How Robotics Could Transform Construction

0:00On this episode of the show, I have with me Shaman Tadiki. He is CEO and founder at Human Family Robotics. We're talking about robotics and construction. We're going to talk about the application of how robotics can impact the construction process. We're going to talk about where, talk about how those projects look a bit different and how the products maybe actually built a little bit differently. And Shaman, I appreciate your time. Thank you. Hey, thanks so much for having me. Big fan of the show and looking forward to to a new conversation. Thank you. Appreciate that. Okay. Human friendly robotics. What do you guys do there? Human friendly robotics. We make robots that do construction. The long-term vision, the long-term plan is, and it's not even that long-term. Within the next four to five years, we anticipate robots will be on the majority of construction sites in some form or the other. So we are aiming to be in those construction sites in that four to five-year timeline. To start, we built a tiling robot, which we had to name Tyler, just seemed appropriate,

1:02but that's our wedge into the marketplace, mostly so I could have the stupid pun of VR at the ground floor. I love it, that's fantastic. Not going to not forget the name of your robot, so it's just going to be stuck in my head, Tyler. I guess, you know, we'll get into a little bit more, but you're looking at the construction industry. It's obviously an interesting industry. There's no lack of work that gets done in it. When you look at the construction industry, why you chose, well, what made that stand out as to a field to apply robotics to for you? Okay, so let's look at some numbers. Manufacturing over the past six years has had a seven X gain in productivity per worker, per person, agriculture, something like 15X, retail, logistics, everything. It's multipliers of productivity compared to 50, 60 years ago. Construction, the after productivity per worker, has declined 50% since 1960, which means if your grandfather

2:03was in construction, he was twice as productive as you are today, or not the same field, and he had worse tools. So it's right there, and then now is the time. There is a massive, massive shortage of people willing to do the work in the US today, where 500,000 worker short of where we need to be, and the youth are not filling in at replacement levels, anywhere close to what they need to. The grandfather was a plumber, the father was a plumber, the son, a tick talker. You know, it's interesting, because the more we're hearing about AI, the more people are like, well, I'm going to go into HVAC, or I'm going to go into plumbing, I'm going to go into the trades. And here we are talking about a product that's actually going to take away some of those jobs from a robotics perspective, because it's not just AI threat to jobs. I mean, there's actually jobs that are currently being automated by robotics.

3:04And I guess when you think of people not wanting to go into construction, I'm still not wanting to believe that people are going to give up certain types of jobs for, you know, going out link brick. I don't know. That's a, that's a, it's a different mindset. And I think I spoke to someone who is kind of working with robotics and farming. And he said, there's a similar problem. There's very few people are going into farming, regardless of if there's job availability. But I guess when it looked, you look at your product and maybe we can get into a little bit of it. So you mentioned obviously some, some use cases. But, but when you get a product like this, which is very much, you know, replacing manual labor, where does the human sit? Is their human loop use case for situations like this or does the human have no role or is there still supervision? Is there an out that people are using? A little bit of time travel, if you will. For most of human civilization, we've been moving physical things in the world somehow, you know,

4:07whether it's physically with our own labor, or we mechanized our primary function in the world was to move things from A to B in some way or the other. Sometime around called the 1800s, we have this industrial revolution, steam power, electricity, all that, suddenly things can move without us actually moving them. Fantastic. Fast forward, another 150 years, 1950s, 60s, 70s. We suddenly started thinking in the abstract realm that our minds were built to think in. Our gift over the animals, if you want to call it that, is our ability to reason abstractly. And the digital age was built explicitly, not just by us, but for us, because everything in the digital age is abstract information. Entire jobs, careers, fields, industries were formed, just shuffling bits. If you think about it, that's kind of all they were, right? The mega mammoth companies of the 1970s, 80s, 90s, 2000s, 2010s were just shuffling bits from A to B. We're now at a point where,

5:14let's call it AI or LLMs or whatever you want to call it, is fundamentally better at moving bits than we are. So, the human and the loop problem isn't that we won't have humans moving things, because again, with the industrial revolution and steam power, electricity, we already took away a lot of that, right? When we were doing plows by hand, then we got the donkey to do it. We never went backwards, instead of, I'm going to do a better. We just accepted the donkey's better and the steam engines better and the gas engines better than that and on and on. So, I think it's the wrong framing to be concerned about the human and the loop for atom moving, when it's happening and happening for bit moving, if that whole long drawn out analogy makes any sense. Yeah, it does, because I think a lot of times we look at history, obviously, differently. You have a great example of the donkey carrying stuff versus I'm just going to go ahead and try to

6:15figure out how to move similar loads. I mean, that completely is a great example. But when you look at the industry and you look at the sheer size of construction, I mean, it's so big, it's so very, there's so many different types of equipment. I mean, obviously, you mentioned the tiling, tiling of the tiler. When you look at what you can maybe bite off, what you can actually start with that moment, I mean, obviously, tiling makes sense. I mean, conceptually, now that we're talking about it, okay, well, you know, there's a predefined area of work to be done. Tiling seems to be part art and science. When you see people do it, I watch people tiled. I'm like, dang, that seems really hard. When it comes to automating that, I guess, not a human versus machine discussion, but who wins out? I mean, is there our component to some of this stuff? Or is it, no, Mayor, we need to lay the tiled a specific way. The machine never makes a mistake. And it is what it is. I mean, how do they, how do the two compare when you see the results in the room? I'm sure I'm going to

7:18get a lot of hate, but the a tile setter in our one of the job is a different person than a tile setter at our eight of the job, because we're humans, we fatigue, muscles, brain, all that, right? It's a different game. Machines, robots, whatever you want to call them are just as effective minute one as they are, minute 5,000 as they open them with maintenance, of course. So if we can be on par with the human, forget about better, the first minute, we can be better than the human by the end of the day. That's a fair point. It's actually true. I mean, a machine does not get tired. I guess when you're looking at machines in these in the scope of these type projects, I mean, they don't get tired. I mean, you mentioned, obviously, have to manage them. When you think about how you're looking to deploy robotics into the field, what does that model look like? Are

8:20people buying the robots? Are they, is it your project? How did they actually get to the robot? And how do they actually put it to work? So great question. So we actually have our first commercial deployment coming up later this month at a luxury helicopter in Puerto Rico. The model is rental or, well, consider this, construction companies don't like apex, but they completely understand apex, right? And not many contractors will go out and buy a heavy machinery for a job. They'll go ahead and rent, you know, I need a digger today, excavator, whatever, because when you get to a certain scale, it makes sense to buy, but most aren't at that scale. And the same is true of Tyler. If you're, and actually we're talking to a company right now, they've got 225,000 square feet of quartz tile coming up this year alone. So for them, buying probably makes more sense than renting, but they're few and far between. So our goal and our go-to-market is very rental or per job or per whatever the unit makes sense for that particular

9:22contractor, which gives us an advantage in that we get to continue to improve the machine. Right? If you buy it as is, version one is not going to be as good as version five, but if we continue to rent it from us, we can make sure you get a version five when it's ready. Interesting. And I guess when you're looking at, you know, various areas and construction, are there certain, I mean, I would call it jobs or skills, is there certain skills that you look at and you realize they might just be easier for, I mean, obviously, Tiling has a pre-distanble talk about that maybe, but when you look at the various, because there's a lot, I mean, drywalling, I mean, I don't know all of them, but I want to pretend I know. But drywalling's one that I know, I mean, we talk about plumbing, we talk about it. Is the goal to start looking at how do you get into all those areas or is it really like, hey, there are truly certain aspects of construction that lend themselves well and let's just focus on them because, as you said, I mean, there's so much

10:22construction that does happen. So you have, you know, the startup model of the north star and the wedge, right? So our wedge into the market is for Tiling, because like you said, it's repeatable, it's known, it's fixed, machines have been doing repetitive motion now since the beginning of machines. But where I want to go is the, the Trinity, you will have licensed creates electrical, plumbing, HVAC. There are no two jobs at the same. So as we continue to deploy in the, call it the more repeatable spectrum, we gain data about the real world and we use that data to get to the less repeatable part of the spectrum. So on the incredibly repeatable part, you have things like Tiling, painting, maybe roofing, bricklaying, and on the very hard to repeat, again, the Trinity, electrical, plumbing, HVAC. And it's no wonder those are three licensed trades. You have to be a licensed plumber, a licensed electrician, right? The internet's, they have to be a licensed painter. Absolutely interesting. And I guess as you're kind of, you know, if somebody was thinking about a

11:26Tyler and we start thinking about, you know, how that gets put into place. Obviously, we are seeing robots being built for different things, right? And they are robots right now. I mean, all these big factories in the Amazon, so Walmart, they all have robots moving stuff around. They do very distinct things. I guess when it comes to construction, I mean, it is a, it is not, it is not as safe. I mean, people get injured on the job. I mean, obviously we won't talk about that there. But when it comes to the actual, you know, the margins of safety, you incorporate in a product like this. I guess, how do you view that given things can go wrong and given the fact that, you know, if we're taking the human slightly out of the loop, this thing needs to work pretty autonomously and safely as well. Let's talk about injury for a second. There is, let's just broadly say, two types of injury. There's the traumatic, something hits something, someone gets hurt in some horrible way. And that's obvious and clear. You know, a thing fell on

12:26someone, a thing went into someone, and that's unfortunate and should be avoided. But then there's the hidden injury, repetitive stress injury, being on your hands and use all day, lifting heavy things all day, being in potentially dangerous conditions all day. The injuries that accumulate over time, construction as an industry is four times, has four times the number of fatalities, then other industries, most other industries, not kind of things like defense and so on. But the 50 percent decline has not been met with a proportional decline in safety. Sure, things are better. We've got OSHA, we've got regulations to protect workers, but it still is a very dangerous field. So if we're worrying about injury, we have to look at the two classes. And I don't have the numbers in front of me, but I think the second class, the repetitive strain, the more thinning injury, is much larger than the traumatic class. The traumatic class is protected by worker regulations where hard hats, the autobutes, these things help that traumatic class. But not

13:30much is happening for the hidden class, the latent class of injury. And that's where I think call it automation, call it robots, call it co-bots, will help a ton. Because the skilled pile center is still needed. Even in a world where we've got Tyler fully deployed and it's doing its thing and it's fantastic and super fast, there's always going to be some edge case that we just can't build for or haven't built for. And today that edge case is literally the edge. Tyler doesn't do the edge of the room. That's still a skilled person with a saw, with precise management skills that understands the geometry of everything doing that. I mean, that makes sense. I mean, obviously there's the certain, when you look at that, I mean, I'm sure you guys have looked at efficiency gains and whatnot. Like what, what does efficiency look like when you put a robot to tile versus a human in an hour or maybe it's tiles laid in hour? I don't know what the metric is, but I'm kind of curious if there are metrics you've seen. We're upbeat and it's very difficult metric is if you're laying six inch by six inch tiles,

14:31obviously, it'll take you much longer to cover the same square footage versus, you know, two by two foot tile. But generally speaking, Tyler is at least three times more productive than a human in a particular day. Our goal is five times. It's really hard to quantify because it really depends on the space. For example, in this job we're coming up from Puerto Rico, it's a windy outdoor patio, which is not great for Tyler because it's not a big rectangle. That's our ideal. That's our dream. So in a big rectangle, we're easily five times faster and maybe less of a rectangle, closer to the worst case too. That's actually pretty incredible because and I guess, you know, with time, right now you mentioned and you mentioned going up to the Holy Trinity of construction. I mean, as you are getting into learning about Tyler, it's going to get a Puerto Rico. It's going to do a job. Are there, and I don't know this, are there industry

15:35regulations? Are there industry parameters that are coming into looking at robotics? Because obviously there's heavy machinery. We know that and I'm sure there's various things in that side of the world, which I don't know enough about. When it comes to robotics, like as the industry started thinking about that, is are there guidelines? Are there, are there bodies that look at these robots to prove them? How does that work in the world? So I might be misquoting. So, in fact, if I'm wrong, but I believe in Europe, they have more standards for what a co-block panic cannot do, parameters that it cannot cannot operate in distances, it must be. I know that in, for example, some Amazon warehouses, if there are robots that are the honest or in speed, they are attempt away from humans, just in case there's an accident. It's not with a person. When it comes to construction robots, I think, and I may be wrong. It's still being discussed, sorted out, thought about. I don't think, as far as I know, there is anything really

16:39coming down of what a robot can and cannot do, shouldn't, shouldn't do, at what speeds, parameters, and specs for certain jobs. When it comes to tile setting, there's the NTCA, which is the National and Tile Council of America. They specify the parameters for how big and gaps and sizes of tiles and all the tolerances and what isn't as acceptable. But as far as I know, there's nothing about robots. Because the pitch to contractors isn't fainting cool, tiling robot. Here's another tool. How do you, somebody mentioned once, you know, building a couple of one thing is one thing and then building 100,000 of one thing's another. I guess, how do you envision, obviously, you got the project and you guys are young company. You got a product. You're going to go lay tile down and Puerto Rico. That's going to be a great use case as well. How do you envision going from handful to thousands and thousands? Because I think that, that introduces, you said, most people will just use them, lease them, rent them, and pay for them, not own them in a lot of cases, which makes

17:44a lot of sense. But obviously, from a supply chain perspective, that introduces some complications from being a robotics company to now having to worry about supply chain distribution, getting these robots there. I don't know if you need to service them or you've got to maintain them, all those things. We've all seen it in Star Wars and similar, the robot base. I'm now envisioning some kind of future scientific sci-fi where you're going to have bays and bays and these robots being looked after. But just out of curiosity, as you scale, obviously, you'll come to that challenge, you'll solve it. But what are you envisioning? How do you see that problem working out? So let's assume that we never expand to any other trades. We only stick with Tyler, and that's all we do forever, and we're the tiling robot company ever. If we were to rent out robots every single day, 365, to do a quarter billion in revenue, we only need to rent out for different robots a day. So you're at the point where we're dealing with millions and millions and millions of units.

18:47We're fundamentally different company, which is not necessarily a bad thing, but there's a problem to be solved when we get closer to being that company. As of today, again, assuming we never break through wood or walls or painting or the other trades, we rent out 1,500 units a day total. If we're renting out 10% of our total fleet, we only need 15,000 units. If renting out 1% of our total fleet, we only need 150,000 units. It's a manageable number. I'm not going to say easy, because of course there's logistics and how do you maintain it and where all these things kept in? Is the droid factory going to be a part of the thing? Or do we have local partners that are trained in our equipment? Just like if you go to a shopping equipment rental place, they rent out caterpillar equipment. Does caterpillar service it? Or does that equipment rental place service it? I've seen both models. Yeah, definitely. Yeah, and I was interesting, because I was thinking about all these image and painting time. Obviously, I'm going to be able to do these autonomous

19:55tasks and obviously, that's a great problem. You don't need that many robots into the market, which is pretty amazing. You're building robots, you're putting them out in the field. If people are listening, they're like, hey, I see all these videos for robots. They're doing all these other things. There's a lot of different tech that goes into a robot. When you're looking at your robot and if people are comparing it to what they see, you necessarily don't need to have the robot that dances and does the backflips and does those things. Your robot's doing specific tasks. I mean, obviously, you can build it to do things. How does that? I'm assuming it's going to make it easier to build, and then also hiring this feels really hard. How are the two work together? So we've all watched Boston Dynamics videos for 30 years of robots doing backflips. And if I ever need a backflip done by a robot, I know to call. But if you solve more repeatable problems,

21:02everything is, I don't want to say easier, but more certain. So when if you're picking up a tile and placing it and navigating a space, I mean, your Rumba has been navigating your interior space now for 30 years, and it's gotten better and better at it. When it comes to picking up things and putting them down, warehouses have had bone grippers and different types of grippers now for 40, 50 years. So I'm trying to lean on Trident True versus the next cool TikTok video thing. Now, not to take away from that because the people pushing those boundaries of what is physically possible do open up frontiers of research and eventually commercialization for the rest of us. Interesting. Yeah, I like it. I mean, obviously, you mentioned trying not to solve problems shouldn't solve. And I think you gave us your wedge and your holy trinity and staying focused on that path, I think, is very challenging sometimes because obviously, if you go, hey,

22:08if I just stretch a little bit more, I can do X. And that little bit more stretch, typically, comes at a heavy cost. I guess, as you're executing, how do you help the team? Because I'm sure you guys are building, you guys are innovating. How do you work with the team? Listen, guys, yes, we might be able to stretch this next couple of percent into this amazing other breakthrough, but for what we're building, that's not necessary. We actually have a sign that I put up in the office here. And the sign simply says, Pareto was right. And Pareto being the Pareto principle, the whole 80-20 thing, whatever new, cool, weird edge casey thing you're thinking about, we'll suck up 80% of our time no matter how small it is. So let's not worry about that. For example, edging, could we solve it today? Probably. We could probably spend a few months figuring out how to cut tiles and do the whole thing. But of a typical floor space, that's less than 5% over the floor. So I might really dedicate the next year to solving the cutting tile problem,

23:12or am I going to worry about 95% of the floor that doesn't need to be cut. So it's coming from software as well, right? That's been two decades doing software and startups. The principle of yagni, you ain't going to need it. Whatever it's you're trying to scratch, you probably don't need to. And you know, your channel, jobs famous, quips and quotes. And you start to realize, you might have been on to something. Learning when to say no is more valuable than just going along. I love it. That's a great stopping point. So I appreciate your time, man. This is very interesting. I'm excited. We'll have to have you on down the road to see how a Tyler got on down in Puerto Rico and in the learnings. But if somebody does want to ask a follow-up question, because I'm sure people are going to hear about this and be like, okay, well, this is not the typical robotics episode, but it sounds like it's actually going to get out in the

24:16field really quickly and do stuff. How can somebody get in touch? How can somebody ask a follow-up? So you can fax me. Oh wait, no, that's not going to work. How about LinkedIn? You can find me on LinkedIn. Probably the only Shimon Sadiqi out there. And the picture as a guy looks kind of like this, but probably a little bit younger. There you go. I love it. Sherman, thanks for coming on. Thanks for sharing with us. Hey, have fun, man. Thanks for having me. Absolutely. All right, that's it for this episode. Be back again, different guests, different topic, until then, two things. One, if you could share this episode with somebody who has an interest in robotics or is kind of curious by how a robot's actually going to get on in the construction field. This is the episode for that. So please share this episode. I think Shimon did a great job of talking to us about what Tyler's going to be doing and his vision and where he thinks the opportunity is out there. And he's probably going to have a whole fleet of robots at one point and we'll have to figure out how he's building that out and getting to other areas down the road. It'll be a great follow-up. So please share this with others. Also, like, subscribe, comment. Let me know how the show's going for you.

25:20Until next time, thank you.