Ad for Sunday's dog food: new fish recipe for picky dogs, human-grade, vet developed. Use code SRS50 for 50% off.
Ad for Chime: fee-free banking with rewards, 5% cash back, 3.75% APY, travel perks. Use code SRS at chime.com.
Brett Eggcock, welcome to the show.
I've been looking forward to this for a long time. The robotics guy.
Let me give you an intro. Brett Adcock, serial entrepreneur and founder of Figure AI, building general-purpose humanoid robots. Founded Vetery, acquired for $100M. Co-founded Archer Aviation, launched HARK. Time's 100 most influential in AI 2024. Married, father of three. And before we get into it, we always start off with a gift.
He didn't give you any tips on that, did he? All right, I got to hit one. What do you think? You can leave this guy here. I got you another gift. Little robot. Mhm. Yeah, no problem.
They're great. This guy's staying. That is awesome. No way. That is awesome. Thank you. Very cool.
Well, Brett, we got a lot to talk about. How many companies are you running now?
I'm not sleeping. I got too many. That thing is amazing. I think it's incredible. I can't wait to talk more about it.
Patreon question from Steven Casey: What safeguards are in place to prevent harm from AI errors in robotics?
We don't want the terminator. Four years ago there was no path for humanoids. Now we have electric humanoids with neural nets. We need to make it work at scale safely. We have intrinsic safety and semantic safety. We've had robots in my house testing for a year. Kids name them and love them. Eventually every human will have a humanoid like a phone.
They like it, it's normal. They talk to it and want to jump on it. We're not there yet but will be in several years.
What's the longest they've been around a robot?
A couple of months on and off. They weren't always around but often. They name it and love it.
What's the emotional attachment?
My kids wanted it, they don't want to get rid of it even if it's banged up.
Honestly, in our lifetime, every human will have a humanoid like a phone and car.
The complexity of the problem, like knocking over a pot of water.
It's a problem fun house: hardware, AI, scaling, manufacturing.
Some people say AI is in a bubble. Poly Market says 18% chance it bursts by 2026. Is AI in a bubble?
Absolutely not. We'll see the most transformative events in the next 36 months. We're scratching the surface. We need millions of robots to make an impact. We'll have AI that can use computers and machines, leading to an age of abundance with trillions of synthetic humans.
What will humans be doing?
I don't want to do busy work. I want to delegate everything to my AI operating system. In 24 months we'll have all that stuff automated. Just the leg work. I want to spend time with my kids and do what I love.
You're basically exporting your brain to robots.
Delegate it all out. We'll have it so good you won't order food or book stuff. Physical chores too. It's a compression algorithm. We have AI at HARK that can use computers and make phone calls.
I've never thought of it that way. I always feared AI taking over.
It's a compression algorithm. We'll delegate all busy work to robots. In 24 months you'll have a humanoid at home that knows you.
Ad for SpotOn: GPS fence for dogs, Nova Collar, 99.3% containment. Use code SRS for $50 off.
Well, I would like to do a life story. Where'd you grow up?
Yeah, let's do it.
Central Illinois. Small town of 700 people called Mikua. We grew up on a farm, third-generation farmers.
That's even smaller than where I grew up (Chilikothy, Missouri, 8,000). What were you into?
Sports, computers early. Started startups in high school and college selling things online. Didn't grow up with money, so the internet was a way to make money. I loved creating things.
Do you have brothers and sisters?
I have a brother, Colby. He runs an AI defense company called Scout. We live a block away from each other.
What do your parents think, from farmer to this?
They drilled into us to run our own business. Farming is entrepreneurial. They're proud but confused. I've been doing crazy stuff for over a decade.
Were you building stuff as a kid?
Constantly. On the farm, software, internet. I always wanted to work on hardware I could touch.
Where'd you go to school?
University of Florida. After, moved to New York and worked on software startups. During college I did side internet things.
After college, I founded Vetery, an AI recruiting marketplace. The hiring process was broken. We matched talent and companies at scale without humans. Sold to the world's largest recruiting company for $110M.
You didn't even have it for sale?
No. We were doing 20-30 thousand interview requests a week. They approached and bought it. I was ready for something bigger.
I wanted to work on flying cars or school security. I chose flying cars. In 2018 I started Archer Aviation. Electric aircraft that take off vertically can fly cheap and safe. I didn't know how to build one, so I hit the books and took courses.
Where do you get the confidence?
I learned. I studied electrification, vertical lift, and winged aircraft. I built an algorithm for electric aircraft sizing. With electrification you can have many small rotors for redundancy. I was at a Hyatt in Atlanta at a design course and met a PhD student from UF.
From University of Florida. I was like, 'Oh, I went to school there as well.' And he said, 'What are you doing here?' I said, 'I'm starting a company to do this and I need to figure out how to build these things.' He said, 'My professor runs a small drone lab with a full building and 12 PhDs. Why don't you come down and meet him?' I flew down that weekend and met his professor who ran mechanical engineering and aerospace. Long story short, I ended up taking over his lab and we started building aircraft in 2018 and 2019 at the University of Florida. I temporarily moved down there with my daughter and wife. We built a new lab off Archer Road and called the business Archer Aviation. I spent the next year and a half modeling and building electric vertical takeoff and landing aircraft.
Yeah, it was a great time in my life. But there was no intersection of people who knew electric, rotocraft, or airplanes — no Venn diagram of overlap.
So there was nobody in the world who understood how to make all that work. I had to learn it from first principles, then moved the company to California a few years in. Things took off. We built bigger aircraft, took the company public within three years, and it's now a six billion dollar publicly traded company. We've designed four or five generations of aircraft at Archer. It was hard, but it set me up well for Figure and other stuff. Going public was one of the hardest experiences of my life.
We went public through a SPAC process, which was all the rage four or five years ago.
It was a special purpose acquisition company, basically a reverse merger. It was hard because in 2018 and 2019, I had never done hardware before. Raising capital was tough, and nobody was funding deep tech eVTOL companies. The big VC groups weren't funding SpaceX, Tesla, or Rivian back then.
Oh [__], are they always behind like that? The mandate for most VCs in the Bay Area is not to do hardware.
Gotcha.
They don't do deep tech like rockets and autonomous vehicles. I don't think there's a single top VC in the US that's invested in a humanoid company as of six months ago. So I ended up going all in on Archer through the IPO. I put all my money into it — I bought a house and the rest went into the company. It was a stressful period. We went public through a SPAC because we couldn't raise enough privately. We had a choice: raise $100 million at a $500 million valuation or go public and raise a billion dollars. We went public and raised a billion dollars.
Wow. You've got a huge appetite for risk, huh?
And we got sued during it.
Yeah, we got sued by Boeing and a big startup founded by Larry Page. I woke up to a front page New York Times article about it.
That's got to be intimidating.
Yeah, it was crazy. The backstory is Larry Page started a company called Kittyhawk about 10 years ago. They did great work on electric VTOL aircraft. I ended up taking the core team of 10 to 15 people from there over to Archer within the first two years. They retaliated by trying to harass us while we were going public. But we got public, had a billion dollars on the balance sheet, and started building aircraft and the service — the app, real estate for vertiports, and the engineering work. It's basically a flying robot with battery systems, electric motors, sensors, embedded software, and control systems. Similar to what you saw this morning — it's a 6,000 lb, four-passenger piloted robot with 24 degrees of freedom, tilting motors for vertical takeoff, and variable pitch propellers. A highly overactuated system that needs really good software.
What altitude is it flying?
About 2,000 to 3,000 feet above ground level, similar to traditional helicopters.
I think a lot about Tesla and EV vehicles. The government seems far behind on AI and gridlock. I've always wondered why we aren't going full EV. With AI processing traffic, a lot of that would go away even without air vehicles.
Yeah. But it's also hard because there are a billion and a half installed cars in the world, and we make about 80 million per year. It takes about 20 years to replace all the cars. And you can't retrofit all of them with autonomous hardware. We're in this transitory state for the next five years or so, where everything will become fully autonomous and trusted. It's a weird period, but soon it will be native and safe.
How long do you think it'll be?
In the Bay Area, you can take Waymos everywhere now. They're all over — in Palo Alto, Menlo Park, San Jose. I take Waymo to dinner on weekends. It's awesome. The car drives human-like and it's great not having a human there. Technology is in early chapters, but it's all here. It will scale everywhere with time — we need to mature the technology and install base. It won't happen in a year or two; it'll take time.
When are we going to see your aircraft?
We have them now. We fly every week in California. The challenge is FAA certification to fly passengers and charge money. We need a type certification, and the FAA moves slowly. The safety standard is one in a billion hours reliability. It's the highest bar, but that's what you need for flying passengers over cities. That's the long pole for us. We're in the middle of FAA certification now, hoping to get it as soon as possible. We're also working with other global entities.
What do you envision 20 or 30 years from now? Do we still have roads? What does the sky look like?
The airspace is three-dimensional, unlike roads. Cities are built around 2D roads and there's no room to expand. Humanity is moving to cities — half the world lives in cities now, and it'll be 70% by 2050. Roads can't grow, but the air can. You can stack an infinite number of routes at different altitudes. The future of travel in cities is below ground in tunnels and above ground in the sky. With vertiports at specific locations, you'll take a Waymo or bike to a node, then fly to any other node on the network. It's exponentially more routes than tunnels. For trips over 20 minutes, everything will move to the sky. You'll be able to live far outside cities and get downtown in under 30 minutes, flying at 150 mph with no traffic.
Will these be personally owned or an Uber service?
It'll be like an Uber service. You'll pull up an app, pay per trip — maybe $40 to go downtown in under 30 minutes. On-demand, you'll get in and fly down.
This will be in every neighborhood, very accessible?
Yeah. Electrification reduces cost and safety burden. A helicopter has hundreds of safety-critical parts; an electric aircraft has none. It's much safer, quieter, and cheaper. I haven't flown in ours yet, but our test pilots love it. Everything's going electric and it's crazy that it works. We're in the right decade.
You're really changing the world.
We're the start of it. Hopeful to make it work.
Let's do it. How did this idea start?
I spent five or six years on robotics at Archer. The meta problem in robotics is building a general-purpose machine to do everything humans can. We built the world around us for our human form factor. The holy grail is a humanoid robot with legs, arms, hands, and sensors. After working on aircraft, I thought it was possible to build an electric humanoid. Electric is important for cost, safety, and performance. At the time, Boston Dynamics Atlas was hydraulic, heavy, and ran for only 20 minutes. We needed to radically transform the hardware and figure out an AI brain. The robot has about 40 degrees of freedom, meaning more states than atoms in the universe. You can't code that; you need neural networks. Our controller runs 200 times a second for balance. Every joint has position encoders and force sensing. It's extremely complex.
I was going to say, it's updating the foot 200 times a second.
Yes, our main computer processes what to tell all the joints 200 times a second to balance and do tasks. If it's too slow, we fall over. It's dynamic — if you power it off mid-run, it falls. Every joint movement affects the whole body.
Starting something new isn't just hard, it's terrifying. Before I launched the Shawn Ryan Show, those what ifs were loud. What if nobody listens? What if this fails? Walking away from what's familiar to build your own thing takes real faith, but it ended up being one of the best decisions I've ever made. Whether you're starting a podcast or launching a store, it helps to have a partner like Shopify on your side to help ease those worries with their expertise and tools. Shopify is the commerce platform behind millions of businesses around the world and 10% of all ecommerce in the US. From household names like Figs to brands that are just getting started today, one of the biggest hurdles is efficiency, and Shopify helps you accelerate that. Whether you're uploading new products or trying to improve existing ones, the platform is packed with helpful AI tools that can write your product descriptions, page headlines, and even enhance your product photography for you. They also allow you to tackle all your important tasks in one place. From inventory to payments to analytics, there's no need to juggle multiple websites. Everything is in one spot to make your operations smoother. And you've definitely seen that iconic purple shop pay button used by millions because Shopify has the best converting checkout on the planet. It helps boost your conversions, which means fewer abandoned carts and more sales for your business. It's time to turn those what ifs into with Shopify today. Sign up for your $1 per month trial at shopify.com/srs. Go to shopify.com/srs. That's shopify.comrs.
So getting back to your original thing. My bet, three and a half years ago when I started Figure, was that this was possible now. I believe over the next 10 to 20 years it will work. I started from the ground up building humanoid robots and AI software. At the time, there was no precedent — no AI had ever worked on a humanoid robot, no electric humanoid hardware was good, and there were no hands. I had trouble getting people excited. I self-funded the first year, got to a million dollars a month burn by month four. We built a 40-person team quickly. I knew how to spin up hardware and software from Archer — electric motors, battery systems, control software, sensors. We built actuators with rotors, stators, gearboxes, sensors, electronics, wiring, and firmware. Motors are designed to run at high RPMs for efficiency, but humanoids require holding forces at standstill, which is hard. We have whole teams working on rotor design, electromagnetics, stator design, gearbox design, sensor design, and motor control. It was an enormous lift. Now we've raised about two billion dollars. We built Figure One, walking in under 12 months — fastest in history. Then Figure Two, which is the second generation robot. In 2023, we did a demo where Figure One made coffee using neural networks — no code. That was a breakthrough. We realized it's all about data to train neural networks on good hardware.
It sounds simple, but give it to a four-year-old. The dexterity on the hands to hand you the bag of gummy bears.
Yeah, exactly.
Can you do it with neural nets on board? Can you take camera pixels and output motor trajectories through a neural network with no code? We did that in 2023 on Figure One. It was one of the first examples in the world. That was game on. Let's build really good humanoid hardware, make it cheap and reliable, and go all in on neural networks. We launched Helix, our neural network stack. Now we're on Figure Three, the best humanoid hardware by far. We have robots unloading dishes and folding laundry. Figure Twos worked at BMW for six months, every single day, with 10-hour shifts. They were building X3s on the line starting in January 2025. I bought the first four — one is at my house. BMW is the best robotics company; they have giant manufacturing robots everywhere. Cars are so big and heavy that you can't move them by hand. The feedback from BMW is great.
Building the car and moving it. The body shop line is automated end to end but with humans. The car is built by machines. There are machines everywhere with special end effectors that switch in seconds, some the size of a chair.
One or two seconds, really fast. They're building a car with robots everywhere. BMW has been privileged to see this level of automation. These machines make what we do look like toys.
We're doing some very complicated things, but car manufacturing is no joke.
What specifically were the bigger figures doing?
We had a body shop line building the rear header. The robot put sheet metal on a fixture, repeated for 10-hour shifts. The robot arm spot welds it. The goal was to see if we could run robots every day. It was a learning experience. The same robot lasted six months. The worry was humanoids couldn't last a week, but from hardware it was A+. From software we did B+ because of architecture decisions. Half the stack used coded controllers for walking, the rest neural nets. But we had scaling issues. So we refactored everything into a neural net. We removed over 100,000 lines of code. Now the robot runs on Helix 2, all neural net. We run 24/7 shifts without faults for days.
It was cool that the same robot lasted six months. From hardware it was A+. From software, B+ due to architecture. We had a hard time scaling. So we switched to all neural net. Now robots run 24/7.
What kind of stuff do they do at your office?
Logistics moving packages, greeter bots walking around 24/7. It's a stress test because the neural nets are new. We run them all the time.
What's a conversation you've had with a robot?
We work on deep memory so the robot knows who you are and has permissions. You can command it, but it needs authentication like facial recognition.
What about permissioning systems?
For example, kids asking for ice cream. We need voice and facial recognition. The robot should have conversations and do tasks. Language is the key UI.
