43:00] can find a lot of really good a, as you're saying, you can find good efficiencies for that. You can pick good materials, you can reduce your cost per foot. and I imagine, actually, I think that'd be really cool, is if you guys develop like this amazing way of putting pipe down and then you can sell to all the utilities ones because your process is so much more efficient than their process for all the piping like that, actually could be a good, eventually your company does this.

[00:43:25] Canon Reeves: Yeah. Yeah. I think it's, we started with okay, what's the bare minimum just to get to market and really go. And now as we're starting to get more and more into the way we work, we've got over a hundred pre orders for these locations and some bigger products. So we're actually starting to get some volumes here, hey, let's go find these areas for us to go and work now.

[00:43:46] and yeah, it's just nobody else has really had the same pressure in the industry, so we can't really shake things up.

[00:43:51] Audrow Nash: And I would imagine that, there's a lot of, I, I can't, I'm, I haven't thought of this too much, but I bet you that [00:44:00] there's a lot of other opportunities for pipes. if you're really good at pipes, I'm sure that there's lots and lots of things that you can do that would bring value to a lot of people.

[00:44:14] just, moving things is obviously one of them, but I wonder what other things there could be, obviously normal utilities, but seems like you guys are gonna be in a really cool spot as you figure this out.

[00:44:25] Canon Reeves: yeah, I think we can go to City. It's something that, when you're retrofitting utilities, what they've done in the past is basically like, we're opening up a trench. we're just, it's just taking a big trench. You go to other companies that are like the power link or anybody else, any of the other utility companies and say, Hey guys, we're going to open up this trench.

[00:44:41] Does anybody want to put stuff in there?

[00:44:43] Audrow Nash: They're like,

[00:44:43] Canon Reeves: does anybody,

[00:44:44] Audrow Nash: Oh, we've been waiting for someone to ask. Yeah,

[00:44:47] Canon Reeves: so cities, cities are really pushing to bring all of the power lines underground. And there's always going to be utilities that get to go in somewhere. And so our hope is that as you really do those bigger projects, you coordinate with all of the parties in that area and [00:45:00] you, line up to where you can, and that reduces the cost for us because now we're getting somebody else in

[00:45:05] Audrow Nash: other buy-in. Yeah. You might make a profit on all parts of the process. I feel like. Yeah. I feel like it's I don't know, Amazon with their ebook or, let's see, what's, a good example? I am thinking of air Airline, I guess airlines and credit cards. It's like they're, have you heard the thing where it's like, airlines have planes, but the main business is credit cards for this kind of thing?

[00:45:31] Or Amazon sends packages everywhere. But their main thing is AWS the web, web services for all the server stuff they have. it'd be funny if you guys ended up like you're a robot company and that gets used to scale, but then you end up just being a pipeline company because you can do so much benefit so much, you can add so much value by doing that eventually.

[00:45:52] And you still have the robots for all the delivery and everything like that. And that would be a sizable, I don't know, part of it'd be [00:46:00] a sizable, profitable part of the business anyways, but

[00:46:04] Canon Reeves: Yeah. I think there's going to be. there's going to be a lot of ways that we create value and it's really lasting value. if you have a pipe going into your neighborhood or a type show restaurant, that can stay there for as long as that neighborhood restaurant will stay there itself.

[00:46:20] these are, a hundred plus year systems. And I think there's going to be a ton of value created by the system. And I think you're spot on that. We're probably going to find some different areas for us to extract other value. So whether that is us becoming, and selling, whether that's us like selling things on that system and plugging in directly, or, value added services off the pipe install, there's a couple of different areas for you to pull that value, but it's, I think the main thing is just to start with creating that value, making something that people really want to use.

[00:46:51] Audrow Nash: Yeah. And the clever thing to me is that the, sequencing that you guys are doing, where you're starting with fast food, by just basically giving them a bigger drive [00:47:00] through, and then you can connect a few. Companies or a few fast food places, or a few restaurants to these and effectively make it easier for customers to pick up.

[00:47:10] and then you can keep branching out. You can go to grocery stores and you can go to normal package delivery like normal parcels. and you can keep going and keep building and reuse it all, for sending anything around. and then while doing this, you're building valuable expertise and something that I imagine most people don't wanna do for this kind of thing, which it's the problem you're solving is really hard and that's why you haven't seen many people come into this space, is what I imagine.

[00:47:40] It's a good moat in a sense. Just the difficulty.

[00:47:43] Canon Reeves: It is. I've definitely been, I had started and sold a company before this one, and I was in the educational robotics space. So I was making robot kits for kids. it's, it's, there's a ton of competitors. And so it's really hard to be

[00:47:55] Audrow Nash: Oh yeah. A hundred

[00:47:56] Canon Reeves: all that stuff.

[00:47:57] And, I felt It's a lot harder to raise money [00:48:00] for. It's a lot harder to

[00:48:00] Audrow Nash: Definitely. It's a commodity.

[00:48:02] Canon Reeves: Yeah, And this has been such an eye opener for me to come work on a problem that is so difficult, but so rewarding at the end. and what's cool about it, it's like the right kind of difficulty. It's not, one of our software engineer, Mac, he co founded Voyage, a self driving car company.

[00:48:18] And it's such a different, like the problems you're solving are so different.

[00:48:22] Modular Design and Scalability

[00:48:22] Canon Reeves: there's not this like long tail of problems for us to solve. It's really once you solve this at a small enough scale, you build the base system, you get good enough instruction, you can really go scale this.

[00:48:32] Audrow Nash: Totally.

[00:48:33] Canon Reeves: because it's just, there's, not really this like long unknown.

[00:48:36] it's pretty cut

[00:48:38] Audrow Nash: de-risked it pretty hard.

[00:48:40] Canon Reeves: And it becomes a spreadsheet problem. It becomes a, okay, there's this many customers here, this many restaurants here. This is how much pipe we need. This is when it will pay off. This is the loan we can get. Yeah. cool, let's go do construction and then go do a ton of work after that.

[00:48:53] Audrow Nash: Yeah, I think what it is to me, so you're saying you started this company, you have other people starting autonomous car companies at your [00:49:00] company. what it is to me that I've seen is a lot of Robotics companies, when they try to go into a different domain, they end up having to be a CEO or CTO or whatever it is.

[00:49:11] They have to learn how to do a robotic startup, and they also have to learn how to do something in whatever that domain is. So there are really two companies fused at once. whereas you guys Pipedream to me, you guys are solving a simple problem. I say simple, it's a like incredibly difficult problem, but it's one problem.

[00:49:31] And then what you have to figure out is just the distribution of it. How do you roll it out? And it's what it feels like is you have a bunch of good options now and you, get to pick what you think is the best one. And so that's like your risk for this. Like it's how efficiently you can possibly scale versus.

[00:49:50] Oh my God, is the technology gonna work? We need this critical part of technology and oh my God, we need to find the market fit. we still, it's 'cause those, you just have multiple problems at once [00:50:00] that are both existential to the company very often and you guys are just building something that clearly people want.

[00:50:06] And your robotic systems are pretty simple. And, I think that's awesome.

[00:50:12] Canon Reeves: Yeah. Yeah, man. It's great.

[00:50:14] Customer-Driven Development

[00:50:14] Canon Reeves: And I think the, my, my co founder, and our CEO, Garrett, he is so great at sales and building relationships with these partners that we have. And what we found is we prefer to sell ahead of building. before we ever started really designing in a detailed way our Instant Pickup product, we had several customers who were like, okay, we want this one, can we get this?

[00:50:38] And that allowed us to develop it with them, but also meant that okay, those sales are there. we just have to go do this like technical stuff. and to me, it takes problems that are existential. And brings them inside. So it becomes, can we make it? Not can we sell it? And to me, that's like the way to run a hard tech company, any company.

[00:50:58] It's I'd much rather [00:51:00] have pressure. Customer wants this, get it done faster than is anybody going to buy this?

[00:51:04] Audrow Nash: Totally. And get it done faster too. 'cause you're burning out of money. That's what I mean, the two problems for this kind of thing. Yeah, this is, what you guys are doing is like literally the advice of many people that I've talked to where they're like, sell it first, then build it kind of thing.

[00:51:21] it just makes things so much easier and you know you have a good problem then too, so you don't have to worry about that. 'cause you already have buy-in from customers for this kind of thing.

[00:51:29] Canon Reeves: And you, the customers we work with, they all have thousands of locations.

[00:51:34] Audrow Nash: Yep.

[00:51:35] Canon Reeves: And so it's like you get that big account and you still have franchises to deal with. And it's not, it's never easy, but you're just selling to that one big customer and then. It's a lot easier to get to get continued sales after that.

[00:51:47] Audrow Nash: Hell yeah. Yes. Sounds super exciting. And I also, I really like when it seems like a company can diversify itself or provide another service. Like when you have something big and amazing to grow [00:52:00] towards, like that's super exciting. So you guys helping other companies lay, pipe and infrastructure, like that kind of thing, that's a huge value that you guys could tap into.

[00:52:13] And it's really cool because it's like, how would you get there if you, wanted to just raise venture capital and get there right at the beginning, it probably wouldn't work. But because you're able to grow in this way, that you're building your own pipes and you have a very good use case for it, and a lot of people are already paying for you to do it, they're pulling for you to do it.

[00:52:33] You get to scale towards that to become like almost a utility company, which is super cool.

[00:52:39] Canon Reeves: Yeah.

[00:52:39] Small Teams, Big Impact

[00:52:39] Canon Reeves: And we've always viewed, we think underground delivery is inevitable. Like we think that there will be a delivery utility in our cities. And we always felt like it's time in the market that will be the determining factor of if we get to be the ones that do it, because this is. This is such a hard coordination problem with so many stakeholders.

[00:52:59] And there's so many [00:53:00] things that have to come together for that network to, that first network to go well, and for the fly wheel to start spinning. That if we burn cash too quickly, and we don't make sure that we're here to see those relationships through, our odds of failing are higher. And so we've always Raised ahead of time, hit metrics ahead of where you're supposed to, and we kept the team way smaller than we should have.

[00:53:23] Like to the point of it's definitely been painful at times, but

[00:53:26] Audrow Nash: Yeah.

[00:53:27] Canon Reeves: hiring only amazing people, you can pull it off. And so you get this really nice long runway, plenty of time to figure the things out and go do sales, and you do it with fewer people and you do it with better people.

[00:53:40] Audrow Nash: Yeah, I've, one of the really cool experiences that I've had working, or actually an internship when I was in grad school is I was an intern at a company called Haldi, which now they're one x they're one of those humanoid companies and the company was like 20 people or something like this. And everyone had a [00:54:00] silo in it because it's like they're a little fast moving company and they were able to move so fast.

[00:54:05] And it was so fun working with them because you can see this like really small team outperforming huge companies that are doing, similar initiatives. And it's just, it's I don't know. So I imagine that's how you guys are with your small team. You have really good people. 'cause they had really good people too.

[00:54:24] They probably still do. I haven't talked to 'em in a bit. but yeah, it's amazing the velocity that you can get with good people working on a problem and they're really focused on it and it's a very clear problem too, which is

[00:54:37] Canon Reeves: Yeah. Yeah. and it's cool because, you just have to get those right, getting those right people's heart, like hiring.

[00:54:45] Hiring the Right People

[00:54:45] Canon Reeves: This has been the most difficult piece has been hiring by far to me. but once you get those right people, and if what is the right level to optimize to the question you hit on earlier of like, How do you know when to stop optimizing that customer experience?

[00:54:58] If you get good at not going [00:55:00] too far, not going too little, finding that right balance for the time of the company, you can just keep everybody moving so quickly. And people, because there's so few people, we're all really close to the customer and the customer's needs. it just means the whole team has brought it on, making high level decisions and understanding like, Where we're going as a company, like I'm huge proponent of smaller teams doing more.

[00:55:21] Audrow Nash: yeah, I love that. I think actually. I'm not a hundred percent sure on this, but it seems like an exciting thing. with all the AI stuff going on and with how distributed many companies are. Like, so for example, if you need something, you can hire it out from a company and you can grow a relationship.

[00:55:38] Like the, circle of concerns of your company could actually be small and you could have a small fast moving team that outsources other things with other teams. and then it's almost like a lifestyle startup in a sense, which just seems like the coolest idea to me.

[00:55:55] Canon Reeves: Yeah, no, it's super interesting. I, I've really come around to using contractors. We've got, we've [00:56:00] grown our envelope of like specific mechanical contractors for software. And so it's if you need that one kind of thing, that's tightly scoped, you allow one engineer. to manage five contractors now that engineers output is multiplied as long as those contractors are also good enough to like not be added

[00:56:16] Audrow Nash: yeah, you have to find good contractors. So it's a challenge in itself, but then you don't have to have that wing of the business, which is cool. Which real, it's like using a software library is how I think of it. I'm not gonna write num, pa myself, I'll just use num pa or any linear algebra library.

[00:56:30] It's I'll just use that one. I'll just use this contractor that does this. We don't have to have that expertise in house. We don't have to do the hiring or the coordinating with the lots of people they can do that. They're doing a good job, this kind of thing.

[00:56:41] Canon Reeves: absolutely.

[00:56:43] Audrow Nash: Hell yeah.

[00:56:43] Remote Work and Contractors

[00:56:43] Audrow Nash: What have you. So for hiring, how have you, how do you go about finding these great people?

[00:56:50] Canon Reeves: We've done a mix of honestly, Twitter has been our best source by far.

[00:56:56] Audrow Nash: that's what I find. not that I'm hiring, but I am [00:57:00] amazed at the quality of people that I meet on X now, but, I. Yeah. It's just, it's unbelievable as a network. LinkedIn, everyone just likes stuff. x it's like real substantive conversations. Really great connections. Blows my mind,

[00:57:14] Canon Reeves: Totally. You want to find you want to find those you want to find those like cool people who are just a little bit different. they're not just A Mechanical Engineer. They're a Mechanical Engineer plus these other things. And it seems like X is like really, it's great at servicing those people that want to do unique stuff and like to follow this kind of unique stuff. really our key for our hiring process has just been flexibility. I really believe in forming the role around the person. And so once

[00:57:39] Audrow Nash: Oh, I love

[00:57:40] Canon Reeves: that, you know, we only, there's a lot of people we interview that could be good and they seem great engineer by all respects,

[00:57:49] Audrow Nash: Yeah.

[00:57:50] Canon Reeves: but you meet them in, They do the work, but they don't have that kind of wow, factor that you're like, oh man, we have to have you.

[00:57:57] And so we take this approach of we're going to hire [00:58:00] slowly. We're going to look at a lot of people. And until we see somebody that is a hell yes, we

[00:58:04] Audrow Nash: I love

[00:58:04] Canon Reeves: them where we are mistaken that we don't hire. And it's painful for sure. Especially on the engineering side, it's like the stuff that we've got to get out, but it's no, we've got to wait until we get that right person.

[00:58:15] and it's, so rewarding when that, those people come together and that team comes together.

[00:58:18] Audrow Nash: I love that. Yeah. I try to do that in life. every decision it's do I wanna do this? It's hell yes or no kind of thing. so it's like a 95% is no and this kind of thing. Yeah. Hell yeah. I love, that. I wonder, that was from a book too. did you pick it up from a book or did their company that

[00:58:37] Canon Reeves: I'm sure I heard it from somebody and just parroted it.

[00:58:40] Audrow Nash: Yeah. What a funny thing. Yeah, I love that though. okay, so you find these people, they're killer people. oh, one thing that was really interesting to me on X, I put out, I've been putting out polls to try to understand my audience a little bit more. I asked a question if people are [00:59:00] founders of Robotics companies want to be founders, or have they already started a company?

[00:59:06] Do they have plans to start a company? Do they have no plans but still want to start a company or no, they're not interested And, I don't know if it's biased to who sees the question or my audience or something like this. I, like, it's highly specific to who's following me, I think. But, I think it was like 80% of people were interested in starting a company.

[00:59:28] 20% had already, or were already founding a company for this kind of thing. It wa it was like 40 or 50% that ha had definite plans or already were a founder. It was just bonkers. It was like,

[00:59:41] Canon Reeves: That's

[00:59:41] Audrow Nash: talking two founders by posting on X, which is just the craziest thing.

[00:59:46] Canon Reeves: and that's all we want, it's as founders, that's all you want. You just want to see other cool shit. You want to meet other founders. You want to hear people like, you don't want the LinkedIn

[00:59:55] Audrow Nash: Ah ha.

[00:59:56] Canon Reeves: Like you just want to see people building cool stuff and you want to support other [01:00:00] people.

[01:00:00] Like I love seeing other founders build things.

[01:00:02] Audrow Nash: Ah-huh. And I think another cool thing with it is a lot of times, even if someone wants to start a company, they may realize the best next step for them is to join a company that is already promising so that they can get that experience, meet the good people, develop their skills. And I think it's a, it's very much the process that you see.

[01:00:24] So it's just such a cool thing. and someone who wants to found a company, I think a lot of times they are doing a lot more than others because they're preparing for this. and so I think like there are a lot of really good people on there, which is just the coolest thing.

[01:00:43] Future-Proofing with Modular Systems

[01:00:43] Canon Reeves: Yeah, with something we've used, almost as like a sales pitch and as a bonus of the office. So of working here, we are so supportive of founders and we've had two people already come and start companies. One of them is still working with us while moonlighting and working on another company.

[01:00:58] But, yeah. Our pitch is [01:01:00] with really killer people, you don't have them forever. They are going to go to that next thing. At some point, a lot of these people want to go start companies. So our pitch is come here, we're going to small team, work on stuff you really care about, have a fun time in your day job.

[01:01:14] but we're going to help set you up to go start that next thing. So whether that's like letting you use tools, space, making connections to our investors, hyping you up on, on Twitter, Whatever that thing is, let's help you come here for a little while, then go start that next thing. And that for us, like that unlocks talent that is just hungry.

[01:01:31] Audrow Nash: Yeah.

[01:01:32] Canon Reeves: but it also just sets them up for a great next thing.

[01:01:35] Audrow Nash: For sure. I think that's so cool. That's such a neat perspective. 'cause a lot of companies, to me, have a needy perspective if it makes sense. Where they're like, you come, you join us, you don't look at any other companies, you don't like, stop doing any side hustles, stop doing whatever. And you just see this and it's but how am I gonna be really good?

[01:01:55] It's I'm gonna try all these ideas for this kind of thing. And I love that you guys are [01:02:00] embracing it where you're like, we fit each other. As long as we fit each other, come do great work for us. We'll try to help you as much as possible. it's, almost like the best approach that you can imagine when dating or something like this, which is just so funny to me that to hear it from a company perspective and it makes a lot of sense and it seems very good to me.

[01:02:22] Canon Reeves: Yeah. Yeah. And it's we're not going to have the cafeteria providing you every meal like Google is, but we can be that launchpad for you. And so it's you just got to own your strengths and

[01:02:32] Audrow Nash: Totally.

[01:02:33] Canon Reeves: into it. Yeah.

[01:02:36] Audrow Nash: like I've heard from a number of startups that they have a hard time recruiting because a lot of big companies, just like a lot of the ambitious people, and maybe they're not founders per se, but a lot of the ambitious people may go to the big tech companies or something like this.

[01:02:55] They'll go to Amazon or Apple or whatever it is. [01:03:00] and so I, and also, those companies will offer pretty crazy compensation packages and things like this, which it's a bit harder. I imagine you wanna run a little bit leaner because you want people, one, you don't want to take out too much. like you don't wanna trade too much equity for money.

[01:03:18] and so people are probably paid a little less than they would if they were at Amazon or whatever. but so how do you heal? How do you deal with competition from larger? Companies, and like the perks they might offer or maybe it's the wrong type of person. I dunno, what do you, how do you think of it?

[01:03:37] Canon Reeves: We got really lucky that most of our stuff is like mechanical. We've got the majority of our engineers are mechanical engineers, and there just aren't really as many positions like that in these kind of big companies. And

[01:03:51] Audrow Nash: you're right.

[01:03:52] Canon Reeves: so software is really hard. if you're a robotics company, that's very software heavy, which is most software robotics companies, for sure.

[01:03:59] You're [01:04:00] competing with those people. But for us, like we've got one software engineer

[01:04:03] Audrow Nash: Wow.

[01:04:04] Canon Reeves: yeah.

[01:04:05] Audrow Nash: I imagine as you guys grow, you'll need more software. Like you said, you're putting the complexity in the software, but I guess you'll be in a better position for. That. because you'll have raised more, you'll have more of a track, whether you record, you'll be more promising. You'll be able to get investment on your terms, and if you're getting investment on your terms, you might be able to pay more without cutting off your arms or legs or whatever for that investment.

[01:04:30] Canon Reeves: Yeah, and you can contract some of it too. So as long as you have the core, as long as you have the core robotic software, orchestrator, all those things, we've got a really, killer software engineer. We actually are hiring our first,

[01:04:43] Audrow Nash: Exciting?

[01:04:43] Canon Reeves: our second software engineer. Very exciting. so yeah.

[01:04:46] But yeah, for the most part, we're just competing for mechanical engineers who are coming from defense companies, from boring company, from just all kinds of places. And they are harder to find. I would say it's much more rare to find that sort of like [01:05:00] full stack mechanical engineer who is just

[01:05:02] Audrow Nash: What's, what do you mean full stack? Full, stack mechanical engineer. What does that refer to? what, is their

[01:05:07] Canon Reeves: it's yeah. It's really somebody who's really good at motion system design, really understands electrical and how to integrate sensors, really understands like sheet metal structures and understands enough about manufacturing to know what are good methods. You can find people who are, they've just been doing ejection molding for the last 15 years, or you can find people who are used to building like big steel structures that never move, but finding people who can like, do all of the sorts of mechanical that we need to do is a bit harder.

[01:05:34] And really, we can teach most of the stuff. What we really look for is people who are just,

[01:05:39] Audrow Nash: Hungry.

[01:05:40] Canon Reeves: love to ship. yes. And you find that softer, because the barrier to entry for shipping is so low that you can find people who just are, they're ripping stuff out. But in mechanical, these iteration times are so long.

[01:05:50] And a lot of people are working at defense companies where they can't even share what they're working on. It's yeah. So finding those people who are just, they, they really, they love the ship. They have a good portfolio [01:06:00] of existing work where you can look at it and know that they've done a lot of great work.

[01:06:04] It's hard to find surprisingly.

[01:06:05] Audrow Nash: I am not surprised at all actually, but just because talking to a lot of people that are interested in the mechanical side of Robotics, it's it's hard 'cause it's, you, see a lot of companies, they need the hardware people until they don't, until they have things stood up. And so it's a lot of like work now and then maybe not work later and so they move around from job to job for this kind of thing.

[01:06:31] And there's not many of them because as you said, most there, there's tons of robotic software engineers and there's tons of demand for Robotics software engineers. but it's not as clear of a path in my opinion for mechanical ones. And it's been harder if I know fr if I have friends that want to go on mechanical path, for them to get like their beachhead for getting started for this kind of thing.

[01:06:58] Canon Reeves: Yeah. It's super tough.[01:07:00]

[01:07:00] Audrow Nash: Interesting. Yeah. And you're right about, with software where it's like you just ship all the time because things are really easy and or really easy to deploy relatively. it's an interesting thing, but yeah. Okay. Super cool.

[01:07:17] Exciting Growth and Challenges Ahead

[01:07:17] Audrow Nash: And you guys, you recently got a round of funding. We talked a little before and you said it's a seed.

[01:07:23] It's like I, I looked online. Is it the second seed round that you guys

[01:07:27] Canon Reeves: We did it in two tranches. So yeah, pretty much did the second tranche kind of this, earlier this year, we've got all of these relationships and like customers and we do not have the people or the space to fulfill them. And so we're now shifting a little bit of our mode of okay, how do we fulfill these orders that we have?

[01:07:48] which is a very daunting task. And then how do we step our foot on the gas and get more orders? And, it's a cycle that feeds

[01:07:53] Audrow Nash: You're in a scaling spot already. And that's pretty crazy because, [01:08:00] what I would think, so I've talked with a few investors. I had one on the podcast, Sanjay, and what he was saying is for scaling, like, you guys, you're trying to put your foot on the gas and you're trying to deploy a lot. but that's like series B almost for this.

[01:08:15] And you guys are just having your seed round. which is very interesting that you're there and starting to scale. help me think about this. Like what, are your thoughts on it?

[01:08:25] Canon Reeves: yeah, I think we're,

[01:08:29] Audrow Nash: which is just phenomenal.

[01:08:30] Canon Reeves: yeah, I think it comes back to staying really lean and, just having that team of killers, but, certainly some more, we'll certainly have some more fundraising ahead of us, no doubt. and when we look at scaling, like this is really the early days of scaling.

[01:08:46] So it's, a lot of still not what you would see in that for more refined state, but like we just moved offices. So this, like today is the first day I'm going into the new office. So we're doing things like this to get ourselves set up to do assembly and, [01:09:00] really scale the production of these things.

[01:09:03] Audrow Nash: Okay. And, so you're, I have it as perplexity. and so your seed round, that was 13 million, your, you mentioned the number of people, you're like 10, 15 people or something like that. Now

[01:09:17] Canon Reeves: Yeah. Yeah.

[01:09:18] Audrow Nash: what do you imagine, like in a year from now or in, I guess it's hard to predict further than that really.

[01:09:25] what do you think you guys will be in a year from now? number of people wise.

[01:09:29] Canon Reeves: I suspect we'll double, maybe more. It depends on how these deployments go. Yeah. I'm really excited.

[01:09:36] Audrow Nash: yeah. you're at a cool point. 'cause it's like now you get to grow, you get to bring people on and you get to challenge, like your challenge from hearing from others. I haven't been through it, but your challenge appears to be.

[01:09:50] Keeping a good culture for everything and making it so you guys are really productive as you're bringing people on and not letting like little [01:10:00] relationship souring really impact the company. 'cause you're still small enough that really matters still.

[01:10:06] Canon Reeves: Yeah. You're still at that size where it just did all that team vibe and that culture affects the everything. and it really goes back to, so one part of hiring killers and having this really high standard for hiring is like not just the performance, but are they good people to work with?

[01:10:23] We have a no smart assholes rule. I don't care how good of an engineer you are. If you're

[01:10:27] Audrow Nash: terrible. Yeah, totally.

[01:10:28] Canon Reeves: it just sucks the life out of the team. And so we, every single person that works here, I just care so deeply about, and they're just genuinely good people. And I think that matters so much when you're in a small team and you're asking these people to show up so heavily, doing it with people you actually like.

[01:10:43] It's

[01:10:44] Audrow Nash: Oh yeah. Oh yeah, totally. Yeah. You could be a family that is you guys are a great relationships, everything's enjoyable. You're going through it together. It's exciting. Or it can be like the company's being successful, but you guys are miserable

[01:10:58] Canon Reeves: Yeah. And it's [01:11:00] what's the point of that? life's so short. I want to love coming into work every day and I want all of our employees to love coming into work every

[01:11:06] Audrow Nash: oh, totally. Yeah. No, that seems really good. yeah, it's, such a funny thing if companies get rich, even though they hate what they do, and you're doing this, people who are founding, it's like a lot of times you want control of your time or you want to enjoy your life as much as possible, and you're taking a very active step towards that.

[01:11:25] And if you do that, you just stand into something that you're miserable with. It's it defeats the point of doing

[01:11:31] Canon Reeves: Totally.

[01:11:31] Audrow Nash: almost entirely, which is just so funny.

[01:11:34] Canon Reeves: Yeah.

[01:11:35] Audrow Nash: Uhhuh, what a thing. So what do you, so you'll probably double, what do you imagine needing. In the next year for this kind of thing? what kind of roles?

[01:11:47] I imagine you're gonna be getting more and more into software as you're scaling, because that is probably how you do the optimization and the like, coordination between all the robots that are at that are out and things like this, but, [01:12:00] and probably people to do the assembly or not assembly, but the, actual, I don't, know if you guys contract that out or how it works, but what kind of roles do you think you'll need?

[01:12:09] Canon Reeves: Yeah. I think a, this will probably be the first phase of the company where engineering is not the department that's growing the most right now. it's 90 percent engineering right now.

[01:12:18] Audrow Nash: I bet

[01:12:20] Canon Reeves: because we're just so engineering heavy, but I, what I suspect will happen is our, our sales and business operations will grow a good bit as well as our assembly and manufacturing.

[01:12:30] next hundred, we're really planning on enhancing a lot of the assembly and building. We anticipate the designs to be changing. this is just going to be something that we learn and grow through. so we want to do assembly and house and move to a bigger place where we can do that.

[01:12:41] So we'll definitely have some hires in there. and then, yeah, all of those kinds We will continue to add. And it becomes easier to say okay, if we hire somebody to go optimize this part of the construction process, once you have enough of those construction processes happening, that person's role pays for itself,

[01:12:58] Audrow Nash: ha.

[01:12:59] Canon Reeves: [01:13:00] quickly.

[01:13:00] So I suspect we'll start to add more people who are just really tightly scoped. Whereas right now, everyone's just so generalist and flexing between a lot of different things.

[01:13:08] Audrow Nash: Cool. Yeah, that's a smart thing to do. And it's funny to view it as they pay for themselves. 'cause that's wonderful if they can shave off, parts of your process and literally justify their own salary, that's super cool. Yeah. What, are you guys all in Austin, Texas? Or do you have any remote people or, I don't know.

[01:13:27] How does it, how do you view remote work?

[01:13:29] Canon Reeves: Yeah, we pretty much always hire in person. we've got. Two employees who are remote, and it's the times when we go to remoters, they're just so killer that we just, you have to have them, but they've got some of the right life reason for being where they're at. And you can make it work. So like our software engineer, Matt, he's remote, but it set us up for all of our systems being really remote friendly.

[01:13:51] So no matter where our robots are deployed, we can do things. I don't know if you saw the, the build list voting poll thing.

[01:13:58] Audrow Nash: I'm not sure.

[01:13:59] Canon Reeves: The Christian [01:14:00] killed it. There is a list of all the deep tech companies and you could go vote. And, rank them. but our software engineer, he, he, we were joking that if it's a startup list, the thing to do is to hack it and, robo vote your way up just as for fun.

[01:14:15] And so we had all of our robots, like we have robots everywhere. We have robots in like Austin and Atlanta and DC. So we actually had all of our systems and our robots voting in this poll. So everything's really set up to be remoted into, but yeah, for the most part, when we bring on people, especially engineers, we wanted to be in person and, just being able to walk to somebody's desk and like solve problems.

[01:14:35] This is just

[01:14:36] Audrow Nash: Really nice. Yeah,

[01:14:37] Canon Reeves: Yeah.

[01:14:39] Audrow Nash: for sure. Okay. Yeah. What a cool thing. It makes sense to me. Software can be remote 'cause a lot of the workflow makes a lot of good sense and a lot of the getting in flow, like getting in a good state of focus, what makes a lot of sense. But then for all the mechanical stuff that you're dealing with, like it just being able to talk and show [01:15:00] that kind of thing makes a lot of sense.

[01:15:02] Canon Reeves: Yeah. Especially having way of a machinist and a build tech. And so it's like being able to just walk over there to the shop and

[01:15:08] Audrow Nash: Oh yeah,

[01:15:09] Canon Reeves: and tell them what you need and being able to see how things come together in their hands really matters.

[01:15:14] Audrow Nash: for sure. Do you, within the office, do you dog food your systems? it'd be very funny, you're sitting at your desk and you like, hit, I want a. but, fizzy water or whatever, and then it comes to you by one of your systems or something. I guess you'd need to do a lot of construction in your office to do that, but

[01:15:31] Canon Reeves: yeah. I think in the older old office, there wasn't quite enough space for that, but we might then do it. We're planning on doing a, an office install, in the front yard, front kind of parking lot of our office. we do want to have one here so that we can use it. Cause yeah, like being able to use it and just seeing where it sucks.

[01:15:45] So you know where to fix it. It's so important.

[01:15:47] Audrow Nash: Oh yeah. Dog food in your own stuff. super important. Hell yeah. And I like that you guys are remote first for this kind of thing. I feel like it's almost like a minor setup [01:16:00] cost, and then it's just so much more flexible from then on though. I think that seems really good. Let's see. So I would like to, is there anything, we missed, about you guys that's really interesting about the like. Hardware setup or your deployment plans or anything like this?

[01:16:22] Canon Reeves: One of the cool things that we See happening is creating a USB port for the store. And what we mean by that is, is when, we think that other autonomous modalities will exist. We think drones will exist. We think self driven cars will exist. Like I, I think that delivery is such a big category.

[01:16:43] There's going to be multiple winners that make sense in different kind of scenarios. Yeah. Yeah. And so we see ourselves as being this sort of like USB plug. where a restaurant or a store can install our system in the back of house so when their workers are doing their flow and putting in orders there's one [01:17:00] flow but let's say you wanted to hand off to a drone well instead of teaching that worker how to go interface with that drone and setting up

[01:17:06] Audrow Nash: Oh,

[01:17:06] Canon Reeves: lot for that drone and like building all

[01:17:08] Audrow Nash: that's so cool.

[01:17:09] Canon Reeves: we can automate that handoff And so now you can put the system in and maybe we're connected to a field that's a mile away, but all of these kind of drone orders are coming to this field.

[01:17:19] We can hand them up autonomously. And so when restaurants and partners install our system, they're like future proofing themselves and setting themselves up to work with multiple modalities as we expand. And it plays into these network effects because as soon as you add one drone hub to this network, Every restaurant that's on that network, every grocery store, every warehouse now has drone delivery without having to like ever take that on themselves.

[01:17:44] Audrow Nash: That is super cool. Yeah. That's another, like we were talking about the other, how you could do a lot of pipe laying down. You could also be this super good interface for all of these other modalities. That's another really powerful model, I think.

[01:17:58] Canon Reeves: Yeah. And you [01:18:00] hit on it with the shipping container, right? Like we have this already at a global scale. We don't have anything sized down at the order level, the individual cart. we think there's room to create that. you can have a sidewalk robot do the last hundred feet from that neighborhood portal down to your specific spot.

[01:18:16] there's a bunch of I think a bunch of options that we'll just. be available if we approach this in a way that's easily standardized.

[01:18:23] Audrow Nash: Yeah, and you, threw the constraints. one, it's pragmatic to choose a kind of standard container for all of your things. but two, a result of doing that is that you then give something a standard. you give other companies a standard interface for a payload, which makes it so that it's very easy for them to potentially move things.

[01:18:47] So yeah, you, could just make it so that you route things. I, could even imagine like factories wanting this for Sending parts from here to there. And part of their process, because [01:19:00] a lot of times they have conveyor systems, but they might have adjacent warehouses where they're doing things and maybe they wanna route things from one to another.

[01:19:06] And there's currently a guy with a forklift that stacks a bunch of 'em and drives them from one to another. Or maybe if they're really savvy, they have a mobile robot that does it. but you could have this really fast feed through, through this.

[01:19:19] Canon Reeves: Yeah, it's like a conveyor belt, like a stream, and your distance is not really constrained. You just go as far as you want.

[01:19:26] Audrow Nash: You are limited by your number of robots is the thing that, but I suppose if you make, like they're in a loop and you can probably make pretty good reuse of them and they go fast. but if you had a lot of end points, I don't know. It's so cool. Yeah. There's so many cool possibilities. You could just, you can basically be an infrastructure for moving stuff and it's that's super, general and cities is one thing and food delivery is part of that.

[01:19:53] but there's a lot of opportunities I would imagine even like, how funny would it be if a banks stop using their vacuum system [01:20:00] and just use the robots? Because the thing is with the vacuum system, if it fails in one spot, it probably takes down the whole system. Whereas this is much more fault tolerant.

[01:20:09] One robot fails, one pipe fails. Okay? Don't use that one. But the rest of them are open.

[01:20:13] Canon Reeves: Yeah, we've got versions of this. It doesn't have to be underground. Really, what matters is that there's a space around the robot. And so we've got a version of this that's an overhead ducting. And so we've talked stadiums about connecting kitchens to different drop points in the stadium.

[01:20:27] Audrow Nash: Oh my gosh.

[01:20:28] Canon Reeves: but, it's at the end of the day, we've created a utility that gets things from A to B.

[01:20:34] And what you send and where A and where B are, super flexible. Yeah, you just have to

[01:20:39] Audrow Nash: you just have your interfaces, your portals, is it portal on the sending side for this

[01:20:45] Canon Reeves: is, yes. And they're always, those portals are different too. And they're optimized for like totally different things. you really have to make something that any worker can go and use easily, and it can still be fast and easily serviced.

[01:20:56] Audrow Nash: and you could, I could imagine that your portal [01:21:00] becomes general enough that they just put a hamburger on a, if it's a fast food one, they just put a hamburger on a tray that then lowers and it boxes it and sends it off. you can continue to optimize these things pretty

[01:21:13] Canon Reeves: yeah. Yeah. And by using a, we use a standard tote that's already used really prevalently throughout the grocery industry. There's systems, there's ASRS systems already built around storing and retrieving these totes.

[01:21:24] Audrow Nash: what's a ASRS? I don't know what

[01:21:26] Canon Reeves: Automated storage and retrieval systems.

[01:21:28] Audrow Nash: Oh, I would never would've known that

[01:21:29] acronym. Okay. Yes.

[01:21:31] Canon Reeves: so it's these big systems that are, a lot of these warehouses are goods to person.

[01:21:35] So they've got to store all these totes and then pull your tote to you to get your item. and so what that means is there's a lot of automation built around this one already. And so we could plug into that. And so you could have an automated warehouse, like an Ocado style where they are storing items.

[01:21:50] Bringing bins, picking the bins, which eventually will be a robot picking them, and then you can have that just send directly into our system. We see the modularity as such a key piece to this [01:22:00] scaling and being an effective utility.

[01:22:02] Audrow Nash: Yeah. And I'm imagining, like I've interviewed, plus one Robotics also here in Texas. They, they do a lot of like partial sorting and parcel placing. I think I. And the interesting thing with that is you see that what they do is they, you get a parcel, like a bunch of parcels, a bunch of boxes, and then they have a robot that puts 'em on conveyors to route them all different places.

[01:22:23] And they come back together on another parcel once they're collected for one thing and they just keep going. you guys could eventually have your own logistics network that includes all that. 'cause that would make a lot of sense. In fact, like you could have bigger payloads and then you could distribute them and then you could route those further.

[01:22:40] Canon Reeves: And if you think about, if you are, let's say we're starting up a 3PL company and we're like, okay, I gotta go buy this warehouse to do this automation. You want somewhere where, that, that warehouse is cheap and you can actually get it and you can get workers there, but you also want it close to where you're delivering things.

[01:22:56] We don't have that same problem. We actually are this kind of [01:23:00] interesting retail arbitrage where we can go to really undesirable and cheap land, build this big warehouse, centralize, because, and then move so much volume through that you can invest heavily in automation for these, maybe it's sorting systems, but then you're, because you've got that pipe running all the way into the city and then you connect to this broader network, it's as if I have this massive warehouse with all this automation in a parking lot right next to your house. And for us, it's free to get things from that warehouse to that parking lot. And so we've just turned the value of that spot on its head because we just had to plug in this big system.

[01:23:33] Audrow Nash: That's wild. Yeah. You guys, you're in such an exciting position, it seems like to me that's sick.

[01:23:41] Canon Reeves: Yeah, it's going to be a really interesting.

[01:23:43] Audrow Nash: movement as a service of items for this kind of thing. And your cost as you lay down pipe, which those pipe lay down everywhere. but there's a lot of stuff in society to move and so that's exciting.

[01:23:56] Canon Reeves: Yeah. I think you hit the nail on the head earlier about there's a [01:24:00] lot of great options ahead of us. And what really matters now is what, path do we select and where do we go? Yeah. Cause it's there are a billion, app moving applications. We've really tried to down select for what things will have network effects.

[01:24:12] if we go to a stadium, will that stadium ever send items and foods into the city? Probably not. Is there an interesting business line there? Yes. And at some point, it really would make sense to

[01:24:22] Audrow Nash: it'd be a good

[01:24:23] Canon Reeves: up a small team. yeah. And there's times when that'll make sense. But when those early days, focus in on those places that are going to add to those network effects and add to the flywheel.

[01:24:32] Yeah.

[01:24:33] Audrow Nash: Yeah, for sure. Very cool.

[01:24:36] Canon Reeves: Yeah, and it becomes a land grab at a certain point, there's probably only going to be one of these systems in a specific area because you want these network effects to build up. And so there's going to be this, once you prove that model out in one metro area, I think there's going to be this huge land grab.

[01:24:52] Audrow Nash: Yeah.

[01:24:53] Canon Reeves: where we really just try to deploy as much as possible because once that portal is in, we can just get, extract value [01:25:00] and create value around that portal for decades.

[01:25:02] Audrow Nash: Yeah, it is interesting 'cause you, but then that's where coming in early is very good because you start to gain efficiencies of scale. and so you can probably do it much cheaper and you've already gone through regulation and things like this. so that's something that's a big moat in a sense, like your technology to me, I'm sure the, portals are quite complex.

[01:25:24] You can add like that, you can optimize the hell out of those. but your fundamental robot and pipes are pretty simple with this. So it's an interesting thing. So then what I would imagine is once there's more entrance into this space, it's going to become more about efficiency unless you just have a cash burning, one cash burning machine.

[01:25:45] But, seems like. Why would they do that? Why would VCs invest in that when they're already, as you guys for this kind of thing? I guess maybe the return that they could get if it's earlier.

[01:25:56] Canon Reeves: Yeah. And the other piece is the, supply side [01:26:00] relationships.

[01:26:00] Audrow Nash: Yeah. Totally.

[01:26:02] Canon Reeves: there's only so many.

[01:26:04] Audrow Nash: Big

[01:26:04] Canon Reeves: Restaurant. Yeah. There's only so many of these partners and we've had contact with most of them. And so if we can build those relationships and show them in a single metro area, here's a working model.

[01:26:16] Let's go scale this up together nationwide. they're a lot less incentivized to go work with somebody else. I'm sure they will at some point. But, If we can move quickly and just get it going, it's, it's once this ball starts rolling, I think there's going to be a ton of momentum.

[01:26:32] Audrow Nash: Yeah. I think your biggest challenge is gonna be, keeping your team culture. to me that's the big risk at this point is you guys need to not destroy yourselves. 'cause you have a, it was like you have a golden goose in a sense. And it's like all the people stuff becomes, like, all of your other risks are out of the way in some sense.

[01:26:51] I know that there's still other things, but it seems like that will be the big challenge of the startup to me.

[01:26:59] Canon Reeves: Yeah. And [01:27:00] that's why when I, walk out of this room and I see our team, it's I just feel so good about that problem because every, everybody there is just already

[01:27:06] Audrow Nash: You like 'em a lot? Yes.

[01:27:09] Canon Reeves: they are going to be, now I'm doing all the interviews, but at some point it's going to be them and the people they hire.

[01:27:14] And it's

[01:27:14] Audrow Nash: Yep.

[01:27:15] Canon Reeves: you, hire people that you think are similar to you in some ways, I think. So like keeping great people. is going to make it easier to keep getting great people and working on a hard problem makes it so much easier too. but yeah, nail on the head, talent and culture are the big things ahead of us.

[01:27:30] Audrow Nash: Hell yeah. But that's a very exciting spot to be in, especially 'cause you're coming from what seems like a strong position already from that, hell yeah.

[01:27:38] Canon Reeves: Yeah,

[01:27:39] Audrow Nash: let's see. So we are coming to our end, to the end of the thing. A what, so you have already started, you mentioned you already founded a company in the past.

[01:27:52] and so now you're founding, you're a founder of, Pipedream, right?

[01:27:56] Canon Reeves: Go founder. Yeah. With Garrett.

[01:27:59] Audrow Nash: Yeah, With Garrett. [01:28:00] Okay. Hell yeah.

[01:28:00] Advice for Founders

[01:28:00] Audrow Nash: What advice do you have for other founders? I don't know. what things would save you a lot of time or what I don't know. what would be good to tell founders I suppose.

[01:28:12] Canon Reeves: Yeah. To me, the number one thing with starting a company is starting something that you can wake up every day and work on and be excited to go to bed at night thinking about, and it's a problem that you. Genuinely care about and the people in the industry you actually want to work with. this is such a painful and difficult process, no matter what, that if you don't love every second of it, it's hard.

[01:28:38] It's a hard job. It's it sucks if you don't love it and love the pieces of it. and the other thing is I, really am a fan of taking Big swings like this, like PipeDream, there's probably like a 2 5 percent chance this like works and scales and is a real thing. But man, that 2 5 percent chance is so exciting, and that world is so exciting where that works.

[01:28:58] So it's like, optimizing [01:29:00] for those problems that are hard, it comes with its own set of challenges, but a lot of those challenges can be controlled internally through engineering or problem solving. Whereas if you're in a space where you're just competing with a ton of people who are all doing the same thing, it's just, it's, yeah, dude, it's a lot different.

[01:29:15] So I, I'm a fan of, yeah, go for the big swing, go for the thing that when you tell somebody, they're like, what are you talking about? That's crazy. And they get excited to hear about it and, go for something that has a brain work to it. Pipedream has a certain brain worm. We'll talk to people and they'll ask a ton of questions.

[01:29:30] Audrow Nash: A brain. What? A

[01:29:32] Canon Reeves: Brain worm?

[01:29:33] Audrow Nash: brain worm. Hell yeah. So you mean they get it in their head. They can't stop thinking about it and

[01:29:38] Canon Reeves: Yeah. And the next, yeah, the next time you see them, they've got all these questions and they've thought through this thing. And they're always great questions. And, just, it's just an idea that has this sort of like mimetic potential.

[01:29:50] Audrow Nash: That's super cool. Yeah. What a thing. So how do you, talking about the problem you always love, Even when you love a problem, at some point [01:30:00] you, and like even if you enjoy the work, there is a point where you can do too much or, you won't like it, but you still have to keep showing up. I find like any project that I start, anything I do after some point, I am no longer wanting to work on that project.

[01:30:17] And I think that's natural because you wanna work on other things and you wanna whatever, but you have to keep showing up anyway. how do you 'cause the advice of

[01:30:30] Canon Reeves: it.

[01:30:38] Audrow Nash: for this kind of thing once you've been in it for so long.

[01:30:41] and everything is painful. Like you've solved all the easy problems and the only things left are the hard ones. how do you keep showing up?

[01:30:50] Canon Reeves: You talk to customers, it,

[01:30:52] Audrow Nash: Oh, that's cool.

[01:30:53] Canon Reeves: for me, yeah, dude, it's when you get on that call and I've been in the CAD all week and Oh, this is like driving me crazy or trying to solve this [01:31:00] problem. But then you get on the call with the customers and they're trying to get it faster. And they're just like, when can I get this thing? And they're telling you about, and usually Our system is so simple and it's hard to create a simple thing. But once you create it, anybody can catch on and start to contribute to the idea. And so our customers are coming up with ideas for maybe we do it this way. Maybe we do it this way. And they are getting excited about it. When you're at your lowest, if you're working with customers you like, it just, and that's the piece, that's to me, the piece that keeps you going.

[01:31:27] Audrow Nash: Their energy in a sense. Yeah, I,

[01:31:30] Canon Reeves: And they've always got a new thing that they want and you got to find the limit, right? You can't just engineer forever, but they've always got a

[01:31:36] Audrow Nash: their things, but you can take the perspective and you can take the excitement and

[01:31:40] Canon Reeves: Yeah. And that keeps it fun. it's like that next thing to work towards and work on,

[01:31:44] Audrow Nash: Hell yeah. Okay. that seems super exciting. So if there is, I. One thing that people take away from this interview, what, do you hope it is?

[01:31:56] Canon Reeves: go do that hard [01:32:00] thing.

[01:32:00] Audrow Nash: Ha

[01:32:01] Canon Reeves: about what you want society. to look like and how you can help create that and go work on it. it's hard, but it's so worthwhile. And, our cities and our society and our world is not going anywhere anytime soon. let's, keep building together and just build interesting things.

[01:32:20] Audrow Nash: Oh, yeah. Love it. All right. Thank you, Canon. This was a blast.

[01:32:24] Canon Reeves: Cool. Thank you so much.

[01:32:25] Audrow Nash: All right. Bye everyone.

[01:32:26] That's it for this episode with Canon Reeves from Pipedream Labs. I don't know about you, but I'm pretty excited about the idea of getting my food delivered through underground pipes. It's wild to think about how this could change our cities and the way we get stuff.

[01:32:42] What do you think? Could Underground Delivery Networks be the future of Last Mile Logistics, or is this just a pipe dream? See what I did there? As always, thanks for listening and I'll catch you in the next one.