Divya Thakur (Founder, Yardbot)

Divya Thakur, founder of Yardbot, joins host Jim Belosic to discuss how he’s building a next-gen landscaping company powered by robotics. Instead of chasing full automation, Yardbot combines human crews with custom-built mowing robots to deliver better results at a lower cost.

Divya shares his journey from software engineering and finance to self-driving cars at Cruise, and how a failed Bitcoin startup taught him one of his biggest lessons: build for customers, not just cool tech. We get into the gritty reality of developing hardware, five generations of robot iterations, and what it takes to run a tech-driven business in a traditional industry.

Transcript

SPEAKER_01 0:00

And to be frank with you, everything just broke. He's like, dude, what are you talking about?

SPEAKER_00 0:03

You're not gonna be able to buy a house for a million bucks. If this whole Senkut Sen thing doesn't work out, we're gonna go back to mowing lawns. That's how I made money for RC cars, which was like my my drug addiction back in the early 90s.

SPEAKER_01 0:14

I was running a team of 120 engineers. And how hard can it be to run a small team of blue-collar l workers?

SPEAKER_00 0:20

You can have a thesis, which is like my freaking least favorite word. Everyone has a thesis now. Hey guys, welcome to the Just Gonna Send It podcast. I'm your host, Jim Belosik, and today I'm joined by Diviator of Yardbot. So Divya, thanks thanks for being on the pod. Thank you.

SPEAKER_01 0:46

Happy to be here.

SPEAKER_00 0:48

Cool. We're gonna dive into uh kind of Divya's background and how he finally got to Yardbot, uh, which is awesome if you haven't seen it. And so it's yard.bot, right? That's the web address. That's correct. Yes. Okay. Logan Logan will put it right here. Um tell us a little bit about Yardbot before we go into your background and everything. Like what's what's Yardbot all about?

SPEAKER_01 1:12

Yeah, so uh Yardbot, or we sometimes call it Yard Robotics, is a hybrid landscaping and a robotics company. So on the front end of the business, we do what's traditionally landscaping. So mowing mowing lawns, weed eating, edging, blowing, and also other activities like mulching and bush trimming for residential and commercial customers. So at the front end of the business, we present as a as a lawn care company. And on the back end of the business, uh we build robots, we design and build robots that we ultimately use for this uh activity, primarily for the mowing side. Um so yeah, it's kind of a company of these two sides, essentially.

SPEAKER_00 1:46

Okay. Um yeah, I I definitely thought it was more just like robots that you could put in your yard, but it's it sounds like it's yeah, a hybrid of an actual like traditional landscape maintenance company. So are your guys are like hauling these things around on trucks and then they're weed eating while these things are mowing?

SPEAKER_01 2:03

That's correct. So yeah, we we have a truck with three robots and weed eaters, Azures blowers. We train our guys to use these robots in remote control and in autonomous mode. So remote control when they're driving it with a remote controller or autonomous mode, let's say you see an open yard, you can put the robot there and it can mow autonomously. And at the on the other side of this, we're we're presenting as a company that provides a replacement for a traditional landscaping company. So for the from a customer's perspective, we are better, our lines are straighter, our price is lower, and on our side, we're we're kind of doing this because we have these robots that we built.

SPEAKER_00 2:37

Okay. That's yeah, that's wild. Um I've I feel like, you know, what, uh 10 years ago or something, I heard about all these companies that were trying to do, you know, a Roomba for your lawn. And you know, I know a couple guys who bought them and uh and they sucked. You know, they would like drive across the street or or they'd get stuck in a ditch or whatever. So it sounds like you guys didn't try to um over-automate, you know, it's a a mix of a mix of dudes and robots is is maybe the magic.

SPEAKER_01 3:05

Yeah. Yeah. So so you know, when you think about lawn care, Jim, the end product is a good-looking lawn, right? So the solution path can be a robot or can be a human being. You know, the the insight that sort of we built is that the end customers, you know, these are like moms, you know, they're not they're they want a good looking lawn. They're not interested in learning how to operate a robot and make it do it correctly. And even if even if it does a perfect job, there's human activity, weed eating, edging, blowing the front drive, you know, front of the house. The robot's not gonna do that. So, you know, ultimately the true solution, what we realized is we people just want a better quality lawn care service. And it's a different mindset to think about this from this perspective versus let's we have a great robot and we want to give it to people. So that's that's sort of a different way.

unknown 3:51

Yeah.

SPEAKER_00 3:52

Yeah, you're you're kind of speaking my language because uh every day I see all these companies that are trying to just over-automate something. They're like, oh man, we're gonna get rid of all of our staff and just have robots. Um, you know, or we're we're gonna have we're gonna automate our forklifts, we're gonna automate, you know, all of our lines, we're gonna have humanoids, we're gonna do all this stuff, and we don't need people anymore. And we're just not there yet. You know, and honestly, I don't know if we'll ever be there. And, you know, one of the things I'm proud of at Send Cut Send is we use dudes where it makes sense to have a dude, and we use robots where it makes sense to have a robot. So if it's repetitive or you know, uh a very discrete problem that's that's uh solvable in the same way over and over and over again, it's great to use a robot. But uh I don't know, not a lot of companies are thinking about you know uh what the mix is like. You know, is did you did you know that was gonna be your your take from day one or you know, like the hybrid, uh, or or were you trying to fully automate it, or how did you end up there?

SPEAKER_01 4:56

I mean, it was just lessons, right? So I started off by thinking if I design the perfect robot at a great price, it's gonna solve my problem. And I built something. And then I realized that it was not, it's not, it doesn't do the full job. So that you know, it was a series of like lessons. Then I, you know, the next iteration of the business was perhaps I can sell this to traditional landscaping companies. So then I started knocking on doors there. You know, the reality is the landscaping industry is a you know, it's a different world. And the kind of talent that works in landscaping, I mean, I came from software, hardware background. So it's a different type of talent. And they'd rather hop on a zero turn and whip through 10 yards than figure out how to sit in front of a computer and run a fleet of 20 robots. That's, you know, it's a different kind of person. So it was a difficult time doing that. And then ultimately I realized, hey, if I can just be price competitive and competitive from a quality perspective, I can just start a landscaping company and you know the robots can make it happen. So that's kind of the current iteration of the business, and this is the this is the one that's been the most successful. So to answer your question, it's just gone through a couple of learnings here and ended up where we are today.

SPEAKER_00 6:00

Cool. Yeah, I think that's that's our our favorite motto here is uh learn by doing. Uh, you can have you can have a thesis, which is like my freaking least favorite word. Everyone has a thesis now. Like, no, just hey, I got a random idea. I have a gut feeling, I'm gonna go test it, I'm gonna go get beat up. And you you just have to know that where you're starting and where you end are gonna be two completely different places, and the most successful people are the ones who can pivot and adapt. So uh it's it's funny. My CTO and myself, um, we both have a background in landscaping. His he worked for like a professional big landscaping company during college. Okay, and Jacob Graham. Um, and then I had a lawn mowing route since I was like 11 or 12 or something. Um so that's that's how I made money for RC cars, which was like my my drug addiction back in the early 90s. Um so he and I have always joked that you know, if this whole Senkut Sen thing doesn't work out, we're gonna go back to mowing lawns and uh and kind of make a new business model for it. So it sounds like you're already doing it for us. So hopefully I don't have to uh you know run a blower all day uh ever again. Uh that that gets a little a little crazy sometimes. Um so what's what's next for Yardbot? Like I I looked at your robots, they're rad. Um I see a lot of send cut send parts on there. Uh what generation are you on? Like what how many how many uh iterations is this?

SPEAKER_01 7:29

So we're on our fifth iteration right now. Um and it's um you know it's a 36-inch deck, which is uh perfect. Most of our customers are sort of subdivision homes, like residential homes. We do about 600 locations in the Huntsville area, Huntsville, Alabama area. Um so we have a fleet of 20 robots today. We're building 10 more very, very soon. We just actually got the order in. Thank you, Sankat San, yesterday. Um, so uh these robots um uh are yeah, so they're 36 inch, 250 pounds, they have a five kilowatt hour battery, uh triple blade, um, and it's got an autonomous system that's lighter and camera based. And let's see, and yeah, it can run all day. It can do one robot can do 10 yards, uh, and we put three, so the guys are you know a little bit, yeah, they they're a little, they have too many, but um and yeah, that's kind of the high level, the little specs of the system.

SPEAKER_00 8:23

Dude, that's that's amazing. Um 36-inch deck is no joke. Like, usually that's uh you know, many horsepower gas engine to to cut all that. And then you're in Alabama where that that's like that's like real grass. That's not this dry ass, you know, Nevada fake grass that we have. So uh like what are what are some of the challenges you've had? You know, you're on fifth generation. Like tell me about the failures of one through four.

SPEAKER_01 8:50

Yeah, so you know, um a critical person in the company is Nathaniel Chong. He's my uh you know, he's head of engineering hardware. So I first started designing them, Jim, and you know, I just 3D printed stuff because I came from a software background. That was my first step into building something. And to be frank with you, everything just broke pretty much right away. So that's when I discovered Sen Consent, to be frank, and I realized I could just actually get metal parts and started building the first couple of generations. The two areas I would say most iteration we've done is the drivetrain and the deck itself. So the drivetrain, uh, we started off with skateboard motors, so e-skate motors, um, which were direct drive. Over time, then we added a reduction with a belt, and uh we realized the grass gets into the belt and tears the belt apart. Next we moved to the chain drive. Well, uh, we moved to gears, actually, and uh we used planetary gearboxes uh for the next iteration. Um we were using five to one, that was not enough, so we went to 20 to 1, but the planetary gearboxes, the pinions were too small, so they would they would shear off. So then we finally moved to a chain design, and that's a current iteration, and it's amazing. It's actually really, really good. So it's uh it's a five to one reduction on a gearbox and then a three to one on a chain, and we do a it's a differential drive. So every every wheel is powered, and uh, we can basically climb any hill or do anything we want with the robot at this point. So it's it's really solid.

SPEAKER_00 10:10

Um what's that make sense? Yeah, no, that's that's incredible. It goes back to like learn by doing. Yeah. Um, you know, you can you can spend six months going through a bunch of iterations in CAD or whatever, or you spend six months you know building five different versions, you'll have a way better product if you actually get your hands dirty and go see what's what breaks. Um I've seen so many people that are like, oh yeah, this is this is gonna be super, super strong. And then yeah, they there's the unknowns. Um oh, I didn't realize grass was gonna get stuck in there, or I didn't realize, you know, that in muddy conditions the damn thing's gonna tip over. So I gotta readjust my center of gravity or whatever. So just yeah, learn by doing. Your your modeling in CAD is cute, but go out and actually like get shit done too. Um what's okay, so tell me tell me what is what is generation 10 gonna look like? Like what's your dream?

SPEAKER_01 11:04

So I think okay, that's an interesting question, Jim. So um I think there's obviously like mean time to failure, which we need to bring down across the fleet. And we're already pretty good, but every year it surprises us. We find new failure modes that we have to have to drive down. Um I think the platform, Jim, I think about it more than just uh mowing platform. So uh the way it looks, you can kind of tell it doesn't look like a traditional zero turn. It's designed to hold a deck, but perhaps hold a spray system, or perhaps hold a something to pick up garbage. I think about it as like a unit of automation that you can deploy in your community and then you know achieve like a better quality of life. That's kind of how I think about it mentally. So um I haven't thought through generation 10 yet. I'm a little bit in the iteration mode right now, but a platform that's super reliable, can run autonomously, and can do multiple tasks across a large area is kind of what what I'm looking to do here.

SPEAKER_00 12:04

I'll I'll tell you that's that's a fun exercise that I do all the time is where are we at now, but then where are we at in 10 years? Or, you know, what does generation 10 look like? I've like throughout my career, I've always thought very small. Um and sometimes that's bit me in the ass a little bit. You know, when I was uh my my previous life, I was a graphic artist and you know had a little tiny advertising agency. And I was like, oh man, I want to be the best agency in Reno, Nevada. And then I was like, oh, maybe on the West Coast, and then the nation, and then you know, global or whatever. And but I always thought in these very, very small steps. And I think it's a good, it's a good way to think, okay, how what do I look like when I'm dominating the globe and then work backwards and then go back and focus on your little small town. So if you can think about, you know, what does this thing look like uh generation 50 and then go back to like generation six or something like that, it'll it'll help your roadmap a little bit. It exercises your mind a little bit, and then and then you'll be discouraged because you'll you'll never make it to that that crazy um you know hundredth generation or whatever. But uh our our uh VP of operations, Brian Wolf, he always says, he says, aim for perfection and you'll achieve excellence, or something like that. It's like aim for the stars and you'll still get the moon or whatever. Like you'll still be pretty satisfied. So um okay, so you guys are in you're in Huntsville, Alabama Alabama. Lots of grass to cut. Um is that where you're from, or how did how did you end up there?

SPEAKER_01 13:39

No, I've I've I've I'm from a lot of places, but the last place I was in was uh San Francisco. But sorry, uh but just Jim, just to touch on that, yes, the the vision for us is can we take this and replicate it across the country, right? But that's still not V10. Okay, just so I'm thinking about you, you've encouraged me to think broader, but I'm still thinking about it. But somewhere in there is replicate this across the entire country. But yeah, I don't know if that's V10.

SPEAKER_00 14:04

But okay, so but yeah, a little bit excited about the sprayer too, because uh from my my you know 12 to 13 year old, 14-year-old self, um, I when I realized that I could use the spreader and put out fertilizer and then upsell that fertilizer, like it doubled my revenue. Um and that was it, that was the first experience I ever had to learn about like margin increase or whatever. I was like, oh man, if I, you know, I'm already there, I'm gonna do it for you know 15 bucks, but I could actually make 30 bucks if I did this fertilizer service and this fertilizer only cost me five or whatever. Um all of a sudden I was like, oh man, I'm making like double per hour or whatever. That was really cool. So yeah, add-ons like that on a on an existing platform, that would be that would be pretty amazing.

SPEAKER_01 14:51

We just started doing it. So we started offering spraying services. So weed control, fertilization, you know, weed control is liquid, so we're still trying to figure out how to get that on the bot. Um, but yeah, I'm I'm I'm aligned to you thinking your 13-year-old landscaper is living in my 40-year-old body, I guess. But yeah. Do you guys have to aerate in in Alabama? Yes. Okay. Yes. So we we have a toe-behind aerator for the robot. So we we kind of hitch it on and we can run the bot. Um it's not as big a service as I I would think, but uh when customers ask for it, we do it.

SPEAKER_00 15:24

Okay. Yeah, and usually it's a good way to destroy sprinkler heads. That's that's the best use of aeration, I guess. Or to locate sprinkler heads that you forgot about because you'll find it. Yeah. Um yeah. Yeah. Well thanks for those ideas.

SPEAKER_01 15:37

Yeah. Yeah. You were in you were in San Francisco. Yes. So just going backwards here. So um let's see. Maybe I'll start. I I moved to the uh let's see. Okay, let's start with San Francisco. So I was in San Francisco for about seven, eight years. I worked at a company called Cruz, which is a self which is was a self-driving car company. It was bought by General Motors. Um I started there as you know, when it was pretty small, about 60, 70 people. I started as a front-end engineer. Okay, that was my previous, previous life as a software engineer. And then uh the company grew really fast. It grew to about 2,000 people by the time I was done with them. And it was um I was uh you know, I was running a team of about 100 engineers and I ran um uh product engineering. So that was all websites, apps, mapping systems, some machine learning systems, some you know, uh collect fleet management, a collection of sort of uh software, software pieces. So I did this in San Francisco, and then during COVID, I decided, okay, you know, I want to I want to live somewhere new. Um and uh I just kind of you know built a spreadsheet of fastest growing cities in the US with great quality of life and low cost of living. And I convinced my wife at that time, like my wife, like, hey, listen, let's move to Huntsville, Alabama. And we've never been, we've never been to Alabama before. But you know, yeah, yeah. Yeah, but remember, COVID was a time of like, you know, rethinking everything, right? So so we flew out out to Huntsville, we liked it, we saw a couple of houses, we're like, all right, we you know, let's get this house. And we got a house, which we could, you know, we try to buy a house in San Francisco, it was impossible for us.

SPEAKER_00 17:08

Yeah, can you can you talk to me a little bit about that? Like tell me, yeah. Like you know, the price of a three-bedroom, two-bath house in San Francisco versus Huntsville, like what does that look like?

SPEAKER_01 17:20

Yeah, so today, uh so I live in a three-bedroom, two-bath or three-bath house. And, you know, I paid, I think, $350 for it uh when I moved here and during COVID. Now Huntsville has seen a lot of growth, so it's probably closer to $500 today. Um I was living in San Francisco, I was paying about $4,000 in rent for a one-bedroom loft, uh, I guess, you know, so really just one room. Um and I tried to buy a house, you know, because I was getting older. I was like, I told my wife, let's get a house. So it was tough, man. It's really hard to buy a place. I remember the funniest thing, interaction, was I went, I was taking a cab in San Francisco to some place, and the cab driver I was asking, Do you like San Francisco? I love San Francisco. He's like, Well, I'm look- I told him I'm looking to buy a house, and um I I can spend up to a million dollars. And for me, that was a stretch, okay, because that's a lot of money. Okay. So I I told him and he laughed at me. He's like, dude, what are you talking about? You're not gonna be able to buy a house for a million bucks. And that I kind of realized like something was wrong. Uh now this was during COVID, things might have changed. Maybe San Francisco's more affordable. I I don't think so, but maybe it is. But um, I think cost of living matters, you know, just to be frank.

SPEAKER_00 18:32

Yeah, if if you can lower your your monthly nut, you know, down to, you know, uh as as low as you possibly can, like your health improves, you know. Otherwise, you're making a mortgage on a million dollars or whatever. Plus, you know, in San Francisco, it's like seven bucks for a black coffee and eighteen bucks for a burrito and all this other stuff. And then you go somewhere like Alabama or uh, I'm a little more familiar with Kentucky because we have we have a facility there. Uh we any anytime I travel, I open up Zillow and I try and figure out like, okay, what's what's my even trade, you know, for my home in Reno versus what can I get here? And I think in uh, you know, we we have a uh three-bedroom house here in Reno, and I think the straight trade in Paris, Kentucky was like 80 something acres of like horse farm with like a barn and all this other stuff. And I was like, oh my God, you know, and that's that's a Reno, which is a step down from you know a big coastal city like like SF or or uh Los Angeles or New York or something like that. So yeah, anyone listening out there, your quality of life will be better, your health will be better, your stress will be gone if you can just yeah, lower, lower your nut and uh uh I don't know, buy a buy a house in a good community. That's small.

SPEAKER_01 19:51

But Jim, can we but Jim, can we say something nice about San Francisco? It has some magic, right? It has magic. It has a it is a magnet for talent, you know, quality of people is incredible. So, you know, it there are some upsides, so it's not a straight decision for everybody, you know what I mean? But no, I agree.

SPEAKER_00 20:10

You know, if you're when you're living in a hub, there's just totally different opportunities. Right. You'll meet people that you would have never met otherwise. Um, you'll have opportunities that you've never had otherwise. Um, you know, your neighbor will be like, hey, you know, I got a job with this new company. Maybe I can get you a job. Like that happens in those big cities. Uh just running into people on the subway or something, it's it's pretty cool. Um, yeah, it's it definitely happens less, especially, you know, if you have a little bit of land and then you don't actually see anyone. Uh which is that's kind of my goal uh someday. I I'm I'm a bit of a hermit uh down deep. But right anyway, so so you're in San Francisco working at Cruz. Okay, so so yeah, I'm familiar with Cruz. That was that was a super cool company, um which makes sense the jump into Yardbot. Obviously, there's there's a lot of connections there, like the LIDAR and and everything. Um, but before Cruise, like you were doing you were doing stuff in like banking or something, right? You were at Goldman?

SPEAKER_01 21:10

Yes, yes, yes. So um so after I graduated from college, I went to Toronto, so University of Toronto in Canada. Um I moved to the U.S. So my first stop was Seattle, Washington. So I moved to I worked at Microsoft. Um that was a short stint, but very educational for me, but it wasn't for me. So I moved out of there. I went back to Toronto and I told my parents at that time, like, hey, I want to I want to try a new industry. I I I for like very quickly at an early age, no, I don't want to be in the corporate, like big company situation. So I want to try a completely new industry. So I decided to focus on financial markets. I try to study them, figure out what to do. Ultimately I ended up in a in a firm called Ready, which was part of Goldman, uh, which did uh electronic trading, so uh you know, options for. And I built systems for like you know routing trades, interfaces for placing orders, different types of you know, order types, like pair trades and different option types and different option strategies. So we we own that interface and that system and our team. And yeah, I joined that company and that that was awesome. I living in New York, you know, it was uh you know, seven, eight years there. And after that, um, I decided this was the first time I kind of realized that, hey, I moved to America, I can do something entrepreneurial. So I decided started a company in the Bitcoin space. And the Bitcoin is 20, 20, 20, no, 2012, 2013. Yeah, it's just coming up. It's just coming up at that time. And for me, um, you know, the interesting thing was that it was kind of like a financial product, but not exactly at that time. It was a technical product, like technology product, but also a financial product. So I came at it from a perspective of like, hey, uh, I'm working in Goldman's ready system, and it's it it routes all these trades from different exchanges and buys finds the best price. And perhaps I should build a system for um, you know, Bitcoin across Mt. Gox and BTCE. There's like all these exchanges back there.

SPEAKER_00 23:11

I don't know how familiar I am with Bitcoin gym, but you know, there's like I I remember uploading my driver's license to Mt. Gox to get an account created, and it was so freaking sketchy. I think I had sent fax at one point, which is like really dating me. I'm I'm 45 or 46. I think 46. Um 41, so yeah, about the same. Yeah. And yeah, we were mining Bitcoin. We had these little miners from um uh Butterfly Labs, I remember. Yeah. And they wouldn't send them to me because they were mining with the miners before they sold them. And uh yeah, and then I I started losing sleep because I think big Bitcoin went up to like 200 bucks or something like that. And I was like, oh my god, this is crazy. I gotta get out. And then uh yeah, sold, sold my stake. So yeah, no more Bitcoin for me. But my my health has been better.

SPEAKER_01 24:00

Okay. Okay, yeah, yeah. Are you are you still do you still have a stake? I have a little bit of Bitcoin, but nothing, you know. My company, my yard bot is worth way more for me. So I'm not I've nothing nothing meaningful. I I mean I sold. That's the that's the reality, you know. You're supposed to hodle. That's the that's the whole idea.

SPEAKER_00 24:19

But uh I that's not for me, man. I'm not a day trader. I'm just not told for it. Those guys are built totally different. Um I just couldn't do it. I think yeah, you you hear about people spending, you know, uh 25 Bitcoin on a pizza or whatever. And right, yes. If I look back at how much I sold, it would be nice to have today, but uh, you know, I don't have a time machine.

SPEAKER_01 24:40

Hard to look, you know, I I ordered some commander you remember Dogecoin, right? It was uh it was a thing back then. So I ordered these uh these little gold Dogecoin things by paying for them the Dogecoin, and obviously the amount that the Dogecoin's worth versus those little you know commander com uh you know, just like coins is is is way different. So anyway, but just to just to kind of go there, you know, I worked on the Bitcoin space. I I but this was a big lesson for me, Jim. It was like I didn't I wanted to design something that I wanted to do, and I didn't care about the customers, okay? This is I know it's obvious, okay, but hard lesson is like I wanted to design a multi-exchange Bitcoin clearing system for the for the sophisticated trader trading Bitcoin. Okay, those things, none of those make sense. Okay, they they in in 2024, 20, or sorry, 2012, 2020, whatever time frame it was, none of those things were needed by anybody. Okay, so I built a system, uh, nobody wanted it, and then yeah, it was a struggle to try to make sense of it. And it was a hard lesson. And you know, uh at some point I talked to my wife, I said, Hey, everything is uh, you know, bank accounts are zero. We you don't have anything going on, so I think I need to get a job at this point. So, and that's where I ended up at Cruz uh and we decided to change the scenery as well. We went from from New York to San Francisco.

SPEAKER_00 25:56

Okay, okay. That's yeah, that's awesome. I I'd love to hear about people's backgrounds and the weird journey, because you know, uh sometimes like you know, a new grad will come up to me and be like, Jim, yeah, how did you do this? You know, you're yeah, you have this this cool company, and you know, man, you must have like known exactly what you were gonna do. I was like, hell no, man. I was I was selling candy on the bus when I was eight, and I was mowing lawns at you know, 12 or whatever, and uh I was working at um Office Depot. Like I everyone has these weird backgrounds, um, but you you learn a little bit, you know, everywhere. So if you just had a straight line, like if I was always in manufacturing or something like that, um I think that that's what people expect. But really, I think it's sampling all these different experiences and then putting to putting them together in a a different way um for a legacy industry is really interesting. So and that's that's exactly what you've done. You've taken background from you know Goldman and the the company was celery, right? But that was the Bitcoin. That's right. That's right, yes. Okay, so and and then cruise and everything, and you applied it all into a very legacy uh industry like landscape maintenance, and and that's why you're disruptive. So um I don't know, a huge inspiration, man. This is this is super, super cool. Um however, I know it's not easy. Uh there's you we've talked about some of the the failures along the way. Can can you tell us any of the the down and dirty behind the scenes of like failures? Um like we we talk about our raccoons in the ceiling. You know, we we bought a a facility, the ceiling was full of raccoons, we had to figure out how to get them out, you know, just crazy stuff that you don't think about. Um when you're like, I want to be a business owner. Uh what are some of the failures that you've had either with Yard Bot or anything that that else that you can talk about?

SPEAKER_01 27:49

Yeah, I mean, I think um the first thing comes to me, Jim, is like uh honestly, people, humans, you know, um the kind of environments that I spent time in at Cruz or in New York and Goleman, it's a very different type of environment than working with landscapers and you know, and it's uh it's just a different world. Blue-collar work essentially is a different world. So because of the hybrid nature of yard robotics, yard yard bot, it's you know, I have to have a sizable force of people. Like we have nine right now that kind of go out and drive the trucks and and take the robots. And, you know, everything I kind of I thought I was I knew everything about managing people. I was running a team of 120 engineers, and how hard can it be to run a small team of blue-collar workers? And it's it's just completely different. Their incentive structures are different and interactions are different. And you know, when you think about a corporate environment, you're like, oh, you know, this would be unacceptable or acceptable. All those things change. So I had a lot of failures there learning about you know what motivates people and what keeps them happy in, you know, if they're not getting paid the same way as as we as people get paid in San Francisco. So uh yeah, I'm sorry, I'm not vaguely being specific there, but you know, it's just a hard learning uh on on the on the blue-collar side. On the robot side, I mean, you know, we've had robots end up in a lake, um, you know, and had to had to kind of pull them out. We've had robots fall apart, you know, the deck falls off, you know, all those kind of things. So we've just had a bunch of fun, random things. Thankfully, every everything is recoverable, everything is everybody's understanding. The customers honestly love the robots and seeing them in their yards. So nothing super crazy. But yeah, it was it was a surprise when the robot was in the lake and they were trying to push it up and they called me saying, Hey, I think we need another to get this out of here.

SPEAKER_00 29:44

It's it's definitely a flex uh to have the robots mowing your lawn. Like I would, I would definitely show that off to my neighbors and be like, oh yeah, no, I got I got robots who come in here. Yeah, I was gonna ask you a little bit about um, you know, how you go about recruiting for these teams, because it does seem like a hybrid. Like you have to, you know, you're you're gonna run this piece of high technology that has LiDAR on it and stuff, um, but then I also need to throw a backpack blower on you and you know go go blow leaves out of stuff. Um can you can you get a traditional landscaper to, you know, uh get on board with these robots, or or do you need someone who is like a little more interested in technology, or what's the ideal candidate there?

SPEAKER_01 30:34

I think it works on both sides, right? So we we have traditional landscapers and we have guys who've basically never mowed a lawn before this. Um the way I think about this is, and this this is something I do carry from my past experiences when when we think about individuals to hire, we think about skills, values, and abilities. And I think I was taught this in New York, I think it was maybe um uh Ray Dalio, but I'm not sure who kind of came up with the system. But it's basically skills being something that's super uh easy to acquire, three to four weeks, abilities taking about a year, and values are something you're born with, right? So values are stuff like do you work well with others? How much how big is your ego? And you know, things like that. So, you know, in engineering, when I when I recruited people, it was always about values only, because I mean there's skill tests and you know and abilities are kind of like you kind of get them over time, but it's about values. Can you work in a team environment, things like that? So that kind of came came, you know, in the blue-collar work as well, except there is a constraint. The constraint is because the just to be frank about it and honest about it, because of the way the system works, they're not paid as much. So money is really critical. And a lot of incentives get wrapped around compensation, right? Uh it's it does so in engineering as well, but in blue-collar work, it really, really matters, right? So, you know, so I still look for values, but you do need skills because we don't, you know, we're we're not gonna be able to teach you how to weed eat in a in the time frame that we're looking for you to perform. And the the key other element of this gym is training, right? So we have a rigorous training program. So every mistake we've made ever, right, is in our system. Every time we forgot a mode behind a fence, every time we hit a sprinkler head, we have a note and photos. So it's like a dream of training material, right? And thank you, LLMs. You know, this is one place they're really good. So we can take all this data.

SPEAKER_00 32:24

It sounds like a perfect application for LLMs.

SPEAKER_01 32:28

It's perfect. So we have a handbook that has the 50 different things everybody needs to be uh needs to keep in mind, everything from like soft stuff like your presence in a neighborhood, yelling loudly or smoking or you know, things like that, to precise stuff like what is edging? What does that mean? You know, how do you how do you screw it up? How can you make mistakes? You know, we have photos and examples of that across every single aspect of the business. And, you know, thank you, LLMs again. You know, we can aggregate all this data and present a solid training program that our guys can take. And then once they do that, we pair them up so we'd send a deal truck out. Our typical, our sort of standard procedures to send one person per truck, but we'll send two out when they're training, and then we send one out. And so far, honestly, this is after a couple of iterations, this has been the best system for us so far. So this is not the a precise answer to what you're asking, but these are sort of all the considerations when it looks to hiring hiring people for our firm.

SPEAKER_00 33:17

Yeah, um, I agree with you a million percent on values. Like I I can teach anyone to run a machine, you know, or drive a forklift or you know, um type into a piece of software or whatever, but I I can't teach you how to be friendly, um, you know, how to how to answer the door when a customer is knocking or whatever. Like your parents should have done that. So we we try and identify those things during the hiring process. Um one of my favorite things to do is like I'll grab a tote of heavy parts in it, and I'll be like, hey, you know, follow me. Let's go out on the shop floor, and I'll walk towards a door that needs to be opened for me. And I'll see if they grab the door. You know, if they just are thinking one step ahead, like, oh, he his hands are full. How's he gonna open up that door? Let me let me grab it for him. And then there's some people that'll just sit there and like watch me struggle. And I'm like, oh, come on, man. You know, like who raised you? So it all goes back to the parents, but um so yeah, we were we were talking about, you know, AI is that's actually a perfect use of this. Um you take all this data, like just log it, just log it, log it, log it, and then use LLMs to go through it and identify these major pain points. So um, how effective has that been? Like, can you identify, oh man, um we need sprinkler head detection in the next revision, or uh, hey, we've identified we need to make the mower deck 35 inches so it can fit behind a gate better or something like that. Like, what are what are some of the big things that AI has revealed to you that that are kind of unexpected?

SPEAKER_01 34:50

Yeah, so the uh so Jim, AI, um the way I think about it is first is you know, there's different aspects. So computer vision and LLMs, right? So on the computer vision side, this is where I kind of learned some things at Cruise. You know, how can we use our cameras to detect things like there's a hose on the ground or you know, like uh there's there's a dog toy in the city.

SPEAKER_00 35:10

I ran over a lot of hoses. Yeah.

SPEAKER_01 35:12

Yeah, yeah. And you know, they cost 60, 70 bucks, so it takes takes the you know price of the mow away. So um those are sort of sort of the foundational things that we have, and we're iterating on, we're collecting data on, we record, you know, all cameras across all our systems. And you know, we're very small. So I do all the software work in the company, so you know, we're iterating on it still. So it's not fully perfect, but that's kind of what the AI uses on the computer vision side for our for our robots. The place where it's actually been incredible to use LLMs in the in the sort of the in that part of the AI is obviously programming. So, you know, when it comes to front-end software development, even back end, like run-of-the-mill crud stuff, you know, database stuff, you know, the low-hanging fruit of are, you know, you're you can be so much more productive today than you could just honestly last year. So that's changed things completely. And then the last sort of area which you're you're touching on is when you have, you know, I call it like context shaping. I don't know if this is the right term in the LM industry, but like you have all this context, you have all this customer data. Can you use it to do, for example, a training manual? Okay, that we can. How about how about we do a sentiment analysis on every um every phone call text and email coming from every customer every single day, so we know by the end of the day which customers are the most happy or most unhappy, and use that to build a model to see which one's most likely to churn because we get the same mistake two times in a row, or you know, the language from uh their text or the the transcription of their phone call. You know, it's just uh something that's honestly never been possible. And it's that gets me very excited. Uh, but obviously it comes with limitations, but that is where I think LLMs have just been incredible for us that we can with it with a super small team, we can just provide such a high quality of service, is what's being different for first.

SPEAKER_00 37:01

It it allows a super small team to you know uh have the same performance as a as a large team. You know, the the way that we see it is uh we don't see LLMs here to eliminate jobs, we see it as a way to make the existing jobs like you know better and better. So then we can actually hire you know more people, but they're every time we hire someone and pair them with an LLM or something, they become a superhuman. So um yeah, it allows it allows the teams to be lean. And I actually think in the future, you know, if we're doing it the yard bot method where you're allowing a small team to act like a big team, I think layoffs will be much more rare in the future because companies aren't gonna overstaff and have you know 10,000 accountants or whatever. They're gonna start with two accountants and then layer on this like superhuman ability and then grow at the right pace and and hopefully you know not have have to suffer from these like massive layoffs that you see at these big corporations. So I think we're going through a churn time right now. Um, but in the future it'll be a little more predictable, I hope. Dom was saying there there's no snow in uh in Alabama. Because we were we're talking about like, could these things what else could they do? Could they could they blow snow? Could they uh you put a dozer blade on the front of it? Like have you ever thought about something like that or in different markets?

SPEAKER_01 38:22

Oh yeah, yeah. We did we were asked to build a prototype for a military base uh for snow removal. Um we didn't so we built something, but we weren't happy with it. So snow removal just has a different I I don't I guess I can go deep in there, but like it it, you know, I'm from Toronto, right? So it snows a lot. My wife's from northern Ontario, so it snows a lot up there. So I'm familiar with snow. The thing about it is it's very irregular, right? So mowing, you're doing it at like a week, two weeks, but when snow comes down at a base, for example, it needs to be cleared. There's no option and all hands on deck, but then it may not snow for a while. So it's very spiky. And um, you know, the uh yeah, I you know, I I'm not uh this part I'm not sure about. The perception system might have to be slightly more sophisticated than what we have, but perhaps you can make it work. Yeah, snow removal is a terrible business. I don't want to say that, but like it's just a different different thing.

SPEAKER_00 39:17

I mean, for our thing, yeah it's hard. I have family down in uh Mammoth Mountain, California. Okay. And if it starts snowing, like that's what you do. You wake up every three hours and you go and shovel. Because if you get behind it, you'll wake up and you you have seven feet and there's nothing you can do. You know, you have to wait till spring, basically. So yeah, snow removal is hard. Or or like, you know, here sometimes we'll get one inch, uh, and it's like just borderline, should you even bother? Um but all these snow removal companies are out there because they're trying to bill for it. Uh, and then on the days when we do get two feet, then they don't show up because they're too damn busy at you know hospitals or you know, these primary customers or whatever. That's that's a hard business. Um but Huntsville, it's so you said you had 600 customers, right? That's right. 600 lawns that we maintain, yes. Dude, that's awesome. Um are you thinking about you know other uh other markets like nearby? Or I mean I guess there's also many, many more lawns to go there. Yeah, what's what's next? Like what's uh what market is exciting?

SPEAKER_01 40:24

So, Jim, the way I think about it is Huntsville for us is sort of like a like a test case. So we're kind of every uh every part of the business from building robots to running them to the sales and marketing, to the operations, to the engineering, like all that stuff is kind of like this is uh you know, like the base case for us in Huntsville and and the financial aspect as well. Are we making money, right? So we add all this together. And for me, the objective is to take this and copy paste this is and in many as many cities as as it makes sense, right? So what's very attractive for us is any place where grass grows eight to ten months a year or more. So everything starting from Texas all the way across to Florida is really attractive for us. Um the next phase of our business would be to take this and take it to one of those cities and then see, you know, what what what things, what strategies kind of continue to work and what don't. But I do the objective is to copy it to maybe let's go to Nashville or let's go to Austin or let's go to Orlando, set up the system, set up the same system, set up, you know, and we'd run it very lean, right? So, you know, we we have a storage unit right now. So, you know, this is our office where we hang out, you know, but the the the law the robots and stuff leave from a storage unit. That's kind of what our setup system setup is. And you know, we can do this in a new city, so very low fixed costs uh for us. And um anyway, but yeah, long story short, yes, to answer your question, geographical expansion is a key element of a landscaping company. You know, all landscaping companies that are big, billion-dollar companies, they're in 200 plus cities and we want to be there.

SPEAKER_00 41:54

Wow. Um what do you think is gonna be the number one challenge uh of the first new market that you go into? Okay, so let's see. Because I have an idea.

SPEAKER_01 42:06

You have an idea. All right, let's hear our idea first. Let's hear it.

SPEAKER_00 42:09

I I think it's I think it's management. I think it's leadership. Okay. Um when we open up a new facility in a new state, yeah. It's that the the operations director, that's what we call it, the guys who run our buildings, identifying that operations director is key. Um they will they will make you or break you. And we've had nothing but good ones. But now as we're we're trying to expand into more cities, um normally we could pull guys off the bench. You know, we're like, okay, you've you've been here long enough, you've figured it out, you you have these qualities. Um you know, if if you're if you're young and you want to relocate, like we'll give you this opportunity. But what's happened is like we've been going so fast that our bench hasn't been able to keep up the pace. So we don't we have a lot of guys that are awesome right now. They just need like a couple more years to learn. Um, so once you deplete your bench, then you have to go from outside the company. And then that's a a big gamble. So we're trying to figure out like what qualities do we want to identify in these cities. Um because relocating someone is really challenging too. Like you're taking a big gamble, you're asking them to do a lot for you. Um, you know, especially if they have like wife and kids or whatever. I'm like, hey, uproot your life and go to this city that you've never been to. Um so if I can recruit people from those existing cities, then it's a little less risky. Um anyway, sorry. Sorry to give you your answer.

SPEAKER_01 43:32

But that's no, no, I I I like it. So I wasn't, you know, to be frank with you, this is a weakness of mine. I was thinking about robots and manufacturing, but humans are always the challenge. Um the only two things I'd say there, Jim, is one, you know, I uh the way you're thinking about it is how I think about it is like, can we take our existing bench and ask them to do a tour of duty in a new city and and bring it up essentially, right? So they go out, the first 60 to 100 to 200 customers they kind of take care of. Till we get local talent that's kind of ready to take on take it on. And the second thing, and this is the sort of irrational technologist problem that I have, is we have so much data. Okay, we have video of every lawn being mowed across 12 cameras and every truck being driven, you know. Can we leverage this data, perhaps, to make sure that things stay straight? But I don't know. But that's kind of the you know irrational optimism I have is we can make it happen.

SPEAKER_00 44:30

No, I I love it. We we spend a lot of time studying um Starbucks and Chick-fil-A and In N Out Burger or whatever. We're trying to figure out how how can you expand, you know, you go from 10 Starbucks to a thousand Starbucks. Like, what does their forward deployed team look like? Um, how are they identifying these new managers who are going to run that location? You know, with Chick-fil-A, they have this crazy, you know, audition hiring process. Um, it's like a lottery basically to get to run a Chick-fil-A. So I'm like, man, how do you how do you build that? Well, it starts with like a really cool brand, a really cool culture that people want to get into. And then so I'm like, okay, so everything that we're trying to do to expand now actually started, you know, eight years ago uh in order to attract that talent. So I don't know what the magic key is. I'll I'll let you know as soon as I figure it out.

SPEAKER_01 45:22

So thank you. But I gotta also say, you're like about 10 steps ahead of me. My problems are much smaller. But uh but yeah, sounds like I mean, congratulations, obviously, everything you Jim, you've accomplished that second step, but you guys are way ahead.

SPEAKER_00 45:37

So learning, man. Overnight success takes 10 years. Um so we we got a ways to go. Um actually, and that's that's kind of a great place to kind of wrap this. Uh I was gonna ask you our final question here. What's what's one piece of advice that can be about anything that that you want the audience to walk away with?

SPEAKER_01 45:58

Um don't give up too soon. That's what I would say. You know, a lot of times I've done something and then said, Oh, I'm gonna give up on this, like Bitcoin. You know, um and then exactly. But maybe not Bitcoin's not the perfect example, but you know, where I've kind of just made a decision too soon, but just gotta stick out stick it out and iterate, and then things can get better. So cool.

SPEAKER_00 46:21

I love that. Divya, thank you so much for your time. Yeah. Um, if anyone in the Huntsville, Alabama area is interested in getting one of these things in their yard to flex on their neighbors, uh, where should they go? Uh just yard.bot. Yeah. Yard.bot. Oh yeah, go check it out. Um editor Logan will put a bunch of photos here of the awesome robots. Um dude, thanks again for your time. But this is where I say my outro, which is actually Dom just hitting a button.