The Rise of Humanoid Robotics w/ Brett Adcock | EP #57

A message from McAfee. I'd say howdy. But I'm not a real cowboy and.

I'm not a real alien.

We're deepfakes and because of fakes like us, it's hard to tell what's real.

Unless you have McAfee.

McAfee's scam detector automatically identifies text and email scams and even deep fake. And it works everywhere, even out on the range. Yee hawk.

You're not even a real cowboy.

If they're faking it, they're not making it past us. Get award winning scam detection today. Mcafee.com keepitreal if you're alignment in charge.

Of keeping the lights on, Grainger understands.

That you go to great lengths and sometimes heights to ensure the power is always flowing.

Which is why you can count on Grainger for professional grade products and next day delivery. So you have everything you need to get the job done.

Call 1-800-GRAINGER click granger.com or just stop by Granger for the ones who get it done.

Goldman Sachs predicts robots could generate $154 billion in revenue in the next 15 years. That was their number. That's impressive.

It's going to be a big deal. We think there's an opportunity to put up to 10 billion humanoids on the planet.

Oh my God, that's amazing. 10 billion humanoids on planet Earth.

You'll see humanoids in warehousing manufacturing factories. Yeah. So the robot is up and it's working now. It's working today.

Call it over. Come on, come on over.

We're really trying to build a humanoid to just insert into the economy and hopefully do really useful and good work for humanity. This is the right decade to make that happen and I think over the next year or two we'll hopefully demonstrate that.

Everybody, welcome to the Moonshots podcast. My next guest is Brett Adcock. Brett is a serial successful entrepreneur. He was the founder and past CEO of Archer, an evtol flying car company. And since then he's been building an extraordinary humanoid robot company called Figure. We're going to jump into the humanoid robot marketplace. How soon you can expect these robots? Expect there may be as many as billions more humanoid robots than humans on the planet in the next few decades. Where are you going to see them? Can you buy them? How much are they? How are they going to impact your life and the world? Check it out with an extraordinary moonshot entrepreneur, Brett Adcock. Let's dive into the episode. Everybody. Welcome to Moonshots and Mindsets. I'M here with Brett Adcock, who is an extraordinary entrepreneur. If you don't know his name, you know his companies and you're going to know his name real soon. So, Brett, a pleasure to have you, pal.

Thanks for having me on.

Yeah. So, first of all, you know, I don't want to go into in too much detail, but I love that you took on one of the key challenges that we've talked about since the beginning of tech, which is flying cars or EVTOLs, electric vertical takeoff or landing. You know, I. I love the old saying, you know, when Peter Thiel said we asked for flying cars and what we got were 140 characters. Well, we've got Archer delivering vehicles very shortly, so congrats on that. And when I heard that you were becoming the CEO of a new company called Figure in Humanoid Robotics, I said, oh, hey, that's the second option. When people talk about we're living in the future. Flying cars and humanoid robots are the two sort of signposts that tell you we've arrived. So full disclosure, everybody. My venture fund, Bolt Capital, is an investor in Brett's newest company called Figure. And let's jump in. So what is Figure? Let's begin there.

Yeah. Well, first, thanks for having me on. So Figure is an AI robotics company designing an autonomous general purpose humanoid. So a humanoid is a robot that has some of the similar characteristics of a human. We have two legs, two arms, hands, and our goal is, over time, to put as many humanoids as humans on the planet to make physical labor a choice.

I love that. So if I were going to ask your moonshot, since we're talking about moonshots and moonshot entrepreneurs, and you're a serial moonshot entrepreneur here, which is pretty cool. Very few of those on the planet. How would you describe your moonshot, your target?

Yeah, we hope. I mean, we look at the world today and feel like most of the world was designed for humans. We have like, in the physical world, like a human operating system. I'm going to leave this door that has a handle, I'm going to grab with my human hands. We have tools, shelves at a warehouse are designed for humans to interact with. So we feel that if there's a general purpose interface to this physical world, it could be a substantial like, or basically substantial benefit to humanity doing all this physical labor that's happening to the world. So we believe that over time we should be able to solve some of these really important problems in the labor force, problems in, you know, doing work for companies, helping out at home. Caring for the elderly. And, and our goal over time would be, is put, we think, we think there's an opportunity to put up up to 10 billion humanoids on the planet.

That's amazing. 10 billion humanoids on planet Earth. So when do you, let's, let's go there for a second. If you had to guess how many humanoid robots there will be on planet Earth by 2030 or 2040, what kind of growth we're going to see there.

I think over the next couple of decades we're really going to be volume manufacturing, limited in how much supply we can get of humanoids into the market. I think if we look at a very long term, three to four or five decades, I think every human's going to want a humanoid. Just much like you have a car or phone. I think there'll be one in every home. I think there'll be billions in the labor market doing all the work that is dangerous, monotonous and boring for humans to do today. I believe over time we'll colonize space with humanoids, we'll care for the elderly. So I think certainly over a long enough period where we have time to volume manufacture, I think there'll be billions of humanoids. And then in the near term we're going to be constrained by how well the performance of the humanoids can be and how reliable they can be in the market. And I think we're really working on that problem now in earnest and with the goal, hopefully within the next 24 months of demonstrating our robot into actual real use, real life applications.

I can't wait. I want to.

There you go.

So many questions. We'll get into it, you know, in success as the cost of manufacturing these reduces and the volume increases. Do you have a vision of what the cost of a fully functional humanoid robot might get down to?

Yeah, I think you look back over time, if any consumer product or vehicle, there's a really high correlation to price and manufacturing volumes. You really want to get up on the experience curve, which is basically, basically every doubling of manufacturing volumes, your prices can fall or costs can fall by 20, 30%. So price really is in a lot of ways a real function of how much volume you're really gonna have to door. I think over the long term, you look at this first order of this, there's roughly a thousand parts in our humanoid today. It's like an electric car might have anywhere from maybe 10,000 parts, be four or five thousand pounds. We have a 150 pound humanoid with thousand parts. I think the cost of this should be less than like a cheaper electric vehicle in my mind, mostly dominated by the actuator, basically motor and sensor costs and compute costs on the robot.

So just to throw some numbers out there, like if it cost 30k for a figure robot and if you were going to lease it versus buy it, you can imagine having a, you know, a lease payment of 500 bucks a month for your robot. And that's amazing, you know. And I imagine a future in which these robots are sort of sitting there on demand, like go run these errands, go do this, go do that. And it's how I'm going to ask one time point and then we'll come back to it. How long do you think before I can buy one and put one in my home? I think obviously factory settings are going to be the first location. But is that, is that this decade?

It certainly feel like, as you said, the first use cases will be in areas that are more constrained and lower variability. So factories, manufacturing things that basically are just much more structured in nature than the home. That'll help us get cost down, safety up. We have a whole data pipeline. We need to go build out further manipulation and high level behaviors and perception policies. But I think we're probably end of decade, early next decade before we're starting, we're starting to see early life of humanoids in homes helping out. And I think it's just going to take some time. We need a lot of maturity across the product that we're going to do through the corporate labor market safety as well.

You know, I live here in Santa Monica and as I'm walking about on Main Street, I'll see these little, what they call coco robots that are six wheeled robots that are rolling down, deliver, delivering six packs of Diet Coke or burgers. And at first they're an oddity as you see them and people are taking photos and then you ignore them as they're walking by. And it's going to be interesting to see humanoid robots sort of enter the live work, play universe that we're in and there'll be like strange oddities and then just like I guess the Star wars universe where they're just every place of all different shapes and forms. Is that what you see?

It's funny you mentioned that because we have, we have a lot, we have a big presence of folks here from Boston Dynamics at the Figure Team.

And by the way, we looking in the back here. So those of you who are watching this on YouTube, is that your factory floor back there?

Yeah. So we have Basically an office here in South San Francisco Bay. I'll give you a little quick tour.

Sure.

But we're about 50 people or so, and yeah, we're based here in South San Francisco, so we have a facility.

I'm so excited to come visit.

Yeah. So it'd be great to have you. Yeah. So, yeah, what I was mentioning before is we have a couple folks from Boston Dynamics here that have mentioned wait till the robots walk around enough and stuff. And it's not going to be as exciting because everything that happens now, the ankle rolls and we're just like, oh, we're watching it. The robot just took first steps and started walking a few months ago. And we're like the whole company surrounding, watching what's going on. And it's such a spectacular thing to see robots in the office doing really useful things that the novelty still surely hasn't worn off with me yet. But I think.

Are you a dad? Do you have kids?

Yeah. So I have a 2 and a 5 year old.

So it probably feels a little bit like that early stage of toddlerhood. Like, look, look what it just did. Amazing. That is so fun. I used to build robots when I was at mit and just having them not smash into the wall, just roller. They were supposed to map the room out in units of their own length and so forth, but the electromagnetic noise would always hit the circuitry and they'd go whizzing off in some direction. It's come a long way. The last few podcasts, Brett, that I've done have been in the field of AI, and I know that you're developing your own AI there, and I'd love to talk about that a little bit because I think one of the things that makes humanoid robots possible is the AI capabilities we have now. But there's another part of the conversation I'd like to go into, which is, well, that having a physical instantiation of AIs in a robot, I think is going to be an interesting part of AI's evolution. Right. An AI that's in a box or just looking through a camera or speaker is very different than an AI that's able to actually go and interact with the world. And there's a lot of individuals who feel like it's the embodiment of an AI that's going to make it ultimately sentient or conscious. I don't know if that conversation takes place there at Figure It All.

I believe, like in the Limit here, we're going to make a. We'll have the ability to make A hopefully a substantial impact into AGI. I think there's this outstanding question that we're all debating now in 2023 is if there's enough words on the Internet to train next word prediction language models to get us to real intelligence. And if that answer ultimately turns into we're not able to do it, then I think the, the longer but the surest path is through humanoid robots that can ingest human data online and then use vision language models to basically ultimately interact with the environment and to ultimately make progress in the AGI fronts.

Yeah, I call it poking at the world and seeing what happens and learning through doing. There's a very famous and I don't know it well enough to do it justice. When Helen Keller was learning language, it was through her tactile sense, through interaction with the world and her embodiment in the world that allowed her to become sentient in that sense. Otherwise she'd be living in a world devoid of data to a large degree. And so I definitely see that the name Figure, I'm curious, was the origin there. You could have had a lot of different names for the company. Where did you choose that?

Yeah, I've been pretty thoughtful over my last three companies starting Vettery, Archer, and figured to really think about the name, the brand and setting up even the basic stuff around the brand, around the mission, vision, values. But so we, I basically spent the first nine months building the brand as well as a team and the product here. So funny enough, when we started, we were, I was like, I just told the lawyers, like put a placeholder in for the C corp. It ended up being called Adcock AI Inc. Think they can maybe like, you know, 90 days later we'll change this and nobody will notice. But it happened to be. We hire like 40 people with like, you know, you're joining this AI Adcock AI Inc. Which is really weird, but we spent a tremendous time on the name. We really wanted something that was easy to say, easy to pronounce, very unique in the category. Somewhere we can build a lot of brand presence around. And there's something about the human figure that we really, really aspire to. And we thought this name was something we could really own a lot that had a lot of depth to it. So we ultimately called the company Figure. And so far it's been great. We came out of stealth in March and the feedback so far has been really good. And a lot of the focus we have over the next year or two now will be product development milestone focus. So hopefully what people see over the next year or two would be a pretty substantial amount of product development on the product development front for the humanoid robot. So with AI systems like low level controls and ultimately showing the robot can actually do useful real world things.

So, you know, I'm going to get into a little bit later your advice for entrepreneurs. Want to take big moonshots like you did in Archer and like you've done in Figure because you know, it means raising a bunch of money, it means getting an extraordinary team. And one of the things I commend you for is the team you pull together and figure. It's, you know, when we ran through it, it was like, wow, that's a rock star team. And then it means being willing to run fast, fail, recover and keep iterating and having enough capital to do that. So we come back to that a little bit and get your advice for entrepreneurs who want to do follow in your footsteps. Let's talk about the actual robot. One second. I've got the stats in front of me and I am curious. You know its height is five foot six inches?

Yeah.

Okay. Well it's, I'm about five, four and a half, so I can almost see eye to eye for it, but it's like you didn't make it 6 foot or, or 6, 5 or 5 foot. Is this like, you know, what's the average height and comfortable to, you know, humans don't find scary. Can still reach in the top shelf. How do you think about the height?

Yeah, it's, there's a very laborious process we went through to get to the height because there's like two, like, there's like kind of almost two divergent things happening here. One is kind of from a physics perspective, you really want the height probably smaller than 5, 6. You want basically you really want the amount of power the robot area to be as low as possible, which means you want the lever arms or the extremities to be as short as possible. Yeah, you don't want like, think about, you don't want a huge arm holding a bunch of weight. It means like there's so much more power needed given the length and distance there of that lever arm. So you really want everything shorter and closer to the ground. And also when your fall, survivability is much better. Like little kids are so close to the ground, when they fall, they're fine. So I think physics is pushing you one way. And then separately, the commercial market, which is like humanoids are going in and grabbing things and reaching over shelves and reaching up high and down low, they really want these really long arms so you can reach across and grab the bin and turn it around and, and articulate it. So from a commercial side they want like this Inspector Gadget type robot that can like, you know, reach really high and have superhuman strength here or there. So it's really this balance, we think 5, 6 is probably plus or minus a few inches of where we'll want to be commercially. I think it was a pretty good first order approximation on the robot. And I think the, the next generation robot that we're designing now that'll be out this year is the same, almost the same exact height.

You know, it's interesting because you don't want to make the arms extra long like an ape because it causes a canny valley type experience. So you want, you want these to actually look humanoid. Is that true?

I think there's, you know, if you look at the uncanny valley and the research around that like as you get really closer to human looks, like there's almost like this trust that builds until this like, until there's like point really close to a human look, that it gets really scary and terrifying. So our view is that we're not trying to look like a human. We're not trying to put facial expressions in or chin or noses and ears. We want to just ultimately have the human capabilities in terms of manipulation and locomotion capabilities and things like that, because that's what's necessary to interact with the human operating system world that we talked about earlier. Meaning we don't have to change anything if we look like a human. We can just go in and do all the warehouse work that nobody wants to do, all the manufacturing work. Go cook home in your, go cook at your home, do the things without any altering of the environment, which is really, really what humanoids are for. Right. You're really trying to build a humanoid to just insert into the economy and hopefully do really useful and good work for humanity.

Yeah, but you also don't want to make it look so strange. Right. There's some comfort in thinking that it's, it looks like humans, proportionally looks like humans. It doesn't have like a third arm or extra long appendages and so forth. You know, it's. If we could, without diving into it. How many humanoid robot companies are out there? Everybody has heard of Optimus and hopefully now has heard about figure. What would you guess? Are there like a dozen or a few decently funded?

Yeah, maybe like half a dozen. Like pretty serious. Maybe have funding, have a team greater than Five groups out there that we would like maybe put on our list. I think that the vast majority of humanoid projects like last 10 years have been all research and R and D. So Boston Dynamics Atlas is still an R and D project. You have a lot of really great labs in the US at Caltech and Berkeley and other places that have like demonstrated some of these capabilities that are, you know, under research. And then commercially there's probably, yeah, maybe half a dozen groups out there we kind of look at as. Are you a commercial group? Are you walking? Do you have hands? And the only groups that we know of that have those three qualities today are US and Tesla Optimus.

Yeah, I run my Abundance360 CEO summit and every year we highlight a different robotics company. And we had Mark Raibert here with Boston Dynamics and his robots a few years ago. And then last year we had a robot called Ameca from Engineered Arts out of the UK and Ameca guide appendages, but it's, I'm calling it her special is facial expressions and movements extraordinarily humanoid. Right. In a way that's eerie and amazing. And Ameca is driven by now GPT4 and obviously Atlas is driven by its own systems and everyone's different. I don't think what people realize is ATLAS as a robot is really heavy and its hydraulic systems are really dangerous. And you've taken a different approach here with figure because the robot is a relatively reasonable weight and I think is less likely to injure somebody. And so five, six. What are the other parameters on it?

Yeah, so we're. Our Target weight was 60 kilos and we weighed in at 61.

Nice.

61. A little over 61 kilos, which is great. I've always been significantly overweight in every hardware program I've ever been on.

It doesn't work well with flying, flying hardware. But this is.

Yeah, yeah, we could talk about that at one point. But like, yeah, talk about the, the mass related engineering problems we had to solve at Archer to make that work were amazing. Yeah, it was. So we have, we have, we have, we have, we have a full charge, full state of charge before charging. Target of five hours. So we want to be five hours on, we want to be off fast charging, close to 2C and back in operations again. We want to be able to do like kind of like fast walking. We don't want to run, but we want to be able to do a, you know, basically close to a couple meters a second in terms of walking speed.

What do humans do? How fast are we as A typical walking speed?

Yeah, maybe like one and a half, two. So false, false walking. Fast walking. Yeah, we have, we have like no intention to do running or sprints and things, but we, you know, there might be times where we need to like, you know, walk a quarter of a mile down a warehouse and we want to do that in a fast way. And then we ultimately have some manipulation in basically speed and reliability and safety targets we also may want to hit internally. But for the most part we should be able to do the majority. The hardware should be able to do the majority of what humans can do today. We'll really be limited by software in our ability to do where it's at today. And what it should be able to do long term is basically just like a software, we're software update away from being able to do that kind of stuff.

Amazing. And I love the idea that your robots can be our software updated on a regular basis to increase their capabilities. The same way your Tesla might be updated on a regular basis. You know, I know Elon and Zucker are planning a wrestling match. I'm just wondering when we're going to have the figure versus Optimus wrestling match as well.

But yeah, I mean, listen, we're like, we're funny enough, we're actually right across the street from like an unmarked Tesla, like Tesla facility here in California, South Bay. So yeah, maybe if this doesn't work out, this whole commercial humanoid thing, we.

Can just A pay per view.

Yeah, we can just figure out how to make money some other way.

Yeah, that's funny. Everybody, I want to take a quick break from our episode, tell you about a health product that I love and that I use every day. In fact, I use it twice a day. It's called Seed Health. Now, your microbiome and gut health are one of the most important and modifiable parts of your health plan. Your gut microbiome is connected to your brain health, your cardiac health, your metabolic health. So the question is, what are you doing to optimize your gut? Let me take a second to tell you what I'm doing. Every day I take two capsules of seeds daily Synbiotic. It's a two in one probiotic and prebiotic formulation that supports digestive health, gut health, skin health, heart health and more. It contains 24 clinically studied and scientifically backed probiotic strains are delivered in a patented capsule that actually protects it from the stomach acid and ensures that all of it reaches your colon alive with 100% survivability. Now, if you want to try seeds Daily Symbiotic for yourself. You can get a 25% off your first month supply by using the code Moonshots at checkout. Just go to seed.com moonshots and enter the code moonshots at checkout. That's seeds.com moonshots and use the code moonshots to get 20, 25% off your first month of seed's daily symbiotic. Trust me, your gut will thank you. All right, let's get back to our episode. So I'm curious, the AI, so the capabilities of figure 5 hours runtime, good walking speed, being able to lift what like 20 kg I think is your target lifting mass, which is, which is good. Is all of the processing on board or are you doing it on the edge of the cloud?

Yeah, so all of our short horizon low level processing is all done on board. So we need a pretty substantial amount of compute and graphics onboard the robot to be able to basically run the onboard computer. It's powering the whole locomotion controller. Yeah. Run the perception systems, all the things we need for occupancy and stuff like that. And there is an ability for us to basically like, basically talk to the cloud for a certain amount of things in market that are not like that are just much lower bandwidth and latency is not as of issue. So things like high level behaviors of like what should the robot be doing next, things like that, those things can be done off board. There's no reason we have to do those on board. But for the most part we want to be able to in like a, you know, like 5G denied environment to be able to do as much onboard the robot as possible. The controller is running at such fast frequencies and stuff. There is just a tremendous amount we have to do on board at very high speeds.

Yeah, I imagine. I love to dive into the use cases. But first the timeline so folks can start. I mean, I assume that there are going to be businesses built on top of your systems. People are going to just like. I know one of my friends, Scott Painter is building a whole Tesla based service where you can buy and rent Teslas through him. And there are lots of different other approaches from what's going on with Uber and electric cars and so forth. But when are we going to see the first in commercial use? What's your expected earliest delivery?

I think the earliest we would be able to have a humanoid in one of our clients that are getting, we're getting paid for it and doing real work would be next year. So 24, I think yeah, yeah, 24. And I think if we miss that, we're not going to miss it by five years, we're going to miss it by a year or two. But I think as of now for the applications we'll be doing and the conversations we're having clients, it certainly seems that next year would be the earliest but possible to be able to do.

Are we going to start to see figure robots being demonstrated? How to put this mark Raibert was putting out fun videos of Atlas and its capabilities before we saw it in any kind of useful commercial business. I'm still not sure what commercial businesses other than defense approaches that that really hefty robots done. But when are we going to see a functioning robot? Your best guess on video or at, you know, maybe CES next January.

Yeah. So the robot is up and it's working now. It's working today. We're walking.

Call it over. Come on. Come on over.

Yeah, sorry like yeah. So we'll be putting out. We're going to try to I figure try to build in public as much as possible and keep the basically everybody abreast of what we're doing. I think it's super important. I think it'd be fun. We want everybody to be rooting for us as well. I think we'll be putting out videos quite frequently every year. And our plan is over the next, you know, two to three months here putting out the first walking videos of our humanoid here in our office. And then down the road we want to do more things in perception and manipulation and other traditional operations. But our goal is to be putting out videos to not doing the parkours and showing like the pure performance of like you know here's better than a human box jump or box flips. We really want to do just like boring work and repeatable.

Yeah, yeah.

Let's get this robot in the warehouse just to do work over and over again. And I think that'd be pretty groundbreaking. So we're like we're shooting for that with our clients now is like how do we you know get it to do work in our lab and as close of representative of what our clients environment would look like. So there's a high transferability next year into our client sites visual systems.

Is it all camera visual or using lidar? What kind of imaging systems are you using on board?

Yeah, we're full vision, 100% vision system, perception system.

You're not going with any other augmentation. So it's what a camera can see is all you need. You've made the same decision Elon made with autonomous cars, it sounds like.

Yeah. Given the distances in. Yeah. We do not think LIDAR is necessary here. As of now, we don't have the same view as Tesla and we don't have a don't look at lidar kind of policy. We've evaluated it, we've mapped our facility here, we've used it for some localization. We think it can be a helpful sensor, but that's not really the right answer. The right answer is, can you get there sufficiently without lidar? You know, having LIDAR on board, like, complicates both the supply chain. We have to fuse that data in, we have to maintain it, we have to fix it, we have to procure it, we have to pay for it. It's a bomb cost. Like, we have to maintain it. So there's like a lot going on with every single thing that goes in the robot. So we have a pretty high bar for adding things to the robot unless they're proven to be necessary or sufficient for the robot to do the operations.

I want to paint a picture of how humanoid robots are going to enter society with you. I've thought about this. You've thought about it more. Just to give a couple of data points here. Right. Goldman Sachs predicts robots could generate $154 billion in revenue in the next 15 years. That was their number. That's impressive. And these are humanoid robots. They're not the robots we're seeing in factories building cars and packing plants and so forth. And I think the other thing, when you and I were speaking early on when I first met you, you know, you made the point, listen, half of the global GDP is labor, and that's your total addressable market. And that's amazing. Do you find people telling you, oh my God, you're going to displace jobs and you're going to cause a disruption, like AI is causing a disruption. And do you remind them that we have so many filled jobs and the labor market is really becoming tough in different places?

Yeah. So if we look at how I think the business unfolds over the next 10 or 20 years, I think it looks very similar to what you saw in self driving cars, where the easier stuff will be demonstrated first. So driving on the highway has been demonstrated at higher safety levels than driving in, say, San Francisco, the city. And it's because in the city it's got a higher safety case, it's more variability, it's less structured. It's probably like, you know, one or two orders of magnitude Harder from an engineering perspective to do that reliably and safely than on a highway. The same thing exists for humanoids. There are applications in the world that are easy to do. You're moving bins or boxes, you know exactly what the bin is, you know exactly the payloads, you know where you're moving it to. You're in a, basically a space that you already can map and you already understand. You can have communication with the manufacturing or warehouse execution system. It's a really like well known or kind of highway driving equivalent. And then there are things that are really much harder which are, you know, cooking somebody food in their home, caring for the elderly. Those are like city driving equivalent to self driving cars. So my, my strong view is that a lot of people have a misconception of humanoids because so many people have been working on the latter. They've been working on this really hard consumer problem. If you look at Google's robot with sorting trash, it's very difficult. Toyota Research Institute's been working in the, you know, in grocery stores and things. Those are really hard problems and I'm glad people are working on them. But for commercial groups we really need to do the easier stuff that's necessary first that we can start demonstrating and building AI data engine into the harder stuff over time. So we're almost like almost the opposite of what the research groups are looking at today. So I have a strong bias that you'll see humanoids in warehousing manufacturing where the talent shortage is the most acute. And as you hit, you know, from the macro perspective, you know, half the world is like GDP is labor. We've had, we're having this huge issue with labor population globally. We have, the baby boomers are retiring. The amount of kids we've had has been in basically like been secular decline for like several decades.

Yeah, it's crazy. People don't realize we're in a, one of the greatest tragedies is not overpopulation, it's going to be underpopulation.

It's going to be a big deal. And so you're seeing that. So we walk into a client site like say a big Fortune 100 company, their first thing out of their mouth is like not, you know, how's this going to like be used with my employees, things like that. Their first thing in their mouth is that last year we saw 140% annual turnover in a warehouse. We, we have nobody who wants to do these jobs. They're dangerous. They're hot in the summer, they're cold in the Winter, the turnover.

Dangerous, dangerous, dull and dirty is the phraseology.

Just like nobody, they can't find anybody that wants to do this. And so we walk in there, they're like, if you can do these things, we will buy your service hand over fist. Now doing the humanoid thing is nobody's ever done it before. So it's the hill to climb. To do that successfully is, is extremely challenging. And we happen to believe that this is the right decade to make that happen. And I think over the next year or two, we'll hopefully demonstrate that here.

So I have to imagine that the large language models that feel like they're coming on just in time for the humanoid robot marketplace, so that I could speak to the robot and have it understand what I want and clearly say, yes, I get it, I'll go do that right now and have a conversation that's meaningful. Are you going to be building your own large language models or are you going to be incorporating other ones? And when does that enter your sort of your build process?

I think the way we're going to get humanoids, say out of the factories and into people's homes, like we're working with humans, is going to be through language, as you mentioned. So we think there's a substantial benefit our business has to using like basically large language models or vision language models to basically help us understand like we need like a semantic understanding of the world and language can bring us that and large language models can bring us that. So we will be building here over time vision language models to really help from like a high level behaviors perspective of letting the humanoids understand what humans are saying and be able to talk to humans, but also be able to infer and understand what they're saying and be able to react to it. And so we will most likely not be building our own language models, but being able to train vision language models on the robot system as it relates to the sensor data that's coming off of there and to be able to do useful thing with those models is going to be something that we're going to have to do internally in our doing now. Internally it's going to be extremely important to build that AI data engine correctly so that data coming off the robot can be trained accurately and the neuralness can be trained correctly to deploy over time. And that's what really drives our interest in getting to market as fast as possible. The more robots we can get into market collecting data, the smarter our fleet of robots will become in the future and the more applications the robot will learn how to do.

I mean, people don't realize that these robots, because their AIs are connected, their data sets are connected. When one robot learns how to do something uniquely or runs into unique situation, it isn't that one robot that learns it. They all learn it. And that's a beautiful, a beautiful thing that loves scale.

It's like my kids, right? Like once you learn how to do something they like, they fail like a thousand times. So like little like reinforcement learning policies and once they figure it out, they like, they really don't forget it and then they just keep building on that. So yeah, once we train a robot on how to unload boxes from a pallet successfully, every robot in the fleet will know how to do that. Once we train the robot on how to unload a truck successfully and to manipulate certain boxes that are damaged or whatever it looks like, then every robot in the fleet will understand that. And so it's just like it's going to be a huge power curve that we're going to be able to. Yeah, it's going to be a huge power curve for us in the future.

Yeah. So there's a huge advantage to get that, get out into the real world and get those models learning. I want to go back to the founding of the company. So you had an incredibly successful exit early on with your first company that was in the hiring space vettery. And then you start Archer really hard, challenge Evtol and build it, demonstrate the technology, take the company public, start getting orders and then you decide you're going to take your next moonshot. I want you take me back if you would. Brett to you've retired out of Archer. Did you know what you wanted to do next at that time or did you start sort of looking around and saying, what's the next challenge?

Yeah, so I think, you know, when I left Archer was a really good time for me to kind of really reset. And you know, what did I really want to work on? And I think, and I still say, I still say I said this kind of like off the cuff to somebody yesterday. I said if, you know, if somebody came to me tomorrow, wanted to purchase a business for $5 billion, I would say no. And you might think that's a little bit crazy, but I think, you know, one of the greatest assets for an entrepreneur is like really loving what you're working on and be able to spend a significant amount of time on a really hard problem and build hopefully a really great business long term. I can't think of a more important business for the World than humanoid robots at scale, serving mankind and helping out. And I think it's a one, a a really hard problem. B it happens to live inside the largest economy on the planet of human label. Three, we think the technologies necessary to do this are kind of exist today and have been demonstrated and much more of like advanced research state. And four, I think in the limit, I think we, I hope we can make some progress towards AGI here at figure so. So from a founding company perspective, I looked at this and said that this is somewhere where I could probably spend the next 30 plus years of my life.

Take me back a little bit more. Did you have this in the back of your mind while you were at Archer? Has this been something like from the childhood you saw Rosie the Robot or you saw Lost in Space? Space and robots have always been a fascination for you. I mean autonomous cars and flying cars are robots of a type. But when did you say, I know when I'm building my companies I have this moment in time where something crystallizes. That would be a next cool chapter to work on. Do you remember that moment growing up?

I've always have been a huge sci fi fan, so I've read all of Isaac Asimov books and kind of in my high school period I really realized that I wanted to spend basically the rest of my life building companies. And a couple of the areas that I really thought were important were robotics, AI, the Internet. And yeah, I think from a, like I've, I think for a long time I've always felt that robotics were extremely important to industry. They were really hard. Like there's not a lot of entrepreneurs like tackling this problem. It's.

No, they're not, they're not. Just like there never used to be a lot going in and building the space industry. Right. There was just the government that did it or in this case GM made giant robots for building cars. Do you come from a family of entrepreneurs? What was it that gave you that entrepreneurial bug? And was this like high school years, college years?

Yeah, so my family. So I actually grew up on a third generation agriculture farm in the middle of Illinois. Um, yeah. So never really been an insider my whole life. Um, my parents, yeah, were entrepreneurs for many generations. I remember growing up, my parents were like, you know, at some point you want to control your own destiny, you need to do it yourself and you need to go out and build stuff for the world. That's really real way to impact. And I think, you know, my awakening was I thought technology was probably the greatest lever arm of my generation. It's the area that I could spend a lot of time in and make the greatest impact. So I've been now building companies for 20 years. I did 13, 14 years of that in software and Internet and last six or seven years have been in advanced hardware and AI areas and to be honest, be able to work on these things. And I know you spend so much time in these areas, it's like, what a blessing, right? It is hard problems and they're just so fun. And we're almost inventing the future.

And so we are. We're predicting the future by inventing it ourselves. I want to quote you, I love this quote. I don't know if you remember saying it, but you said we have the potential to alter the course of history and fundamentally improve millions of lives. It's time to build. Couldn't agree more. Right. And like when I tell people when you're setting your massive transformative purpose and taking your moonshot, like stop building another photo sharing app and go do something that's a hard problem that is going to change the world and make the world a better place. So thank you for that. You know, I'm super passionate about longevity and healthspan and how do you add 10, 20 healthiers onto your life? One of the most underappreciated elements is the quality of your sleep. And there's something that changed the quality of my sleep. And this episode is brought to you by that product. It's called eight Sleep. If you're like me, you probably didn't know that temperature plays a core crucial role in the quality of your sleep. Those mornings when you wake up feeling like you barely slept. Yeah, temperature is often the culprit. Traditional mattresses trap heat, but your body needs to cool down during sleep and stay cool through the evening and then heat up in the morning. Enter the pod cover by eight Sleep. It's the perfect solution to the problem. It fits on any bed, adjusts the temperature on each side of the of the bed based upon your individual needs. You know, I've been using pod cover and it's a game changer. I'm a big believer in using technology to improve life and eight Sleep has done that for me. And it's not just about temperature control. With the pods, sleep and health tracking, I get personalized sleep reports every morning. It's like having a personal sleep coach so you know when you eat or drink or go to sleep too late, how it impacts your sleep. So why not experience sleep like never before. Visit www.eightsleep.com that's E I G H T S L-E-E-P.com moonshots and you'll save 150 bucks on the pod cover by eight sleep. I hope you do it. It's transformed my sleep and will for you as well. Now back to the episode. And so you retire from Archer. You hopefully, I'm sure you did do financially well because you've been personally investing in figure a sizable amount of your own net wealth and using that and recruiting some of the best people. And whenever I'm looking at investments through my venture fund or co investing as an individual, I'm like, okay, how much are you putting into the deal? And, and so you, you're putting a significant amount to the company and because you believe in it and you've built an amazing team. What was, how did it begin? What was the first thing you did? I mean this is, this is sort of a one on one training for the entrepreneurs out there that want to take on a moonshot. And I don't know if you want to go back to the days of Archer beginning or figure beginning. But what's your advice to entrepreneurs who want to take a shot at the gold ring here?

The first year of building companies is like one of the, it's like one of the best experiences that I never want to do again. You know, it's like, it's. They're just so hard. And I remember for like, for nine months here at Figure, I was like living in this like really cramped wework phone booth in Palo Alto. Just like cold calling everybody in the space, like talking to every humanoid robotics person anywhere in the world.

Were you educating yourself or recruiting or Both?

Both. Like trying to get answers, trying to understand, trying to find the best people in the world, trying to get referrals, trying to find this, you know, offshoot small lab that nobody's ever heard of, where I could understand actuators better or you know, locomotion controls better. Trying to find the rare book here or there that could help educate me on, you know, whole body inverse dynamics or MPC controllers or whatever it would look like. So I just spent an enormous amount of time, same story at Archer for the first year. When I started Archer, I was in a room just making phone calls, reading anything I could get my hands on and trying to figure out how to, how do I get this built. I think as an early founder, the most important thing is to show that you have a product or even a minimally viable product in the making that's like what? We're all here to ship product or services, like that's it. And so the most important thing you could be doing is getting to that point in a lot of places, a lot of companies. It could be like, you know, raising capital, hiring some people. In my case for both Archer and figure was putting my money where my mouth was. So I put millions in both companies the first year and I went deep into bringing a team together and also deep into like getting myself up to speed on how this works. So at Archer I basically moved back down temporarily to University of Florida where I started undergrad engineering. I partnered with the aerospace and mechanical engineering lab at Archer at University of Florida. That lab that was building drones at the time was off of Archer Road. So I called the business Archer Aviation and I basically built a 4,000 square foot Archer Aviation EVtail lab at the University of Florida which is still there today. And I basically built three, four generations of electric aircraft down there with a small team of PhDs and that really helped me understand the technology, understand certification decisions. We had to make decisions Peter on like do we put a pilot in the aircraft or make it autonomous. These are like you know, business decisions that affected timelines and certification, all this. And so yeah, the early first year at Archer was really me, you know, self funding, getting the initial team set up and then also building small gen like basically subscale versions of what we have now which you know we have I built now 6,000 pound five passenger EVTOL aircraft and you know fast like you know four years before that we were building 20 foot aircraft that were kind of more hobby grade than what we have now. Yeah. So and then you know, maybe fast forward to figure. I spent a large percentage of my time understanding the technologies, basically designing out and architecting what the first generation robot would look like and then building the team up. And I did that self funding the whole way. So I basically felt like I didn't have to answer anybody and I could just move at lightning speed and make decisions really quick.

How you describe this is exactly I've known Elon for 23 years from the early days of SpaceX and it, it sounds exactly like his early days at SpaceX where it was like find the textbook, read the chapter, learn the stuff, interview people, begin building. And then one of the other things I think that's interesting early in the early days of a startup is understanding the highest value trades and what the limiting parameters are. Right. To help you decide, you know what. So for figure what have been the limiting parameters that drove you to make one decision over another. Has it been battery life, material, weights, AI?

Yeah. So we spent a considerable amount of time understanding the requirements. We're very like, we have a very strong axiomatic design process here. I have a very strong philosophy for it. And then we do a lot of like what you call like trade settings. Like basically like what are the right decisions we need to make? I would say at the highest level, the battery and actuator side are very mature. We have enough energy and power density out of the actuators and the batteries to do what we need to do with humanoids. I think the locomotion controls of balancing and walking robots are really mature for folks that know what they're doing. And I think a lot of the bigger trades came down to the availability of either software or hardware off the shelf that we could purchase to make this work. I think the limitations in actuators, middleware, operating systems, batteries, control software, some cases of perception that I thought would be easier to procure off the shelf and put into a robot were really not the case. I think I got most of that stuff wrong. I think there are no good actuator solutions on the market. There are really not any good battery solutions, there's no good control solutions, there's no good middleware operating system solutions. Sensors, there's some off the shelf COT sensors that are fine. And almost all the electronics for the next generation robot we're building ourselves in house. And that's not because we want to, it's because we're being forced to.

So that's a fascinating. So is it fair to say you were a little bit naive getting in and then you discovered the realities of the difficulties and had to solve them because you were already heading in that direction?

I think if I knew how hard Archer was and how figure was, you know, who knows if an entrepreneur would have started those businesses, you know, like, they're certainly extremely hard. Yeah, I would say definitely did not understand the maturity of the supply chain there. I don't think a lot of people really understood that though too. I think a lot of robotics startups think that hardware is just like easily to procure and it's really a software issue. It's. It's really not the case. You really, in order for good software to work, you really need good hardware and good hardware is like I think harder to find the good software. I think the hardware in the space is especially for Evtol aircraft and I think even for figure, the hardware is, is really hard and I think A lot of folks think that the hardware is there and it's just an AI software game. And I think that could be. It couldn't be further from the truth.

Yeah, I've heard this so many times for entrepreneurs. I know it's been true for myself, where, you know, everything looks rosy from a distance and you think all of those solutions are there, you just have to cobble them together and it'll work. But you find out, no, there's a ton of work to be done, but you've spent so much time and money already, you can't turn back and you've got to dissolve them and you have to work. You know, when you're going through hell, keep going is the old saying. So that is fascinating. And then I just did a podcast with Palmer Luckey. Do you know Palmer, the creator of Oculus. And he's building Enduril, $10 billion defense company, which is amazing. And you might look at all of the hardware. I laughingly call him the real Tony Stark. And it's amazing what he's been building there. But they're a 60% software, 40% hardware. But from the outside they look just like a hardware company. But they've had to build both together. What ratio do you see internally now?

We're probably a little bit bigger overall in all software if we include controls, middleware and autonomy than we are hardware. Our hardware team is 15 or so. So yeah, definitely software would be a little bit bigger. Software will definitely be as the biggest part of the company long term, in the limit. We think of figure as an AI business, so we'll have a large autonomy team. And there's a, there's a very significant AI data engine that we need to build here long term. But the hardware stuff can't be overlooked. Like if you really want to play in Humanoids today, you're going to need to develop your own actuators, electronics, battery and then almost all the software. There's really not a COTS or commercial off the shelf solution for this. If you want to play in the. If you want to do it. A high performance, high reliability, high safety and low cost. There's no other way to say it. I mean we have a term here which is like the only way out is through. And we use that a lot because like, you know, in hardware it's like death by a thousand cuts. It's like bringing up the robot. It's just like problem after problem after problem. And it's like things look good and there's like mountain more problems and yeah, it's just. It's brutal. I mean, being in soft, being in software for so long and then getting into hardware. Hardware is just. Hardware is just hard. It takes a long time to get stuff. It's a long iteration cycles. It's really the attack time of iteration cycles that kill you here.

Yeah, no, I like to joke. It's an overnight success after 11 years of hard work.

Yeah, totally. Yeah.

So let's allow people to live in the future a little bit here. We're going to see these robots. There is a ton of capital going in and, you know, I put my bet on figure and I'm excited and hopefully will massively be successful. But there are a few other companies as well, I think, obviously, as you've said, going into the labor space of warehouses and environments that are dull, dangerous and dirty. So if you were going to be. I'm not holding you to this and hopefully willing to tell me, but, like, what kind of robot production rate are you hoping to achieve in the next few years? And will they all be going into warehouse settings? Packing, unpacking, trucking, logistics? Is that the first sort of circle of capabilities you're circling up?

Yeah, we're really spending most of our time now on like, logistics, fulfillment. And we spent a decent amount of time now at the larger car OEMs in the world.

Interesting. What are you gonna do for them?

There's just tremendous amount of people at these facilities. We just went to a facility, a large OEM that you would know in the US last week. They had, you know, close to 10,000 people on site. There were, I mean, a lot of stuff that we could do to help them out that were. They were having a lot of problems in not find enough people. A lot of these were dangerous. They were working next to other machines. They're doing like tons of spot welding, so you can like smell the fumes, you know. So, yeah, there is like, there's different things of so large fulfillment and logistics areas of these facilities because they need to do, just in time, inventory. They need to have. This facility we saw, you know, had, you know, roughly 4 to 5 million parts that were touched by humans every day in one facility or one location at a facility. Yeah, it's a lot. So think about the amount of fulfillment you have to do a lot of touches or human touches that need to happen there. So there's. For fulfillment areas, there's also a lot of overall just like sheet metal being moved around. Right? Like a lot of sheet metal being moved to different machines. Those being spot Welded and that being repeated over and over again at hundreds of stations.

And they've got to deal with, I mean, are they operating 24, 7 or are they operating 8 hours a day?

No, they're operating almost like 21, 22 hours a day, two 10 hour shifts. Yeah.

So I mean the robots can in fact work, you know, lights out, meaning they don't. They can operate on 24, 7 basis. There's no drug testing, there's no vacations, there's no, you know, insurance. I mean it's, it in one sense for the type of job, if there's a good product market fit, they're the ideal laborers in that regard. Going back to the production rate, do you see hundreds of figures being produced in the next few years? What do you hope to get up to by the 2030 timeframe?

Yeah, we think about the businesses like we need certain stages of maturity to unlock the next kind of almost the next phase. I think the big stage we're in now is can we show a robot, a humanoid robot can be useful in a client scenario in a real way. Like can it hit the performance, can it be safe, can it be reliable? Is it going down all the time? And you need tons of human interventions. That's not helping anything. So it needs to make ROI sense and it needs to be safe and reliable. Ultimately, if we can prove that, then even in these very specific class of problems like of moving boxes and bins, we think there's an ability to ship tens of millions of humanoids. I would say it would take us decades to do that. Take Tesla and Ford a little over a decade for each company to put a million cars in the road. So if you want to put a million robots into the world, it takes no sooner than five years, maybe no, no longer than 10 or 12 years to do that based on historical precedent. Now, I think the manufacturing processes for this will be very different and I would say less complex. There's roughly maybe 10,000 parts in a Model 3 car. We have about a thousand in our robot here. And it's like, as we mentioned earlier, a lot less weight and mass. So I think we can manufacture at pretty high volumes as we relates to later in the decade. But the next like two or three or four years, Peter is going to be figuring out can we make a useful humanoid?

Sure.

And I think we really got to get there.

Yeah. And I have every confidence knowing you and the team that you've, that you've built, that you will get there. I imagine the same way that AI is Going to code AI. I assume that you're going to use robots to help build robots too. There's got to be some feedback cycle there.

Yeah, I left this big auto group last week and I was like we are going to. For anything a human is involved with in the, in the manufacturing process. Just substantial. Like the facility we saw had you know, giant robotic arms everywhere, hundreds of them. And it was about as automated place I've ever been in my whole life. And then there was like another 10,000 humans at the facility. So I think, you know, we, we want to have only humanoid. We wanted to design a manufacturing process where only humanoids are building humanoids.

I love that. A von Neumann machine at its best. So what's next? You hit logistics, you hit warehouses, you hit delivery services. What do you imagine the next big market would be? We'll start to see these humanoid robots.

I think you're going to scale into the commercial labor market for well over a decade before you even try to touch some other things in there. We have like healthcare areas, real estate areas, construction. There's other areas of like retail that we look at I think are extremely large markets. And then there's all these other markets that would like don't exist. Yeah, for humanoid robots or human like that we could basically go into like. Yeah, I think you know, like what. Why do the major tech real estate brokerage, like real estate places like Zillow and stuff not you know, have any human brokers on their platform is because their multiples get killed. I think it'd be pretty cool. I think it'd be pretty cool going to a website, booking a, a house visit and then having the humanoid greet you by unlocking the door and in a staged house and basically selling that whole house digitally. It's a trillion dollar market. It's nothing that the tech guys want to touch because it's just so manual today. But there are just many industries like that that can be done through teleoperation or other things that'll be born out of this technology.

One of the markets I look forward to that is so desperately needed is supporting our, our aging population. Right. When you've got a mother or father and rather than, you know, one of the things that's terrible is when you take your aging parent out of their home and put them into an old age home, right. There's a very rapid fall off there. They disoriented, they don't feel comfortable, they don't have privacy. But there is a vision where these robots are taking fantastic care of, of of humans as they age. What's the, what's the breakthroughs that are going to be required to enable that to happen?

Yeah, so this, this is an area that's pretty close to my heart. My family's actually involved in independent and assisted living facilities.

Beautiful.

Yeah. One of the hardest things you mentioned is like, nobody really wants to leave their home for, you know, the personal home they've been in for decades for, for one of these facilities, a really tough process. And we're having, you know, like, we're basically having this huge amount of people getting to like later stage retirement and you know, being able to do at home care would be like a very substantial life benefit to let people age in place at home. You know, this is really just a maturity of a technology itself to be able to make the reliability and the safety and the cost get to a point where we can do these things on the whole. Like the robot from a hardware perspective will be able to do like almost all this work that would be, would be needed in somebody's home. There will be a maturity here of trust of the product and some of these other aspects I mentioned that'll be important for us to mature in the commercial market. So maturing these, with these big corporate groups that have these big labor areas, making the robot more intelligent, more dexterous, higher reliability and then ultimately doing higher volume manufacturing to get cost down will be important to enter this like, elderly area to let people age in place. So I think of this very similar to the consumer conversation we had earlier. It's going to happen like a decade from now. It's going to be very substantial, maybe even a bigger business in some ways than, you know, the consumer side of things. But it's just going to be the second chapter in the book with the first chapter being the commercial market.

Yeah, I, I believe that. When you get into the home and you get into the elderly, do you imagine that future? So the first robot's called figure 01. Yes. So are you going to just keep the generations figure 02 and iPhone 3 and iPhone 27?

We made room to go to 99.

Yeah, that's good. I got that much. So as you get to figure 5, 6, 7, do you imagine there may be a humanoid facial personality that you add for comfort and because, I mean, there is a value in having a robot who's got facial emotional expressions that you can feel they connect with you and so forth. And I think as the AI becomes, as we head towards AGI, that ability to recognize the person you're serving their emotional situation and to convey emotional response thoughts on that.

There should be no reason why we we couldn't do that. Like we our head today has basically a full wrap screen in the front that can convey information like what the robot is doing, maybe a prompt, things like that. And then we have sensors, we have basically camera sensors and other things in the head. So there should be no reason we can't display the right information back to the end user to make them feel comfortable. Whether we're a caretaker or we're actually doing work on site at a big corporate fortune 500. So certainly today language and natural language processing is good enough to basically have a conversational understanding with a robot. Getting that the visuals in a place where it's comforting is certainly possible. We haven't spent a lot of time on it just given how early we are in the business, but I don't see any reason why we can't give the consumer that experience.

Hey everybody, this is Peter. A quick break from the episode. I'm a firm believer that science and technology and how entrepreneurs can change the world is the only real news out there there worth consuming. I don't watch the Crisis News Network. I call CNN or Fox and hear every devastating piece of news on the planet. I spend my time training my neural net the way I see the world by looking at the incredible breakthroughs in science and technology, how entrepreneurs are solving the world's grand challenges, what the breakthroughs are in longevity, how exponential technologies are transforming our world. So twice a week I put out a blog. One blog is looking at the future of longevity, age reversal, biotech, increasing your health span. The other blog looks at exponential technologies. AI 3D printing, synthetic biology, AR.VR blockchain. These technologies are transforming what you as an entrepreneur can do. If this is the kind of muse you want to learn about and shape your neural nets with God, go to dashmandis.com blog and learn more. Now back to the episode. I want to hit on my third favorite market and use case. So actually let me ask you what your favorite third market. So we've talked about the industrial logistics, you know, manufacturing and we talked about healthcare as another as a broad brush. What's your third next big market that you're excited about And I'll show you mine.

Yeah, I mean I would love to really help on the consumer household side and caring for the elderly. I think that's such an important business long term. I think everybody will have a humanoid as like an assistant to do things. And I think another one that really doesn't get a lot of attention is I think planets will be colonized by humanoids.

Ah, now you're hitting the one.

I was going to say space exploration. Yeah, just like so excited about. We're in this like such a great time for space exploration and it's being enabled by this infrastructure that's being put in place for basically launches and rocket launches. And I think, yeah, I think humanoids will be a really great tool for humanity to help colonize and set up colonization facilities in places like the moon and Mars.

When I land there, I want the pina coladas ready, I want the bed preheated, I want all the resources dug out. I don't want to do the hard labor there. I want it pre done.

We'll be ready for you, Peter.

Thank you.

Get the pizza oven.

You know, it's interesting. Years ago, one of the companies, one of the moonshots I took, swung and missed was at a company called Planetary Resources. We were going after asteroid mining. And you know, I was warned it was too early and maybe it was, but what was wrong was I didn't have enough personal capital to see it all the way through. And when we missed a financing, it really, it tumbled and we had two launch failures and I couldn't survive all three of those. And I'll take a, I'll take a swing at it again. But with $200 million of disposable capital to spend on that and not have to wait for someone else to decide because as you know, there's a huge advantage of being able to just fund it beginning yourself and not have to convince everybody. Because a lot of moonshot entrepreneurs will spend 70, 80% of their time raising money trying to convince person after person versus doing the hard work. And in the asteroid business, having humanoid, I mean space exploration is so enabled by able to send robots out there to do the work and prep the materials. And it's a massive multi trillion dollar marketplace. So excited for that part of your business.

Yeah, I'm excited for you to get out there and do round two here.

Yeah, me too. I am curious. I am a space and tech geek. I know you are too. So I'd love to review what are your favorite robots? What did they get right and what did they get wrong in the TV and movie world?

Oh, like not like the Sci Fi world.

Yeah, the Sci Fi world. Yeah. Let's take it. Let's take it. Science fiction. Like, you know, you want to go to, you want to go to Star Trek with Commander Data, you want to go To Star Wars? Do you want to go to Lost in Space? Do you want to go to, you know, the Jetsons? What are the robots that like. Yeah, man, they got so close or. That was really stupid.

It's, it's like so funny that like, it's almost like the last couple decades, it's like the next couple decades is like making all these sci fi movies that we grew up with and novels real. It's like you look at flying cars coming, we have rockets going, colonizing planets, we have home robots hopefully coming, we're.

Extending healthy lifespans, we're going to longevity.

We're almost like predicting the future like 50 years ago. And I don't know, it's just like so funny how all this is like in so many ways coming true today. And I think if you just like, you know, even harmonize, if we're going out in the world, we're seeing humanoids out there doing work, it's gonna, which I think could happen in like the near term, like in our lifetimes. I think it's gonna feel like 50, 100 years has pulled forward into the present. It's just gonna feel crazy and it's gonna be hopefully spectacular.

Yeah, I think so. So I'm going to, I'm going to, I'm going to hold your feet to the fire here. Commander Data at Star Trek. Do you like Commander Data as a robot?

I wasn't the biggest Star Trek guy growing up. I was more in, you know, Isaac Asimov side of things and few others. But yeah, I mean, Commander Data is great.

Okay. A positronic brain is a useful thing to have and so we'll see that around figure 07 probably. All right, how about RTD2 and C3PO? What'd they get right? What'd they get wrong?

Yeah. What's your view of this?

You know, I like C3PO, but you know, that sort of lever arm on little, you know, actuators sticking out of their arm. I didn't like that very much. And R2D2, I just, I don't think you're going to build that shape and form and find it very useful. I guess in the back of a speeder it might be useful.

Yeah.

I remember one of the robots that really influenced my life early on was in Lost in Space in the original TV series. You're probably too young to remember that, but they came out with a recent one that looked pretty cool. So any other robots from the visual.

World like the, the R2D2 story, it's like I feel like everybody is, like, coming at me now and saying, like, what? You got to put some wheels on this robot? Like, why are we walking? Why are you dealing with that complexity of wheels? And you mentioned the form factor is probably not right. I just, you know, we have. I have a hard time seeing a fixed wheelbase, which is even, like manipulators working extremely well in the market. There's a lot of people that have chased this for. For quite a long time. I mean, if you go into a warehouse that. That robot that needs to have, like a z axis, it needs to have like an elevator moving down, you need to pitch it forward, like in the back, so you can reach the back of a shelf or something like that. And then you're basically getting roughly to the same complexity and actuation and degrees of freedom than you would a humanoid. And so, yeah, it's funny, the R2D2, it feels like everybody's trying. The skeptics are all trying to force me into an R2D2 form factor, which is kind of funny.

Yeah, well, I'm glad, you know, there's. There's a level of purity in actually going after the human form. Exactly. So I am curious about one other thing, actuators for muscles. Because when I was a kid, I remember reading about actin and myosin and muscle contraction and so forth, and always hoping and wondering, would they come up with a material that when you apply an electric current, it contracts like a muscle does. Right. And that. That would be the ideal actuator to replicate a robot with versus a rotary function and a screw function and so forth. Did you look at that? Are we. Are we getting any kind of electromechanical muscle tech coming our way?

The human body is just so spectacular. Like, the way that our muscles work, even the joints, like our, you know, ball and socket, like, you know, I say shoulder has like three degrees of freedom. So we, you know, for figure we have, you know, pitch on roll here, we have to do through three different actuators that are almost like, you know, activated serially across the kinematics. So, you know, imitating and getting to where the human is at in terms of degrees of freedom and efficiency is just extremely hard. Like, we're going to be off by a decent amount for a while. We've looked at a lot of different technologies, including a lot of hydraulics and other applications outside of just like, rotary or linear electromechanical actuators, and really have a hard time hitting any of our requirements for maybe packaging or mass halos. Yeah, I mean like, it's almost like this debate of people like, well, what about this or that? It's like we have no problem hitting any of these with electromechanical actuators. Like the, we have enough, we have enough energy and power on board, we have enough degrees of freedom, we have the right speeds and torques out of the actuators to make this happen. There's just like, there's, it's, it's just sufficient to make it work. And at high volumes, we get the cost down quite a lot. There are some areas that we're spending a decent amount of time on, on more of the academic research side that we think are really interesting, but we think they're just a little bit far, far enough out where they're not applicable to being able to put onto a humanoid and to do useful work the next few years.

As we wrap this up, I want to ask you again to serve our entrepreneur listeners here. You start a company, what are the most important things that you did in starting this company in terms of creating culture, hiring the right people, like, lessons learned. This is your third really big major success and you know, a lot of people, hopefully one out of 10 might be a success. Not, you know, three in a row, hopefully, but you had to learn some lessons and like, like, I'm not going to screw that up next time or I'm going to make sure that this is page one, line one of the company. So what's your mentorship for entrepreneurs for doing something like this?

Yeah, I really don't have these like heuristics of like, hey, you just got to do this and it's going to work. I think, you know, this is like problem solving at its finest. It's being extremely robust with these decision making processes. I definitely have like the playbooks that I've been operating for a long time that have seemed to be successful here at Figure and Archer in my past companies. You know, first is really identifying a really useful idea that, you know, can, can A, can work and B, like it satisfies the personal goals of why you set out to do it. And for my case, it's really not about the money, it's about making a much larger impact as I can while I'm alive. And I think this is an industry where I can make a extremely large contribution to humanity and it's just going to be a really spectacular future if this works well. So I think it really aligns well with what I'm trying to do on a personal perspective and on a Mission two is, I think getting the right team in place is probably really important. If you work backwards. What is the goal of the company? It's to ship a useful product or service. You need a team to go do that. It's generally not especially in stuff that's harder. It's not one person sitting around. It's a bunch of maybe the best minds in the world working really hard together to make this really happen. So I'm a very strong proponent of building the right team from the get go. I'm very strong in terms of trying to set the right direction for the company. You really have at a company, especially early stage, you really have two levers. You have a compass and you have a speed. You really want the compass to be dialed the right way and you really want to hit the gas as hard as you can. So I still.

But in that order?

Yeah, you have to. You don't want to hit the gas the wrong way. It's like really tough. It's really like, you know, you're early enough, it's fine. But when you get bigger, it's like a big ship with a small rudder. It's really tough to change directions. So I figure in both Archer, but a figure. I wrote the master plan, which looks online, it's like a ten year vision document. We have a culture doc, which I also wrote in terms of like defining our culture and like what you should expect if you come over here.

Can I just pause? I want to pause there because those two things are what I wanted to call out of you. Right. I think having a vision document, right. That is clear and defining, that aligns the organization from the beginning, your master plan. And a lot of times if an entrepreneur is a founding CEO, founding entrepreneur, if you're not clear about that and if there isn't that compass heading and you are hiring high horsepower individuals, they can start tearing you in different directions. And that alignment is so critical. And then the second thing you said, which is equally important, defining and creating the culture you want from day zero. So is probably the most important thing I think you said here.

Yeah, I think like, it's funny, I'm almost thinking the other side of this, what I just said, which is a little counterintuitive. But I remember like, you know, 15 years ago I'd read so much literature about blitz scaling and how to maintain a good culture. And as a young entrepreneur, you look at this and say most of everybody dies. They don't make it two years. And I'm sitting here reading how to blitzscale and how to set the right culture 10 years from now and how to set the right direction. And I can't even feed myself, let alone make the company work. And so a little bit of this is having done it before, it is important. At the end of the day, what is maybe the. The last thing I'll add to this section would be we got to get out and ship product. And I think at the very core of what I love about, you know, building Figure and Archer and Veter, we had a strong mission belief that we need to ship product and ship it fast. And I think if you come here, one of the big shocks somebody might have from a big corporate somewhere else is that we move incredibly fast and we want to ship product as quick as possible and recursively make it better. And I think the rest of the stuff is almost like these guardrails to help that out. Like, the culture, the master plan, the people, everything is supporting this, like, middle, like. Like, almost like river that's flowing and it's flowing fast. So we can control the. Where it flows to, but we really want to ship product really quick. So I think my advice to a lot of entrepreneurs in the early days, like, the most important thing we could be doing as an organization is focusing on product in shipping product, not working on pr, not working in some ways, like getting the brand perfect or getting the right article in TechCrunch. And it's really about, you know, getting a useful product or service out the door. And so I think my overwhelming advice for everybody is get out there and ship product, and that'll be, like, the biggest, the most important thing you could be doing as a founder or entrepreneur.

Yeah, it's hardware walks. Bullshit. Hardware talks. Bullshit walks. Right? It's. The ratio of something to nothing is infinite.

Got to get out there and build it, right?

Yeah, for sure. Brett, where do people find you on social media? Tell us where they can learn more about you as an entrepreneur and CEO.

Yeah. Personally, I'm basically trying to build in public with Figure in my life as much as possible on Twitter, so you can find me there.

And then your handle at Twitter is what.

It's my. It's Adcock underscore, Brett.

Okay, perfect. Thank you.

Yep. And then professionally, like, you know, figure AI is the website, and then my last company, archer.com are good ways to kind of like, get, you know, like, a better understanding of the companies I've built. And. Yeah, and if you're looking to try to make an impact here, I'd really, like, put a plug out to apply to come to figure. We're trying to look for the best and brightest and hire the best talent in the world. It's always a top priority for us.

Yeah. And you've done that again. Congratulations. So blown away by the team you put and really excited to come and play and see and touch and maybe we can do this podcast again from your facility with one of your robots playing in the background next time we're ready.

So let's get you out here and see some robots.

Awesome. Everybody. Brett Adcock. Figure. Thank you, buddy. Pleasure to have you on this.

Thanks for having me.