The Adam Sand Report - Thursday, August 27, 2026

What's going on everybody? So we are up for an interesting live stream today. So for those of you who might be new here, my name is Adam Sand. I own roofingbusinesspartner.com, business that helps other roofing companies grow. And today we're talking about robots on roofs. We're going to make a case for why it's possible. And the reason that I want to talk about it is because that's what I'm doing. And it's not, it's not something that, uh, like I'm not putting a robot on a roof today. There's a much longer tale to how that actually works. I'm going to walk you through why I think what we're doing is possible, and what the evolution of putting a robot on a roof looks like. And so if you want to see what the future of the roofing industry might look like, you might want to stick around because I'm going to make a case. I'm going to show you guys how we're getting there, how— what— how you can know that it'll be true, and what little pieces of information might help you feel a little bit more okay and comfortable with it, and why we're not like— why the world won't start with robots on roofs. But But I'm going to go through the process of actually getting robots to be capable of roofing and how the technology to do that exists today so that the robots of tomorrow will be able to install a roof. And so I'm going to do that by making a number of demonstrations. I'm going to show you some videos and some stuff that's going to kind of help you guys understand these concepts. And I'm going to try and save you guys what is, you know, a few hours of YouTube. And I'm going to kind of compress it down into some facts that I think you need to know. That'll kind of help you skip the line. Um, I do these live streams at 9 o'clock at night because I fully expect them to suck. I don't, uh, I, I don't expect to have thousands of viewers. I do this because I want it more or less for the historical record, right? Um, it's about documenting the journey more so than actually trying to say, all right, I need to get a bunch of people to watch these videos. It's that in the future I want to be able to say, you know, to people who also had dreams that, you know, there was a time when these live streams got 4 views, 16 views, 112 views. And back then, you know, I still did it even though it was hard, because everybody out there has a dream. Everybody else wants to accomplish that big thing. And what I don't love is that I only get to see my heroes after they've accomplished it, right? You don't see Jeff Bezos, you know, from the early days of Amazon. where he was on his hands and knees putting books in boxes. There's a funny story that, you know, in the early days of Amazon, he's sitting there putting the books in the boxes, putting the books in the boxes, and he stands up and he stretches his back and he looks around to order— now like several hundred millionaires— he looks around, he says, oh guys, we need, we need knee pads, this is brutal. And one of those guys in the original garage with him goes, no, Jeff, we need desks. And that would have been pretty funny, you know, that you were so zoomed in on the problem that you didn't even think like, oh man, we don't even have any desks. And how cool would it have been to be able to go back and see that Jeff Bezos, if he was able to like rip open his phone and be like, man, you wouldn't believe it, the craziest thing just happened. And he would tell you that story. And you don't get that from a Bill Gates or Jeff Bezos or a Mark Zuckerberg or Alex Hormozi or any of those guys. And I— and all of them say that they wish they documented the journey back before anybody thought they were cool before they, you know, before they had the fame. And that's what these live streams are about. So we're showing you what I'm working on before it necessarily hits the mainstream. And robots on roofs, I think, are something that's, you know, kind of worth talking about. And so we're going to get into it, and I hope you guys enjoy the journey. So I want to start with something called Moore's Law, because we taught rocks to think. And so we're gonna— if we're gonna talk about robots being on roofs, we have to think about where technology goes. And what this is, this is something that was posted a long time ago saying that the number of integrated circuits, essentially transistors on a computer board, would double approximately every 2 years. And that has held relatively true. If you draw a line through all these dots, you will see that from 1970 to 2020, the number of chips that print transistors has effectively doubled every 2 years. And there's been periods where that slowed, and we're actually in a period Or that's slowing down. But I find a lot of people don't really understand the size and scale of what that means, because there was a new Mac Mini released, and it's an Apple M6. And all that means is a 2-nanometer chip. And nobody really has the concept of, like, what does a 2-nanometer chip mean? And I want to talk to you a little bit about how cool it is that we've essentially taught rocks to think. And so I'm going to take you to a part in this video where it kind of explains that. This is a video from Marcus Brownlee. Okay? And he essentially snaps his fingers and goes down 100 times in size to actually take us inside what is a modern-day computer chip. So he goes down to 1/100th of his size and shows how big he is, right? You can see he's standing here on the keyboard. And then he opens up that, that keyboard, and you can see where he's standing on this on this chip, right? And what this looks like, and that there was the M5 chip, which is a 3-nanometer chip. So the new one is a 2-nanometer chip, came out less than a year later. 2 nanometers is what, 30% less big than 3, right? And so he goes and he snaps himself down even further. Yeah. Okay. Okay.

Wow. Welcome to the scale of Micrometers. I am now 200 micrometers tall, which makes me about twice as tall as this human hair is thick. This is already the size of some of the tiniest stuff we're able to see using our own eyes. But yeah, a typical sheet of paper is about 100 micrometers thick. Looks pretty tall to me.

So this might sound small, but it gets crazier, right? Because what he's doing is he's taking you down journey to become even smaller, right? And so he snaps his fingers again.

To me now is a transistor from 1968, 20 years later. Moore's Law says that that's enough time for the number—

So that's a— that's in 1968. Now remember, this chart that I showed you starts in 1970. This is 1968, the size of that chip. Now, if we go back here into this video, Waze, okay, when we talk about the size of a transistor, he talks about it here. I'll show you guys where it is. Where is it? He showed right here, okay, right here in the video.

And the M5 chip should be right over here powering the whole thing. Now, this is the size of a single transistor in 1948.

So look at that, 1948, a transistor is this size. Now what does this have to do with robots on roofs? I promise I'm going to get there. You got to stick with me for a second because if you're a roofer or a roofing business owner and I'm talking about robots on roofs, you know, you got to kind of follow a path of some stuff that maybe you haven't been paying attention to. So that's transistor 1948. Where he's talking about here is 1968. Okay, so that's a transistor, 1968. And here, whoop, this here, Moore's Law is 1970. So think they're getting smaller yet, right? Now what does all this have to do with robots on roofs? Well, we're gonna skip ahead a bit. So he goes down another 100 times. So that's that human hair.

100x shrink. So now it's getting ridiculous.

I am now So that there next to him is the bacteria. So he goes in and he talks a little bit more about how, how small he is, but then he snaps down another 100x. Okay.

2026 M5 MacBook Pro is about the size of a garage for me. Now I'll note that this most likely isn't 100% accurate because Apple's not going to share their exact chip architecture with me. Making a YouTube video, but still the smallest parts of it are still smaller than me.

So think about it, the transistor in the, in the Mac M5, when he snaps down 100 times smaller. So think about where he was standing next to that bacteria, what he's standing on.

And at this scale, you can see individual atoms, which are the size of marbles.

So think about that for a second. We are now today human beings manufacturing something at scale, meaning we're putting thousands of these phones, okay, into people's hands— or not the thousands, sorry, millions, millions of phones, right, like this size, in people's hands in 2025. And the smallest pieces inside of it are the size of atoms, and they're the size of marbles, and we're manufacturing things that are the size of a bicycle or a desk. Okay. Now, obviously, because that's a transistor right there, remember how big it was before. And so he makes a point to bring you back to the original size. And if you're in the chat, Whether you're watching it live or later, right?

But if we still made—

If you were to guess, if you'd have had to make an iPhone out of the size of the transistors that were made in 1948, how big do you think it would be? So if you were making an iPhone today out of the, out of the, out of the transistors that were made in 1948, how big do you think the phone would be? Put your answer in the comments so you have a chance. So again, if they made an iPhone today with the transistors that were made in 1948, for the sake of your own brain exercise, how big do you think the phone would be? Some people guessed— I talked to everybody about this earlier— some people guessed the size of a bus, right? The size of a gymnasium. Well, here's the crazy fact.

18 billion transistors that it has today. Think about it. Maybe the size of A city block? Or what about the size of the entirety of Central Park? Keep in mind, you need to fit all these billions of transistors. Wouldn't it be crazy if we carry around the equivalent technology of a phone the size of the city of Manhattan? Well, the reality is these are all still too small. Your phone with the technology from 80 years ago would need to be the size of the state of New Jersey. To fit them all. This progression of technology has given us a computer the size of a U.S. state 270 miles long and shrunk it down to fit in our pockets.

So we have built this thing to this level, this technology to this level. So then the next thing you need to do is you need to understand how the heck are we doing this? And again, I'm going to get to robots on roofs, but how the heck are we doing this? Well, we're doing it with a process called photolithography. Okay, and so it talks about how microchips are made in this video, right, which is called The World's Most Important Machine. If you want to know why the whole America defending Taiwan, China wants Taiwan, you want to know where all that comes from, there's a company called TSMC that makes all the microchips, and they have the They have the process of making all this stuff, and one of the most important machines in the world is this photolithography machine. Okay, and it's done— there's a very interesting story near the beginning of this that, uh, that talks about how these things are made. And essentially, there's a mirror that's made in this machine. So here's the mirror. This mirror, if it was the size of Germany, Okay, the imperfections in this mirror, like, so as you think about the surface imperfections in this mirror, would be the size of a— would be a millimeter. Now I know if you're American you work in freedom units, and that's like a 16th of an inch, okay? But that's how perfect this machine has to be. And so what it does is it's working like a kind of like a welder.

Fill in those etched lines. As a last step, you wash away the remaining photoresist And now you've made a single layer of the chip. We've simplified this cycle down to the main steps: coat, expose, etch, and deposit. It repeats for every single chip layer, and depending on the chip, there could be anywhere from 10 to 100 layers. The bottom layer is the transistors. This is the most complicated layer, requiring hundreds of steps that all need to be perfect.

So think of that— you're using a mirror to reflect light to laser etch into sand at a scale that is atomic, right? Where the parts that make up this microchip, right, that we're using today are on the scale of atoms, right? And this whole video goes through the whole process of how they make this photolithography machine. I highly recommend watching it. I'll make sure that the The videos make it— the links make it into the, into the comments below. So if you want to watch this, it's actually pretty fascinating because this, this machine is incredible. But what it is, is it's what allowed us to build transistors that get smaller and smaller and smaller. Like, as I showed at the beginning in Moore's Law, every 2 years the number of these connected integrated circuits doubles, right? Essentially, the size of these transistors gets even smaller. And keep in mind, when I showed in this video, right, when I showed him standing next to that, standing next to that thing, and he said that on a 5-nanometer chip, he would— the chip was roughly the size—

Most likely isn't 100% accurate because— and at this—

So he said these are about the size of a garage, right, is what he said back here. Like, that's the individual transistor. 14 billion of them, right? That thing you see in the background here is an E. coli, okay? Bacteria. These dots you're seeing here, this is the size of atoms, right? And we're making 14.7 billion of those inside your iPhone, and they have to work. We've manufactured something that works on the scale of atoms, and it just got 33% smaller less than a year ago. Later. So what does this have to do with robots on roofs? Right? Why does— what does this Mac M6 Mini have to do with anything? Well, the concern that everybody has when I talk about robots on roofs is they say, Adam, that's nowhere near possible. It's nowhere near possible. Well, I would hope that by now, anybody who's watched this video, what is it, 16 minutes in, that you've now realized that we as human beings have found a way on a scale of selling millions and millions and millions of phones and computers and laptops, iPads, multiple brands, have figured out how to like manufacture rocks that think where we, where we make 14.7 billion individual parts in the size of a thing no bigger than, you know, Smaller than this stack of Post-it notes. And inside this space, 14.7 billion individual things that work— transistors that go left, right, left, right, on, off, on, off, 1, 0, 1, 0— 14.7 billion times. And they work, and we sell millions of them. And you hold it in your hand, and if it was built at the same size of 1948, then it would be the size of the state of New Jersey. And you have that computer in your pocket right now. You might be watching this video right now. So I would hope to think that now, 17 minutes in, that, yeah, we've gotten pretty good at making stuff. Okay? And so the thing that has to happen in order to train robots to do any job is you have to have data collection. All right? You have to collect data. And so right now, there's a cleaning company in San Francisco that sends robots. To your house, okay? And this company right now will come clean— Tao Robotics will come clean your house, okay, for $30, okay? They will come clean your house for $30, and these robots will get out and actually clean your home for you, all right? And what's happening is they're doing this because they're trying to collect The 6 million hours required to train the robot. So these ones are operated by human beings. So there's human beings behind the robot running the robot. They're, they're, they're running it just like from a remote control. But the idea is, is that they're trying to do it. They're not the only people in the game. Okay, there's a company that will pay you to wear a camera and do chores. Okay. Adjust your gear. This is it. Calibrating hands. My new job. Recording started.

Training robots to do household tasks.

So right now, this company will ship you this ridiculous headgear. And if you are willing to wear this headgear, as long as you are doing this household chore, it is tracking your hand movements as you work.

Analyzes the hand movements.

Finally, all the accepted video is converted into code that can be used. So this company will pay your wife, okay, to wear this goofy headgear that she can take her phone and it'll pay her $20 an hour to wear this thing, and as long as it sees your hands— let's say 72% of the hour it could see your hands doing the work— you get paid 72% of that $20 an hour to track your hand movements. All right, and this isn't the only thing. Okay, if I look up this, there's, um, post of people in India wearing cameras on their head to train robots. So 3rd world countries are getting in on it too. Okay, people are getting paid to wear these things. You look at— there's this post that went viral on X right here. You might have seen it of this guy. They're wearing this, operating these sewing machines, training AI with these head-mounted cameras. Now you might think, oh, fine, this isn't coming my way. You'd be wrong because These companies I just showed you have just been wiped out, and they just got started because Figure has now deployed this, and they're already out there in the world. They've already had 264,000 app downloads and 44,000 weekly active users wearing this headgear to, to do tasks. And so you— this simple piece of headgear is training this right now, and people are using it. It's a lot more comfortable. It's a lot easier to wear. And if you don't know who Figure is, Figure in packing or Figure in shipping video, Figure in shipping video robot. Okay, this is the robot that went for 300, I believe, hours straight. 200 hours straight, basically sorting these packages, and it sorted 250,000 packages in this 200-hour test. It walked over, charged itself, and then they, they did this when they were done. Okay, so they celebrated this whole thing. And so this thing here, the company that made this is called Figure. They're the ones who started Figure Index. And the whole idea is that they're doing this to collect the robot data. What is the problem? Okay, that what they're wearing, what they're doing is not adequate to gather the knowledge to put robots on roofs, but that's not their thing. Figure's not out to train robots to do everything. That's not what Figure wants. What Figure wants is for robotics to be Ubiquitously accepted throughout the world. And so how are they going to do that? Well, they can count on us to be lazy. They know that if they can get you to put Terminator in your house next to your kids and your dog, because you are letting it do your laundry, you're letting it do your housework, you're letting it feed the dogs, you're letting it do the dishes. letting it like do all of these like random little tasks around the house, and you can get one for $20,000, people will buy it. And when the rich sister who dated the soccer player and now married the dentist has now got one of these robots, her other sister is going to be jealous and buy it because they're going to want it. And it's not that far away because robot, you can buy With rent money, that's not probably the best way to do it, but it'll probably still find out. There it is, Neo. Okay, so Neo is a home robot that you can pre-order right now. That they're like essentially it's like the Kickstarter campaign. But if you want to buy one of these things, right, $499 a month, you're telling me that you're not going to take one of these things. From any number of the companies. You've got Optimus, you've got Neo, you've got Atlas from Boston Dynamics, you've got, um, you've got Neo, you've got Figure, right? And for $499 a month, you're not going to put one of these cute little cuddly things in your house that's gonna clean, do laundry, wipe the floors, feed the dogs, like clean up after the kid when he knocks spaghetti on the floor? You bet your ass you're gonna do it. And when you Do that, when I do that, when we all do that, when we suddenly feel comfortable with these things in our homes, we will be comfortable with the idea of them working in the farmer's fields. They're going to be— they're already working in factories. Okay, if you look at Figure BMW, okay, this robot is already working inside a BMW factory. Okay, now again, it's doing the same kind of tasks that you do in the home. Moving stuff around, picking things up and putting them in boxes. It's just moving parts around inside, instead of a BMW plant, just sorting through parts and putting parts in parts bins. But this is already happening. Now, how does this become robots on roofs? Well, it's very simple. It's gonna need more data collection. And this is what we are doing at Foundation Projects. So if you went to Foundation Project, And I'll show you a little secret. Okay? Because a lot of people will be uncomfortable about this. And because I know people are uncomfortable with the idea of robots and stuff like that in people's homes, I wrote this little AI manifesto. There's this little glyph in the bottom corner. It's called a lambda. Nobody knows what it does. If you click it, it doesn't do anything. Right? I got this from the movie The Net way back when. But there's a special trick you do if you, if you type Borrowed Rocks. Or triple-click this 3 times. Your screen will glitch out and you'll come here. And this is my AI manifesto. I wrote this because I wanted to make sure that my stance was known long before these robots show up on people's roofs. And so it's kind of hidden. But this talks about everything that I have planned. But to train robots on roofs, you have to go a step ahead. We have to use these, these Manus gloves, and these Manus gloves allow us to track movement. And the reason that they're good is because the module that actually collects the data is interchangeable, meaning you can tear through the gloves. Like, anybody here who's in roofing knows you're going to tear through these gloves in no time. Right? But the most expensive part of this, there's a piece that actually attaches onto it, and that piece that actually goes onto it, that's the more— that's the piece that you can take off and put back on right here. So this detachable top module. So then what this does is it collects the data. This is what actually makes it so that your hand movement is tracked, and it actually has tensile strength. I can feel Like what you're touching. And this ultimately is what allows data to go back, right? So that you can actually understand what is happening with the robot unit, right? And you can train hand movements. The other thing you have to do is cameras for robotic training chest mount. It's the other piece that we have. Is it's not just this piece right here. There's actually a different one right here. So roofer wears this and this. So then you get 2 perspectives. You get 2 perspectives of what's being done. So then that gives you depth perception. So when you have the gloves and the 2 cameras, you now have enough data point collection to actually train a robot to do it. How much do we need? 6 million hours. 6 million hours of training data. And so in order to do that, you're going to need 300 roofers For 2 years working full-time. And what roofer is going to do it? The roofer needs an incentive. And so that's why one of the most innovative things we've done actually has nothing to do with the robots at all. The robots is technically the— I wouldn't say the easy part, but the part that's already baked in. Because the real challenging part is actually giving people a reason to do it. Because as you'll see in my AI manifesto, I have a piece here about where guys sawing the tree that he's sitting on. And that's what it'll feel like if you're asked to wear a headset to train a robot to do your job, right? That's what it was like driving a Tesla. You drove a Tesla when you weren't— when you were driving and the full self-driving wasn't driving, it was collecting the data about you driving, and then now it drives itself, right? This is the same way that Bedrock Robotics is doing it right now with heavy equipment. They're putting sensors. It always starts with the sensor, right? You put the sensor in the field, you collect the data, and then eventually you can train them, train using artificial intelligence, which again, the rate of intelligence, when we go back to Moore's Law, remember we talked about the speed of inter— these integrated circuits doubling every 2 years. The capacity of artificial intelligence is doubling every 6 months. So think about what happened between 2020 and 2025. We're now manufacturing chips that have 14.7 billion transistors on a thing the size of this Post-it note that does a job. It's a light switch that turns on and off, and that light switch is measured on the scale of atoms, and 14.7 billion of them work consistently enough to sell millions of them every single year to people in the form of iPhones. And that is the byproduct of them doubling every 2 years. Well, now the intelligence is doubling every 6 months. And so it always starts with the sensor. The same thing with Palantir, right out there in the, uh, in the military. Where did they start? The first place they started was putting sensors in the battlefield. All right, you put sensors in the battlefield first, and then they made technology to support it. So the first thing that will happen is in the home. What do they do? Right? We go back to Figure Index. They put sensors in the home, right? They put sensors on your head, and that sensor is monitoring how you do work. And then the next thing is the sensors in the workforce, and then you're going to have robots out in the workforce. All they're doing is building the platform. They're building a platform. There will be a Tesla Generation 5-C made for construction with the heavy-duty hinges and the rubberized knee pads. And that's what we are doing. We have a robotics lab in Austin. And right now you can't talk a lot about what you're doing because frankly America's a little behind. And so you can't really talk a lot about what you're up to because the last thing any of these companies want is embarrassing to happen while they're still in a global competitive state. But what we have is an innovative way to make sure that the roofer gets to partake in the benefit of contributing to that knowledge set. That's the most innovative part of what we're doing because the robotics part's kind of baked in. But what really needs to happen is to make sure that there's an innovative way That's attractive to investors that will make it so the roofers will have a long-term stake. And I'm calling it the Roofer Pension Program. So I can't go into that much, but essentially this is what I wanted to talk to you about today, was make a case for that robot— robots are not as far away from being on roofs as you might think. We will have them on a roof in 3 years. And so that's all I want to talk about today. And if anybody has questions or wants to learn more, go to foundationproject.net. If you're a roofer, if you're a roofing business owner, we are looking to partner with, uh, roofing companies between $5 and $15 million. So if that's you, we'd love to talk. And you can go to foundationproject.net and you can click Map My Exit because we would love to partner with you And if you would like to be a part of this new world, there's a, there's a chance for not only you, but your employees and your roofers to essentially have a stake in this future. Thanks so much for watching. We'll see you in the next one.