He built a farming robot that shoots weeds with lasers— then closed $10M in pre-sales before shipping a single one. | Paul Mikesell, Founder of Carbon Robotics.

Paul built a startup that was acquired for $2.2B. He worked on AI at Uber and Oculus. But when a farmer told him about some of the problems he was facing, he quit and went all-in on farming. 

He built a robot that attaches to the back of tractors, uses computer vision to identify weeds and lasers to shoot and kill them. He sells each robot for $1.5M. He's sold nearly 100 so far— & generated nearly $150M in revenue.

Whether you're working on a hardware startup or not, you'll want to listen to this episode to see how to do proper customer discovery, how to raise pre-product, and how to get millions of dollars in pre-sales without having to ship.

Why you should listen

• Why hardware is hard-- but easier than it's ever been.
• How to tell a story to raise $10M pre-revenue
• How to set up milestones for fundraising
• How to generate $10M in pre-orders before shipping product
• How to get a sense of product market fit even before the product is fully developed

Keywords

Carbon Robotics, deep tech, AI neural nets, deep learning, weed control, farming, manual labor, lasers, research and development, funding, seed round, prototype,  Series A, pre-orders, manufacturing challenges, revenue

Timestamps:
(00:00:00) Intro
(00:02:09) A Background in Deep Tech
(00:05:50) Problem and Solution Interest
(00:07:35) Find the Passion
(00:07:37) Get Behind the Mission
(00:09:39) Back to Seattle
(00:11:40) The Epiphany
(00:16:50) A Human Affair
(00:18:00) The Money Flows to the Best Storyteller
(00:21:45) A Farmer's Weed Control Dilemma
(00:25:56) Weed-Shooting Lasers
(00:28:11) Funding the Prototype
(00:30:13) Hardware is Hard But Rewarding
(00:36:46) Why Laser Weeding Works
(00:41:19) Series A
(00:42:51) Robot Autonomy
(00:45:42) Pricing
(00:48:12) Telling the Story After It Happened
(00:51:44) Fighting Through Disaster
(00:54:50) Finding True PMF
(00:55:41) One Piece of Advice

Transcript

Pablo Srugo (00:00):

So this is probably one of like the coolest startups that we've had on this show. Paul, the founder of Carbon Robotics, built a robot that sits on the back of tractors and uses computer vision and AI to find weeds on farms, and then uses lasers to shoot down the weeds and effectively help farmers grow plants without having to use chemicals or herbicides because they don't need them anymore. Because this tractor is shooting down all the weeds, he was able to get $10 million worth of pre-orders before shipping the first product. And to date, he sold nearly 100 of these machines at an average price of $1.5 million. We rarely have, I don't even think we've had any hardware startups on the show, but this one is definitely worth checking out. Before we start today, could you just do me a really quick, really small favor. Can you just take your phone outta your pocket, open up Spotify, open up Apple Podcast, open up whatever layer you're using, and give the show five Star.

It's not just for me, it's for all other founders. If you like the show, if you get value out of the show, when you review the show, other founders are way more likely to find it. So this is like your philanthropic moment of the day. And yes, if you're thinking to yourself, is this really for me? Is he really talking to me? He must be talking to somebody else. No, I'm actually specifically talking to you. So please take a handful of seconds. If you like this show, rate it, give it five star, write something. Whatever you do, you're gonna be helping, not just me, but so many other founders. Thank you. And I hope you enjoy today's episode. Paul, welcome to the show.

Paul Mikesell (01:37):

Yeah, you bet. Thanks for having me,

Pablo Srugo (01:39):

Dude. So you might have one of like the, the coolest sounding companies, <laugh>, that we've had on the show. 'cause You have robots that that kill weeds, <laugh>. So like that, that's a good, good place to start with. I mean, we'll, we'll, we'll dive into kind of how it all, how it all happened, how, how you got there, building hardware. I know because I did it or tried to do it in my last startup and it's exceptionally hard, harder than an average startup. But let's start there. Like maybe let's just start with your, you know, walk through kind of how you got there. But, you know, let's start with your background. Like what, what were you doing before you started Carbon Robotics?

Paul Mikesell (02:09):

Well, yeah, I mean, my background is all Deep Tech. I have a computer science degree from the University of Washington, which is, you know, up, up in Seattle. And so I've been a, a software engineer originally, and then have done a lot of work on the electrical side and just other things around robotics for quite some time. You know, mostly as kind of a hobbyist, all the electrical stuff started with in the UAV space, you know, back when it was first starting to look interesting in 20, 2013 timeframe, something like that. So anyway, I've always kind of had this blend of software and hardware, but most of my work projects and always been pure software. I started a, a what was essentially a distributed file system company called Isilon Systems. And we sold that company had a, a $2.2 billion IPO in in long, a while ago in 2006. Hmm. And

Pablo Srugo (03:05):

And you were the founder?

Paul Mikesell (03:06):

I was the founder of that company, yeah. And and, you know, wrote all the first code and filed the patents and built the engineering team and, and all of this. And then I was at Uber relatively early on. I spent some time at Oculus in the VR space. 

Pablo Srugo (03:23):

What'd you do at Uber, by the way? I'm curious, what, what deep tech are they doing there on the side?

Paul Mikesell (03:27):

Yeah, yeah. So originally it was scale, helping to scale out their backend system. 'cause We were, when I started there, we were basically on a single database system, but then I got in really into ai, neural nets and deep learning there at, at the, for the last probably 18 months or two years that I was there. And, and spent a lot of time on these AI systems. It was right when TensorFlow was, was kind of the main framework, but Pie Torch was starting to become a thing

Pablo Srugo (04:04):

And Uber was using it for what? Like just for, for like optimizing their own, like, their, their, their marketplace.

Paul Mikesell (04:09):

So the, the most, I think, visible aspect was the self-driving car stuff, but all of the idea around what was happening in the world around, not just the cars, but we had many other robotic or semi robotic projects that were trying to figure out what was going on in the world around them. And all of that stuff was all some form of neural net. Uber was an interesting business because almost the entire time I was there, we were supply constraint, which is to say constrained by the amount of drivers. And so there's an incredible amount of optimization that goes into predicting what's gonna happen, making their lives easier, and also doing the matching in a way that would allow better utilization of the drivers. And so that was there, there was a lot of that stuff. Uber actually had at one time had had quite a large investment in ai, and this was in the Travis Kalanick days, which, in which Uber was I think, a lot more innovative and a lot more willing to try new things, right? It was in the Travis days.

Pablo Srugo (05:20):

Do you work closely with Travis?

Paul Mikesell (05:22):

I will tell you, I was a fan of Travis in the way that he ran that business. And, and a lot of the ways that he liked to operate, I reported directly to the CTO and the CTO reported to Travis.

Pablo Srugo (05:35):

Makes sense.

Paul Mikesell (05:36):

So I, you know, I, I liked him. I had, I was in a number of meetings with him. I feel like I learned a lot of things from him. I don't know if it, if I would say it was close by

Pablo Srugo (05:44):

The way, maybe a tangent question, but after founding a company that goes IPO, what makes you decide to take a job anywhere?

Paul Mikesell (05:50):

Yeah. Well, I mean, really the motive, my motivation with founding Isilon was the same as my motivation with joining Uber, or joining Oculus or founding Carbon Robotics. That motivation is driven entirely by the problem and target solution interest in the market and the product. And so I don't really, if, if I hadn't founded Isilon Systems or if I hadn't founded Carbon Robotics, these are, those are the places I would've wanted to work at anyway. So, at least for me, I feel like the better companies are the ones where you started it because you saw a problem and you had a great solution you wanted to activate on whether or not it was actually your idea originally, or whether or not you founded it. Right? So I felt like Uber was the same thing where I was living in San Francisco at the time but, and, and had previously lived in Seattle and, you know, had, had been in, I I you would call metropolitan, you know, major metropolitan cities or, and just saw the, how hard it was to not have a car and want to be able to get around the city.

Paul Mikesell (07:01):

And, you know, there's all the issues with parking spaces, right? How much of our cities are taken up with just parking, driving is a huge issue, you know, especially being a young gentleman, you know, trying to go out on the town and you have a couple drinks with your friends or some ladies or whatever, and then you wanna get home, what's the safest way to get home? Calling a cab is, you know, nobody's idea of a good time. And so the Uber thing for me was like, oh, I was, I was a fan of Uber. I was using Uber. I saw value in that technology.

Pablo Srugo (07:32):

What years are this like that you're working at Uber? Oh,

Paul Mikesell (07:34):

I, I went to Uber in 2014. I mean, I have very fond memories of everything I've done, and I think that's an important aspect to somebody's career, right? Is to take a job because you're passionate about what you are doing and working on, not because there's some interesting title you're shooting for or something like that.

Pablo Srugo (07:51):

Well, I think to the extent that you're seeking like meaning in life, which is probably like the number one thing, you know, solving a problem is a great way in the world of work to go and find that. And whether you solve a problem, like you said, because you're working within another company or because you decided the best way to solve that promise to do it yourself is secondary. A hundred percent. But I feel like that way you're unlikely to be lost. I mean, there's a lot of people who go out, they work with Fang, maybe make a bunch of money, and then they're either bored or dissatisfied or, or whatever, right? And the bank account's not gonna help.

Paul Mikesell (08:20):

You bet. Totally agree. Totally agree. Also, you can make more money. I mean, if it just comes down to dollars and cents, you can make more money going to an early stage startup helping to grow that company than you can at something like Fang. There's certainly more risk in that, but I mean, I don't know, you know, I've been at, you know, five startups and had four been through four IPOs. So you can, you can actually do a good job of this. You just, what you need to do is make sure that the company is something that you believe in and you understand why you're there, and you get, you get behind the mission. And I don't think I'm special in my ability to choose good companies. I'm, I may be special in my my ability to have conviction.

Pablo Srugo (08:54):

Well, you're a pretty good picker, you know, four to five <laugh>, maybe you should be a VC <laugh>.

Paul Mikesell (08:58):

Well, I mean, you know, but, but I think a lot of it is you, you have, have, have con conviction around your beliefs, right? It's like I can identify what I believe in and then activate on that and put my whole, put my whole time into it. Being able to do that can be challenging and scary. But once you kind of, once some of those earlier bets start to pay off, you realize you've made good decisions and you start to trust yourself a little more. I, I would hope kind of everybody would have that opportunity at least a couple times in their career to feel like I have, you know, I had a strategy, I, I activated on it. I threw my whole self into it, and it paid off.

Pablo Srugo (09:33):

So what happens after Uber? Like, what, what happens kind of between Uber and Carbon robotics in, in 2018?

Paul Mikesell (09:39):

I wanted to go back to Seattle. This is controversial, but not unique. I felt like San Francisco was getting kind of played out, honestly. I was, I was, I wanted to find something else interesting to do. Everything seemed like it was kind of a copy of a copy of a copy some other, you know, SaaS for SaaS companies doing SaaS deep learning and, and ai, it became a big thing at that point, right? This is 2018, but then every company was like, oh, we'll help you optimize your marketing spend or whatever. I was like, this is, I can't, I couldn't care about this any less, you know? So I was like, okay, I'm, I'm, I'm gonna get outta San Francisco, it seems. So, and then just coincidentally, Oculus had a group up here in the Seattle area called Oculus Research that were doing some cool things, combination of hardware and software.

That was kind of a good experience for me because I applied as a software engineer, and then I did the full interview loop, you know, a guy called Michael Abra, who if you're into computer graphics or the history of the Windows UI system or, or, or vr and, and, and specifically Oculus, you may know the name, but he interviewed me personally with a bunch of coding questions on the whiteboard. And that was and he is known to be a tough interview. He is a tough interview, and I feel good, you know, I felt confident about the fact that I was able to go through those interviews and just say, Hey, I'm gonna, I'd like this project. I I want to come write software. So that was cool. That was very fun. I loved it. I had a great time doing it.

Pablo Srugo (11:09):

And this was before the the Facebook acquisition?

Paul Mikesell (11:12):

This was after the Facebook

Pablo Srugo (11:14):

Acquisition? This after. Okay. Yeah.

Paul Mikesell (11:15):

Okay. Yeah, it was, it was, it started off as Oculus Research, and then it became Facebook Reality Labs. I wasn't really a fan of the rebranding. I've, no, I'm not. So I'm not really a fan. <Laugh>, you wanna, here's a funny story. I went to the I was like, yeah, I'll do some vr, this will be fun, right? I was just looking for some fun code to write, and I was into vr, I liked vr, I thought it was a, I liked it as a gaming platform. I could see the other opportunities, other things that could be done for it, vr, ar, you know, et cetera, <laugh>. But Facebook, I mean, I, I didn't care at all about Facebook, right? What do I, what do I care about Facebook, right? That's not, so anyway, so I I did all the interviews. I got the gig <laugh> up in Seattle, and then they go, okay, but you gotta go down to Palo Alto and do your Facebook onboarding.

Paul Mikesell (12:07):

Right? Okay. You know, cool, do that. But I, so I go down there and the next morning, you know, you go to the onboarding thing, and, and so they've got everybody there and it's, I don't know, maybe a hundred people, a hundred people, 99 of them are brand new Facebook employees. And then I'm sitting there and I'm the Oculus guy, and they go, okay, what we're gonna do now, as part of our onboarding, is we need you to connect your personal Facebook accounts with the corporate Facebook things, right? So that you can I don't know, get whatever benefits come with that or whatever. Just I, that's how they track their employees or something. So I, so there, you know, everybody pulls out their Facebook app, it's like, beep, beep, beep, connect it up. And I was like, well, I don't have a, I don't have a Facebook app.

I don't have an account <laugh>. And they're like, how do you not have a Facebook account if you work at Facebook? I was like, Hey, I'm an Oculus. You know? So anyway, that was kind of funny. It was this big, like, to do <laugh>. They were a little pissed at me. And I was like, look, man, I was like, you know, I don't, I don't care about, I don't give a shit  about Facebook. I'm you know, I like the Oculus stuff. Anyway, that was funny. But I will say, as a, as an environment, Facebook took, took really good care of their pe care of their people. I have respect for the way that Zuckerberg, I think treats his folks, his software engineers. That was my experience. O Oculus, I, you know, I wish them well. And I, and I like the VR space quite a bit.

It was in the long term, not for me. I think I, I maybe got a little disillusioned about the distance between the reality of the abilities of the current gen hardware, you know, versus where I wanted things to go. And also, it's a big company. I mean, Oculus is part of Facebook was becoming more Facebook and less Oculus every day. And that, so that wasn't, that, that wasn't my style, you know? So that was, that was the deal there. And so then I went and really just wanted to figure out, okay, so the, the epiphany was the ai neural net systems that we had, and we knew how to build, could understand the world around them, right? This was, we did some of it at Oculus. We did a lot of it at Uber. This is a neural net that can, that actually knows what's happening around it.

Paul Mikesell (14:16):

Through sensor fusion. This technology is, is, I knew how to build it. Is, is the frameworks are good, the training we know how to do it, right? You need to get training data. But, but these techniques are, are are there. And so I just happened to be selling an airplane at the time, and the, the person who bought my airplane was a farmer, and we just got to talking about everything farming. So this is for, this is maybe a lesson in enjoy your times, meeting new people, because you never really know what might come out of it. And so this farmer, this is a guy called Shea Meyers, his farm is called at Yhe. They're over in Idaho. And I spent a lot of time talking to him about farming and what they were trying to do on their farm. And what I learned through this process was that farmers are a really strong business people.

Paul Mikesell (15:11):

They have to be, because all the ones who aren't, have gone out out of business, and even some of the ones who are really good have gone out of business, right? It's a tough environment. They're incredibly innovative and willing to try new things. So that means not afraid of automation, not afraid of computers and robotics, right? And that there's a huge opportunity there because nobody was really focusing on it. There's, there's been this, there's a divide between traditional venture capital money, which comes from places like San Francisco and Seattle and where farming happens, which is in places like Salinas, Bakersfield, Cuyama, you know, El Centro, right? These names that if you're just in tech you've never heard of before, there's a distance there that is geographic. There's a, and there's a physical distance. But there's, and there's even like a, in both, in all of these examples, there's even a mountain range, you know, in between west coast college town, VC, money, you know, mountain and then east side of the state, flat, much hotter rural lot of agriculture. And those two sides don't really meet that much. And so what that means is all the Silicon Valley money is just kind of floating around, you know, making more, what I was complaining about before, making more, you know, SaaS software for SaaS companies selling to SaaS companies. And it just kind of goes around and around and around. So there's this giant opportunity over here, and there's all this money over here.

Pablo Srugo (16:50):

It's an interesting observation just on the, on the money piece, just to go off on tangent, because, you know, you would think that investment should be rational, and if there's opportunity somewhere, somebody will find it. And I think in the long term, that's probably true. But the reality is at seed stage, which is what you need to get the thing going, so much of this is about this kind of sense of conviction, right? Like the person writing a check has to really understand, not just like, like the team and all that, but like resonate with the problem set. And in there it's a, it's a full human affair. And so obviously you're gonna resonate, you have biases and you're gonna resonate the problems that you understand. And if you're, you know, living in SF your whole life or wherever, like Toronto, you're in Canada, whatever it is, like, you're not close to something like farming. The likelihood of somebody pitching you that and it making sense to you is much lower than like, you know, the SaaS force, some SaaS company, because they just heard another board meeting that some SaaS company was struggling with that problem set. And they get it. And so totally get it. That that totally makes sense to me.

Paul Mikesell (17:46):

Definitely true. So you said some interesting things there, right? So one of the things you said is the, the VCs want to invest in this thing 'cause they just heard another company talking about it as a need. That's definitely true, right? The, the problem space that they see is a problem space they experience, and that usually comes from their tech companies that they're invested in. So they're like, oh, I heard that this company had a problem with their SaaS software and this other company's developing SaaS software for that company. Boom, easy investment, right? And, but that means the horizon of opportunity for them is limited because they only see that world. You're totally right about that. The other thing I will respond to from what you said, the idea that somehow this investment capital and these decisions are rational is immediately once you start raising venture money or look at some of the investments that have been made <laugh> you'll realize that that is not the case. <Laugh>,

Paul Mikesell (18:38):

If you ever feel like, I don't know if my idea is good enough to get venture funded, but you believe in it, just go read about some of the other weird stuff that they've invested in and then you'll get, you'll gain your confidence back. <Laugh>, how's, how's that? You know? Yes. Yeah. So it's, it, it, so a lot of it is storytelling. It really is raising money and presenting the problems, fa presenting problem, clearly outlining the opportunity, and then how your solution maps onto that. And I think you could, doing that is, is the most important. And there are a lot of good ideas that never get funded. Not because there's something wrong with the idea, just because the founders never figured out how to tell the story appropriately. And that's, that is really a lot of it. That's what raising money is all about, is you're, you're trying to take somebody from where they are and where they are by definition is not at your solution, right?

Paul Mikesell (19:28):

You're bringing that to them. So you're trying to take them from where they are and this journey to understanding where the opportunity is and then how what your doing fixes that, addresses that. And so you, you gotta remember that they don't already know the answers to these things. And I think sometimes people, when they pitch VCs, they feel like they have to just pump their <laugh> their pitch decks with so much information. So that, I don't know if they feel like that's necessary for credibility or they feel like they wanna sort of tell the whole story and not leave anything out or whatever it is. But you've gotta remember that you're, you're telling a story here and the all of the little details, those leave that for follow up questions. Leave that for due diligence. Make sure that your, your story arc makes sense. And when I review, you know, pitch decks for people trying to get fundraising, I mean, that is always the biggest mistake I see is people just, their pitch deck is like chock full of information, but you lose the story, you lose the thread of the story there. You've gotta make sure that story is clear. And so for, if anybody's hearing this and they're trying to raise money and they feel like they're, it's a, you know, they're having a tough go of it, go, maybe go back and look at your deck and think about is there actually a story here, or it's just a bunch of slides with

Pablo Srugo (20:53):

Data. Absolutely. I mean, that's a, it was like, I was reading recently, that was one of the, the Don Valentine quotes, which is like, the money flows to the best storytellers. And, and honestly, at the seed stage, it's, it's totally true. For better or worse, I just feel like that's just how it works. And, and I think to your point about pitch decks, like it's much easier to add more and say more than it is to refine and cut back. And, and, and refining and cutting back is where you end, end up getting to like the essence and the simplest way of communicating what you want to communicate. And once you have that, you're, you're well on your way to telling a good story.

Paul Mikesell (21:22):

That's right. Is that a Don Valentine quote? I didn't know that, but I would believe that. I would believe

Pablo Srugo (21:27):

It. Yeah. I don't know the exact words, but it was something around that the money flows to the best storytellers. Yeah. And so walk me through, so I mean, you've, you've met these, these farmers, you feel like there's an opportunity there. You feel like AI could do something for them, but we haven't even talked about what the problem set is or even how you go about it. Like you're still at Oculus at this point, or kinda what's going on?

Paul Mikesell (21:45):

Yeah, <laugh>. Yeah, I'm still at Oculus. So I knew that we, we had a list of 15 different ideas for farming. This was after several conversations with this Idaho farmer. And we had this one long, it was like three hour long thing. We just went through all these ideas and the top of the list was weed control, killing weeds in the field. So the problem is you're growing crops and it's an organic environment, so you put your seeds down, everybody knows this, you water 'em and fertilize 'em, boom, you get your crops. Well, that's kind of true, but also everything else in the environment wants to take that water, use that fertilizer, use that nitrogen to grow and seeds blow into the field that come into the water supply of other plants. And when other plants show up, they take those nutrients and they also crowd the space and they make it difficult or impossible to harvest, right?

Because you grow a field of something, if you've got a bunch of other stuff mixed in there, it makes it very hard to harvest, right? You don't want to get a, a bag of, of salad. You don't want in your bag salad to get a bunch of weeds in there, right? Things that you wouldn't typically eat in a salad. That's why farmers try to grow these things in a way where it's a, it's a dense environment of the thing that they want so that they, all of the water, all of the nutrients and the harvesting is efficient. Okay? So weed control is a big problem and the farmers have to do have to do it. It's a mandatory requirement of growing food. 'cause Otherwise you just get a mess. They have a couple of options. They have herbicides, which we're all aware of, which is a chemical process of disrupting this, the plant growth cycle.

Paul Mikesell (23:25):

There are many different ways in which they do it, but all of it has some residue runoff, what's called herbicide drift. All of these chemicals have some negative health effects for people. And the long-term, negative health effects for farmers. 'cause Farmers are the ones who are exposed to these chemicals in the largest degree, right? They spend their whole careers exposed to this stuff, and they're the ones who suffer the health effects the worst. And the production of these herbicides has an incredible amount of greenhouse gases that gets released as well. The production of these herbicides is detrimental, destructive to the land around it. It's not good. But we do all of this because we don't have a, a better way to control these weeds. And then the other thing that people do, they'll just throw people at the problem, right? So they hire farm labor almost entirely immigrant workers who are on these H two A guest worker visas. For those folks, it is a really difficult, dangerous job and, and can be a dangerous environment and situation 'cause they are having to travel across the border, get their paperwork in order. There's a lot of cost, time and expense spent on this travel. And then they're far from home doing this work, which is not, you know, which can be dangerous. I mean, we've had field workers die from being out in the field in the hot sun pulling weeds. It's not a, it's, you know, it's not a great job. 

Pablo Srugo (24:53):

By the way, how did you find out all of these problems? I mean, you had that one discussion with the farmer, which I think opened the door. But I'm curious on the process of you figuring out all of these sp specific problems that farmers are having when it comes to weeds and weed control.

Paul Mikesell (25:07):

We had several conversations and then I spent some time out on his farm with him. So it was more than just talking. We, you know, we actually went out there and spent some time. And, and also he introduced me to other farmers. And so then I spent, before I built a,

Pablo Srugo (25:20):

This was you, you had already quit your job and you kind of were doing this full time or as you were working?

Paul Mikesell (25:25):

Yeah, at this point I had basically, I had quit Oculus and I was just spending my time flying around and visiting farmers. Yeah.

Pablo Srugo (25:30):

How long did you kind of fly around for? How many farmers did you speak to?

Paul Mikesell (25:33):

I think the initial set was probably three farms spent, you know, a couple days at each of 'em, right? So it wasn't like months and months and months. It was definitely, but it was definitely a couple of weeks. Pretty intense process of discovery. Yeah, that's right. Took a lot of pictures, took a lot of video, you know, learned a lot of stuff.

Pablo Srugo (25:50):

Were you asking them by the way, like high level, Hey, what, what are your proms? Or did you kind of start going in already on, Hey, how do you do weed control?

Paul Mikesell (25:56):

At that point? I had this thesis about weed control being the, at least the first thing that we wanted to target. So then at that point it was talking to 'em about, Hey, how do you do weed control? And seeing if I was hearing the same, the same experiences, the same problems, the same limited solution set the same difficulties. Okay. And it started to track, you know, I was hearing the same thing. You know, I, so I, so I had, I saw this problem, I knew how to build a neural net, an AI system that could in real time with cameras identify what was happening. And so now the question is, well, how do you get rid? How, what do you do? How do you get rid of the weeds? And lasers was really just thinking about the ways in which you could destroy plants effectively.

I don't even remember, I just had the idea for lasers at one point. I think it was because, you know, deploying heat directly to a plant in the most efficient way. And I knew, I knew a little bit about lasers, although not that much. But I, one thing I did know is, is the laser is a way to put targeted energy directly at a point. You know, we had done some laser stuff at Oculus, of course, maybe that was where I learned the basics of some of this stuff. But, so I knew this general concept, and then it was just time to learn about the specifics of laser technology and would this really work very quickly? I figured out how to take optics and control the beam path in two dimensions and put the beam where you want. The first version of that thing was something I built in my home office.

Paul Mikesell (27:26):

I still have the videos of it where I was basically doing like a little mini laser show on my wall just with mirrors and servos. And, but once you, you know, once you start getting that working, you imagine, well, what if this laser in, you know, instead of 20 milliamps was hundreds of watts, and what if these servos instead of these cheap, you know, UAV servos were actually scaled up really fast industrial servos. And what if instead of raspberry pie, I've got big computers with GPUs, you know, this is a solution that I could build. Would that pencil out? Would that work? And so the rest is just exploration a lot of math and trying stuff pretty quickly. We got something built that we could take out into the field and shoot some real weeds in, in real farm

Pablo Srugo (28:11):

Fields. And this was all, this was pretty funding still at this point.

Paul Mikesell (28:14):

So what happened is I quit Oculus, I did the basic work to identify the problem. I had built very simple versions of the software and had shot some weeds in my backyard and decided I wanted to go try this. So I put some of my own money into it, and I got a little bit of venture funding just enough to basically get an office, hire a couple people, and kind of build a first prototype. So our first, that was our seed round. Yeah, that was 900,000 in a seed round is what it was. 400,000 of that was from me. 400,000 of that came from this fund called Bolt, which was a, a hardware focused early stage fund.

Pablo Srugo (28:58):

And what was the objective with that? 'cause That's a huge difference between hardware and software, is, you know, this idea of an MVP, I mean software. You build something, you put it out, you get some traction. What's the objective with a $900,000 round when it comes to building huge machinery?

Paul Mikesell (29:11):

Yeah, that was enough money to basically get a prototype built and put it in a field and see if it worked. That was the whole objective. So hire a couple, hire a couple people, build a machine in this case, something that we manually pushed all the lasers, the computers, the servos. And

Pablo Srugo (29:27):

By the way, that machine, you were, you were pushing it, or was that like, is it on some sort of like, moving system or how does it move around the farm?

Paul Mikesell (29:34):

We were pushing it, we were manually pushing it. So that was me and a guy. There's a guy called John May, who is our vice president of product today who started off as a mechanical engineer. And so that was me and John May pushing that thing in the field. So at the time, John did all the mechanical work and I did all the software work and the electrical was just like cobbled together. But we got that thing working, we pushed it around, so, you know, yeah, me and John out there pushing this thing, <laugh>, we called it at the time, we called it vertical cart because it was a cart that you pushed and the laser was mounted vertically down, pointing at some optics and all that stuff. So yeah. So we're out there pushing vertical to cart.

Pablo Srugo (30:13):

This is what people don't realize with hardware. Like, hardware's hard because it, it forces you like the whole cycle. I mean, you push out software, you're in your, you're in, like wherever you are, you push it out, like it breaks, you fix it like it's back up and running. <Laugh>, this is like, you go out, you push your cart around the field, it's too hot, it, you know, falls, whatever, <laugh>, and you gotta deal with all these problems and <laugh>, it's crazy. The cycle is so much slower and so much more painful, right?

Paul Mikesell (30:35):

It's exactly what it was. Yeah. And it's like, oh, the camera lens is like not able to focus at the right distance, and we have to move where it's mounted and the, the one of the wires, you know, stripped out and we have to go find a soldering gun and try and remember where it goes. And I

Pablo Srugo (30:54):

Remember, I have got to tell this story because I'm sure you've had things like this similar, like I remember <laugh> with Gym Track, this was our first ever, we shouldn't have done this, but we for whatever reason signed up to go to like a conference and have like a booth at a conference way before we were ready. And I think it was just like a timeline, like, let's just make it happen. And we worked backwards from that. And I remember the hotel room before, like the conference day, and I had to, I drove there, I think I had to, I ended up leaving, I was going from Ottawa to Toronto, a four hour drive, and I ended up leaving like 10:00 PM because I had to grab some like piece that the tech team had forgone and it brought it over. And then they were like <laugh> in the room with a soldering gun in the bathroom, just putting things together. It is just, it's chaos. <Laugh>, it's complete chaos.

Paul Mikesell (31:36):

Yeah. It's total chaos. Yeah. And it never, and it, you know, and the thing, and it's usually the stuff you weren't thinking of that breaks, you know, it's usually the things that you were worried about, the important part. But nothing works without the simple pieces and the simple pieces, the simple pieces break, you know, that's the thing. Yeah, hardware is hard, but it also, I think it's more rewarding because everything you do in life is the result of some physical process, right? You and I are talking because you have a microphone and a camera, and you're on some computer. I have this computer, I'm gonna we're out on a farm trip right now, but we're going to, you know, drive some trucks around later and go see some machines or whatever that makes the food that you're going to eat in your grocery store.

Paul Mikesell (32:22):

So most of the things you interact with are physical, are physical and some of the largest companies in the world produce physical goods, right? You've got Apple, you've got Tesla, you've got Nvidia, right? All of this stuff around you. It comes from some physical process, but the, but the software companies are the ones that, you know, people think about maybe when they think about startups. So why is that? Well, part of that I think is just that when software really became a thing that we were capable of, the, there was this just this huge opportunity, you know, I'm talking about in the, in the, in the nineties now, the early nineties where it's like, oh, there's actually businesses you can build here. And it took a while for that to catch up, and I don't wanna say saturate, but at least catch up and get to some amount of maturity.

Paul Mikesell (33:12):

So I think we're at the point now where everything from now on, it's gonna have to be more of a combination of software and hardware. We're, we're, I'm not saying that there will be no innovative pure software companies. Of course there will, what I'm saying is there will be a higher proportion of hardware companies, because software is sort of caught up to the, to the point at which it's potential is starting to get achieved. And now actually I see a deficit in, in hardware because these AI systems are capable. Like I've, as I've been kind of trying to point out of understanding the world around them, and that's what enables the hardware to do more, more interesting stuff. The hardware's gotten commoditized, right? It's gotten very cheap to buy stuff, right? You can, you can go buy yourself a, a raspberry pie right now. You can go buy yourself, you know, any kind of servo you want to control anything you want, you could go to, you know, DigiKey right now and buy any, any manner of electronics for, you know, you could buy a, a little integrated circuit for 80 cents that could do amazing things that 20 years ago would've cost you hundreds of thousands of dollars, right?

Paul Mikesell (34:22):

You can design PCBs to do very complex things for you electronically with free open source software right now today and have it produced and show up at your house, right? These things give you the hardware platform where your software can run. And, and in the end goal is to do things in the real world, right? And you think about what you're gonna do with your time today. There are so many opportunities now for the hardware enabled with these AI systems to do the kinds of things where previously it was human activity that was either difficult, dangerous, rote boring, right? So I'm talking about picking weeds with our company, but there are many other things, you know, I like to see people do stuff like dish washing robots, right? Food making robots, window washing robots. Or I made a version of a window washing robot at one time that climbed and used these little suction cups and stuff.

Paul Mikesell (35:20):

These are all the kinds of things that are gonna start showing up now because the commoditization of the hardware, the abilities of the software and the AI systems, we should be able to move to a world now where a lot more of what's happening around us in the physical world is, is happening based on robots built by startup companies. And I, you know, and I hope we'll continue to maintain a leadership position in that, at least in agriculture, we'll probably at some point be moving into other things. So, I mean, I'm just, I guess your, your, your observation was that hardware is, is harder, and I

Pablo Srugo (35:55):

Hardware's hard, but it's also never been easier <laugh> in a sense.

Paul Mikesell (35:58):

Yeah, there you go. That's exactly it. You summed up my last like 15 minutes of soliloquy with that one phrase, <laugh>.

Pablo Srugo (36:04):

I don't have original ideas, but I can summarize other people's <laugh>. Well, I was gonna, I was gonna ask going back to that so that you're pushing this, this this card around killing weeds and, but high level, I mean, obviously there's issues, but it, I I assume it's working.

Paul Mikesell (36:18):

Yeah, yeah, it's working. That first laser that we were pushing around was killing weeds in the field. I wouldn't say it was fast. And it, and it wasn't easy, right? The pushing the cart around was harder than manually picking the weeds, but we could see the opportunity there and the capabilities, and we learned a lot. So one of the first things I thought that you, to kill the weed, what you do is you shoot it with the laser and you just destroy the whole thing, right? That

Pablo Srugo (36:41):

Was, and does it go down like that was, I was thinking about like the root itself. How do you get at that?

Paul Mikesell (36:46):

Right? So this is the thing is that we did this version where we just, we realized pretty quickly we weren't gonna pen physically penetrate that far down into the ground. And we, but we were obliterating the whole, the whole weed on top of the soil and they were dying. So why is that? Why is that? And so what I wound up doing was I went down to the University of Washington, which is my alma mater, and spent some time with the, the horticulture, an area in the biology department that knew about plants and and just started asking questions. And the cool thing about our public universities is there's a lot of really smart professors there who love to spend time with curious people. And so this very nice lady, and some of her colleagues spent some time sitting down with me, and I just kind of described what we were doing and what we had been seeing.

Paul Mikesell (37:39):

And she explained to me the way that this works and why she thought it was working. So the idea is that all of these plants have a Mary stem. But what is that? So you, so the way that you, like as a person, you have stem cells, you've heard of this term before, probably stem cells, and they, what are stem cells? Well, they're the undifferentiated growth cells that can become just about anything, right? So when you're a, when you are, you know, earlier in your development, when you're a small recently fertilized egg, what you have is a bunch of stem cells that will become the different parts of your body. And they're, and they're activated through hormones and, you know, heat and the environment, all the stuff. But basically that's where you come from in plants. It's the same thing. These Mary thematic undifferentiated growth cells, that's what is the Mary stem.

Paul Mikesell (38:35):

And what we were doing was burning out that Mary stem. So that was the original understanding of why laser weeding works. And so now you will hear people talk about using lasers to shoot the Mary stem on plants. We are the only ones doing it. We're the only ones who have a version that is commercial and works. But people have tried stuff in universities and other things since then, and they have experimented with this idea about burning out the Mary Summit. That's become sort of the way that people describe Y laser's works now. But I think the first time anybody in the world realized that was me talking with that lady at the University of Washington and her explaining to me about the way the Mary stem works. And then I brought her a bunch of examples of plants we had shot, and she got a microscope and like a little knife, and she was cutting the plants open, showing me what was happening.

Paul Mikesell (39:26):

That was cool. So I learned a bunch that day about why this was working, and then we got our own set of microscopes. And as we were shooting things, we started going through this and dissecting and looking at what's happening. So here's what happens. The Mary stem is in there, it's got all these plants. There's a, there's a carbon cycle, some chemicals inside the cells of the plants itself. There's a chemical called rubisco, which if you, if you've ever learned anything about photosynthesis, is, is an, is an important part of the carbon cycle of the plant. And when you, when you explode those cells with the lasers, the cell membrane explodes and you've leaked all that fluid all over the inside of the salt. So the cell walls are no longer intact because you've pumped all this energy in there. The intracellular fluid is all over inside there, right? So the, the outer skin of the plant is now sort of holding together this massive fluid that's leaked out of all the cells. And if you hit all the Mary stem cells, now the plant can't grow anymore. Wow.

Pablo Srugo (40:21):

That's crazy.

Paul Mikesell (40:22):

So yeah, the roots are still there, but there's nothing, there's nothing left to grow. And so then the, some of the roots will die off and they'll become more nutrients for the stuff you're actually trying to grow, all this stuff. So that was what happened. But maybe this is another example of enjoying and appreciating the time that you get with people around you who know things that you don't know. So that was the, that was the, that was kind of maybe the second time that happened in this carbon robotics journey, right? The first time was with that farmer. And then the second time was just writing an email to the University of Washington saying, Hey, can I have some help somebody tell me about plants? And they were like, sure, come on down. You know, let's, let's and so people are generally friendly and curious, and if you engage them in the place that they know about, they will tell you all kinds of things just for the enjoyment of sharing their knowledge. And you should take advantage of that when you can, because you might discover something amazing.

Pablo Srugo (41:19):

And so you've, you know, you now understand how, why the laser's working and it is working, like you said. And so you have a prototype built, which I think was was the objective that, that you mentioned for that, that seed round. So do you go out now and raise a, a larger kind of series A, because you know, I mean, building hardware was expensive.

Paul Mikesell (41:35):

Yeah, yeah, yeah. Went and took a eight point, 8.5, 8.4, I don't know, eight point something million dollars series A and started buildings.

Pablo Srugo (41:43):

And now the objective is kind of use that to build a production level or a handful or some amount of production

Paul Mikesell (41:49):

Level. Yeah, we went through a pretty quick succession, three different generations of laser weeder. And we got to the point where we had done weeding in a, you know, in a real field that was where produce was going to market, but we didn't actually have the machine that we wanted to sell

Pablo Srugo (42:11):

Yet. What was the, the thinking by the way? Like how did, how was this machine supposed to operate? What was the form factor you contemplated?

Paul Mikesell (42:17):

The basis of the whole machine is a computer vision system, cameras, laser control system, and then some clever optics that we've designed. So we have all that system, but then the question is how do you package it and deploy it? So at the time we had autonomous self-driving robots that would do that. And they would drive up and down the field and you can maybe still find some videos of this on the internet. We, I stole a bunch of photos of a bunch of them in the field going around killing weeds with these lasers

Pablo Srugo (42:45):

Robots that you had manufactured within cover robotics or those were kind of outsourced and you put your

Paul Mikesell (42:49):

System No, we built, we built, we built 'em ourselves. And those worked. But there were all these issues around the autonomy piece. There's regulations, there's safety, there's situations where the autonomy just wasn't quite there yet. And so we, and we realized that the, the requirement for autonomy was gonna hold us back if we tied that, if we coupled that together. So what we did was we removed the autonomy, we, we made the laser weeder a thing that would fit, that would be pulled outta the back of a tractor. And then we just focused on laser weeding itself.

Pablo Srugo (43:22):

What year was this that you made that switch?

Paul Mikesell (43:25):

We decided to make the switch in 2021. The first ones came off the line in the middle of 2022. So we made that switch in about a year.

Pablo Srugo (43:32):

And how did you think about the, the go to market side of this? Like as you're building, I mean from 2019 to like 2021, so two and a half, three years you're building, what are you doing on the go to market side during that time?

Paul Mikesell (43:42):

At that point we were just experimenting. We were, we were, we're probably talking to something like, let's say 10 new farmers a month. So we're not really heavily engaged in the go-to-market at this point. We're really just trying to learn more about farming environment. But we were building up a, call it a database of contacts that we could come back to when we felt like we had some product market fit, but we knew we didn't have product market fit yet. And so the idea was, was not to start selling, it was really to figure out how to keep working on the product until we had something that we thought we could sell. And it wasn't until we decided to split out the autonomy from the laser weeding itself that we realized, oh, that was gonna be the, that was gonna be the hit, that was the way to do it.

And that that separation, that split moving laser weeding off the autonomy platform, that was actually John May's idea. That was the first guy I told you about, that he and I pushed that robot around. And so what we did was we built that, that separate pull behind machine. The idea was, and the idea, the idea was see if we could get people to pay us money for it, for what we called pre-orders, which is you pay us some money, you land on the order sheet, and then we're gonna ask you for the remainder of the money when we're ready to deliver. That was the whole concept there. And and people started signing up and giving us money. And even though we weren't gonna, even though we knew we weren't gonna be able to deliver for like a year, at that point, we did demos with the existing autonomous machines, right? So we would demo with the machines we built, show them how laser weeding worked, and then offer to sell them the tractor pull behind version of that. And I did some of those. John May again, the same guy did a bunch of those. And at one point we had stacked up, you know, all these orders and by, so by the time we were ready to roll into manufacturing in 2022, we already knew where our first, you know, 10 laser leaders were gonna go,

Pablo Srugo (45:39):

Yeah. How are they, how are they pricing? How did you figure out pricing? Pricing

Paul Mikesell (45:42):

Is more art and science. You know, you need to make sure that you're selling things for enough that you have enough margin on the product to sustain your organization, right? Otherwise nobody's gonna have anything. Right? But, but you have to price it in a way also where people can afford to buy it, right? So we priced it for an ROI really, we, we focused on making sure that farmers would get their money back within three years on that investment. So the idea is whatever you spend, the savings that you get from that needs to be equivalent to the purchase price within, within three years. So that was our pricing philosophy. These things cost millions of dollars, right? Our, our, the, the amount of money that we spend building it is significant as well. The components aren't cheap. And so it, it's, it's really a question of is the ROI there?

Pablo Srugo (46:30):

And so when you talk about having like 10 machines in pre-order, that means you have like $10 million in, in pre-orders, something

Paul Mikesell (46:36):

Like that. Yeah, that's right. And so that was the, so then we're rolling into 2022, hadn't built, hadn't finished any of these things yet, but we had all these orders. So then it was like, oh, we gotta build this, you know, <laugh>

Pablo Srugo (46:48):

And are you running outta money at this point? 'cause It's been quite a while. I mean, in a million and a half, million dollar A is big, but it's not unlimited. I

Paul Mikesell (46:55):

Mean, in startup life, and particularly hardware startup life, you're al you're always running outta money. You're just like, dude, you're always like three months from bankruptcy. Like continually <laugh>

Pablo Srugo (47:07):

That. Yeah, that's true. <Laugh>,

Paul Mikesell (47:10):

We got good at telling the story. Like we talked about storytelling's important. We had I think a good, a good team with credible backgrounds, you know, people that you could get behind and bet on. And the pre-orders were certainly helpful in fundraising, right? Seeing that people would put money down. So I don't, I don't, I'm not saying fundraising was easy. It wasn't, but it was, it was doable. And, and yeah, I mean we were always, we always had some point where, okay, we're gonna run outta money, right? We always that and that point was always closer than we liked, you know what I mean? But at the same time, you, you know, you see a way through and the way through is continued to sell. So whatever phase you're in, right? Well, for us then, it was continue to get these pre-orders, continue to build the case that does a real business here and then go fundraising.

Pablo Srugo (47:56):

And it seems like the milestones were pretty, pretty kind of clear, right? It was like, raise a seed round to build a prototype, raise a series A to get it close to production, get some pre-orders, raise a series B to really kind of go out and deliver on those pre-orders, you know, and, and everything you need.

Paul Mikesell (48:12):

It sounds that way because I'm, I'm telling you the story with the, with the backbone of, of history clearly in place. Sure. But it's a lot mess here. What I tell you is that when you're living through it, it of course doesn't seem that way, right? You, yeah. You, you, you, you have, you know, 20 different ideas that you think this might play out. And then years later when it works out you go, oh, you know, we, it worked out 'cause we did this one thing. But it's like, okay, but you were also, that one thing was one out of like 20, right? And you didn't know that one thing was gonna be the thing. And and of course that's true. And, and that's been, but that's been true for every company I've ever been a part of, right? Like, even, even, you know, we talked earlier about Isilon, that company that, that I founded that has this multi-billion dollar IPO.

Paul Mikesell (48:55):

And if you hear them tell the story now about Isilon, they'll make it sound like that. Well, what we knew is, you know, A to B, to C to D, and it was totally clear, you know, let's go. That's not what happened at all, man. It was just like, people are trying their best, but the future and path are not clear. And there were, there were definitely moments at Isilon where we were, you know, 30 days from cash out, like, you know, and, and what everybody forgets is, is the times at which it looked like the end, you know, where your VCs are telling you they're not, they're, you know, they don't really know if they wanna put any money in anymore. I mean, that definitely happened at Isilon. There were definitely times where our VCs said, you guys are done.

Pablo Srugo (49:39):

And, and also people will tell, they'll tell, looking back, the one story arc that worked, that ended, that ended with you being all these little tangents along the way just kind of get cut out so you get this simple narrative, right?

Paul Mikesell (49:50):

Yeah. I mean, we could have made any number of simple decisions that could have dramatically affected things one way or another. And we really were just feeling our way through it. And so, and, and you know, like the story's not written yet on carbon robotics, right? We're not a public company yet. We, we seem like we're on the path, you know, we're, we keep doing what we're dealing, we'll probably have an IPO in the next year or two, you know, we, we will, yes, we'll have a billion dollar outcome here that looks more likely every day. And, and what does that mean for us? It means we continue to become a self-sustaining company that continue, continue to build great products based on our experience and backgrounds in ai, AI and robotics. There were several times at this company where we could have made one decision or another and just ended up in the dirt, you know, like ended up dead and, and, and several of these decis, like if we hadn't decided to cut, decouple the autonomy from the laser weeding, if John May hadn't felt strongly enough about that idea to come tell me why that was important, or I had dismissed his advice on that, and we'd probably be a dead company right now.

Paul Mikesell (50:57):

You know, we'd probably be at least struggling, right? And so it's just like, it seems clear now because it's worked out, but it was not clear at the

Pablo Srugo (51:07):

Time. And instead, how many, how many machines are now out in the wild destroying weeds?

Paul Mikesell (51:13):

We will ship number 100 in, what is it now? July, so, no, in the next probably six weeks number 100 will go out the door. Yeah. And you know, they're, they're, they're, they're a million and a half dollars. So we, you know, we're, we'll cross 150 million in, in, in revenue in, in not two thou not too long.

Pablo Srugo (51:33):

That's great. And, and by the way, going back to that, to that time, like those 10 that you shipped, just, just back to your messiness point, like how messy was that, or how well did those work out in, in the field?

Paul Mikesell (51:44):

So we, we brought on a partner, a contract manufacturer in Detroit. And I will tell you that that was not clean at all. That was super messy. Nothing was right, everything was wrong. It was way understaffed. All the parts were built wrong, nothing fit together. The component manufacturers that we trusted completely failed us. Like the people that we had this one group that was supposed to build us a chiller and an AC unit that we depended on. And basically none of those worked. They were all just total garbage, you know, it, it, it was a disaster. But we, so, but we fought through it. You know, we spent, we spent months and months and months engineers, you know, basically living out on the manufacturing floor you know, doing these whatever 18 hour days, like figuring out what was working, what wasn't working, tearing stuff apart, rebuilding things, you know, continually.

Paul Mikesell (52:40):

It was really painful. But we got 'em, you know, we got 'em working, we got 'em shipped. We refined our process. We learned how to get better and better about it. Some fun times, you know, from those days we, at one point everybody was so frustrated, we had this, this chiller thing that just didn't work. It never worked. You know, this company told us they knew and they were experts and whatever. We should have tested it better, I'm sure. And you know, you can say that, you always say that about the stuff that didn't work. I should have tested it better <laugh>, but you ignore all the other things that did work that you probably could have spent time testing as well or whatever. But we still have this video. Like we took this chiller and just everybody out there on the floor, we all just, I don't know who started it. We all just grabbed these like giant pieces of tubing pipes and just like beat the out of this thing for like, for a while. But it was, it was the, it was not the things you expect, you know, it wasn't the optics, it wasn't the servers, it wasn't the computers and the gp, it was like the chiller and the AC line and the high voltage routing of the cables and stuff. That was the crimp, the crimp connectors in between, you know, different harnesses, right? It's all that kind of stuff.

Pablo Srugo (53:50):

Yeah. For us it was many times it was Bluetooth, like just Bluetooth dropping data packets and just screwing this over. Like, you serious <laugh>.

Paul Mikesell (53:57):

Yeah. It's like Exactly. Yeah, exactly. And it's like, you're like, how are we supposed to pretend like this technology is working? You know, like, like Bluetooth is a great example.

Pablo Srugo (54:08):

I think. I I always say I think it's the worst like ad scale technology that we've all chosen to depend on this, just no, never actually got there. It's crazy.

Paul Mikesell (54:16):

And it's gotten quite a bit better. I think people forget how bad it was. In fact, Bluetooth is an example, right? Your headphones you're using right now are Bluetooth headphones. But if you had a generation of those maybe six years ago, you would've, they would've been out of battery by now. They wouldn't have worked. They would've dropped out several times during this call, right? And so that is a, that's a perfect example of technology that looks reasonably stable and usable now, but really had to, <laugh> had to evolve its way to get there. Yeah.

Pablo Srugo (54:41):

Cool. Well listen, we'll, we'll stop it there. I just have the two final questions that we always end on. The first one is, and I'm really curious on the answer here because of, you know, how, how you had to build what you had to build, which is when did you feel like you had true product market fit?

Paul Mikesell (54:56):

It was when we first started selling those machines, before we had built anything at all. 'cause We had done the math to know we could build them for a, a certain price and we had done the math to know if we could sell them for the, for this other price that we could have a real business. And when we started selling those things based on concept alone, that was when we knew we had, we hadn't even finished building the unit yet. It was, it was when we were able to sell for millions of dollars these machines based on concept and design that

Pablo Srugo (55:25):

Will do it. And the last question is, if you could go back to the beginning of starting Carbon Robotics or, or even before then, or maybe even in general with kind of one of the most important pieces of advice you'd have for yourself or for founders what, what might that be?

Paul Mikesell (55:41):

Treating the people that have worked so hard to make this happen? Make sure that we treat them well and make sure that they know how much we appreciate them is the most, most important thing for me. It's always been more important the way we make decisions than the actual decisions that we make. So what that means is I would rather have an organization where all of the people who knew what they were talking about, were able to come together and have a voice and make the decision even if it's wrong, right. Than a company where maybe you make, you get to the right answer, but everybody hates working there, right? And so, and, and so why is that important? Because I'm talking about scalability. Now. If you have the kind of culture and environment where good people can use their expertise and skillset to make good decisions, then that can scale. If you have the kind of environment where it's sort of a dictator setting a direction that's not scalable. So I always try to lean on this concept of scalable decision making that doesn't, and on, and I'll tell you, that doesn't mean everybody has a voice. It means the experts who know what they're talking about get the authority to make those kind of decisions. And that's incredibly important to me. And it's more important than the company or anything else. Perfect.

Pablo Srugo (57:08):

Well, thanks a lot, Paul. Really appreciate you jumping on the show. This has been great.

Paul Mikesell (57:12):

Yeah, you bet. Look forward to seeing it. Hopefully everybody hears is learn something about what they're trying to do or about our company or feels like they at least have a good time hearing it.

Pablo Srugo (57:24):

If you've listened to this episode and the show and you like it, I have a huge favor to ask for you. Well, it's actually a really small favor, but it has huge impact. But whichever app you're listening to this episode on, take It Out. Go to a product market fit show and leave a review, please. It's going to help. It's not just gonna help me to be clear. It's going to help other founders discover this show because the algorithms, whether it's Spotify, whether it's Apple, whether it's any other podcast player, one of the big things they look at is frequency of reviews. It's quantity of reviews. And the reality is, if all of you listening right now, left reviews, we would have thousands of reviews. So please take literally a minute, even if you're just writing like great podcast or I love this podcast, whatever it is, just write a few words. Obviously the longer the better, the more detailed the better. But write anything, leave five stars and you'll be helping me. But most importantly, many other founders just like you, discover the show. Thank you.