Optimus, the Tesla bot, has been in development for 2 years and still can’t really do that much. There’s another dream team of engineers and roboticists who are building a humanoid robot that they call the “kinder, gentler” robot.
In this TechFirst we chat with 2 investors in Figure.ai: Jesse Coors and Gregg Hill.
They think humanoid bipedal robots could be one-to-one with humans in numbers “pretty quickly.” Corporations will want thousands, and most of us will want at least one.
We also dive deep in the Figure.ai robot as I ask them these questions:
- Capability?
- Timeline to usefulness?
- Battery life?
- Cost?
- Where does a humanoid robot fit in the world?
- What does it do to our economy?
- Where do you see it helping most?
Read my story on Forbes, or keep scrolling to watch/listen and skim the transcript …
(subscribe to my YouTube channel)
Subscribe to the audio version of the TechFirst podcast
And … pick your favorite podcasting platform:
Apple Podcasts
Spotify
Google Podcasts
Overcast
Other podcasting platforms
Transcript >> Humanoid robots: ‘we could be one-to-one with humanoids quickly’
(Usual caveat … this is AI-generated. Watch or listen for a higher level of accuracy.)
John Koetsier:
Will we have useful humanoid robots soon? And if so, will they not be from Tesla? Hello and welcome to TechFirst. My name is John Koetsier.
We all know Optimus, the Tesla bot that has been in development for a couple of years, still can’t really do all that much, but there’s been some impressive progress. Well, there’s another dream team of engineers and roboticists who are building another humanoid robot. And they say it’s the kinder, gentler version. We’re going to talk to two investors in the company right now.
One is Jesse Coors. The other is Greg Hill. They’re VCs that invest in innovative products that are powered by deep tech and open up new spaces. Welcome Jesse and Greg.
Jesse Coors-Blankenship:
Thanks for having us!
Gregg Hill:
Thanks!
John Koetsier:
It’s a pleasure to have you. Maybe let’s start here. We’re gonna get into the bot. We’re gonna get into what it can do or what it’s gonna be spec’d to do, timetables to something useful, all that stuff. But let’s start here. Why did you invest in Figure?
Jesse Coors-Blankenship:
That’s a great question. Humanoid robots have been… in research for a very long time, maybe about 20 years. I think we all can remember Honda’s Asimo, but it’s really the convergence of technologies that were not available then that have come out of labs that make a robot, a generally useful humanoid robot, practical and really able to go into markets now. Like AI control systems, lightweight battery technology, electric actuators, all those things back when those experiments were happening.
John Koetsier:
Right, right.
And Asimo of course reminds me of Asimov and the three laws of robotics and all that stuff. But any other reasons, Greg?
Gregg Hill:
Yeah, I mean, honestly, this is a dream market, you know, for VCs.
I mean, I think maybe you saw the recent interview with Elon Musk, where he’s speaking to his investor base, where he said the majority of long term value for Tesla is going to be optimist. I think that’s like one of the most powerful statements I’ve ever heard in VC is essentially like, think about Apple coming out and saying, the majority of his long term value is not going to be iPhones, right … that’s a $4 trillion company.
John Koetsier:
It’s going to be the Apple Vision Pro!
Gregg Hill:
Yeah. It’s just amazing and the fact that just Parkway, we were able to lead a round in Figure, arguably the best robotics team ever put together in the largest market in the world. So we’re just super excited about it and the science is done. So it’s just a matter of getting the road about up and working and simulating and getting it into a 3PL and further.
John Koetsier:
Interesting, interesting.
Always have to take Elon Musk’s statements with a grain of salt, because he’s absolutely right.
If you can get a working humanoid robotic platform functional and useful, I mean, that’s off the charts, useful for the economy, off the charts, market space, depending on cost, obviously everything like that.
But of course there’s a ton of things that Elon Musk says that Tesla is gonna do, and it doesn’t quite do or it takes a lot longer to do than he actually thinks it will. Let’s talk about this robot first. What is this robot? What’s it give us some idea of the specs, capability, and its design parameters.
Jesse Coors-Blankenship:
Yes, happy to tackle that. So Figure One from Figure AI is their first robot. It is a humanoid robot, which is an electric power, electric actuator robot, which is really built to be slightly less powerful than an average human and meant to do physical labor tasks like moving packages and boxes in a warehouse. And that’s really where we’re starting out.
Over time, we would imagine the capabilities of figure to be able to do more conditions of work. But we’re starting out in a very limited set right now in warehouses for fulfillment centers. Typically a job you’d call picking, moving one box from one bin to another or a package from here or there.
It’s something right now that… There’s a lot of turnover in the market with the folks that do that work right now, and there’s a very big labor shortage. So we’re trying to fill that gap.
John Koetsier:
Yeah, in some markets where Amazon has its distribution centers, they’ve turned through 130%, 150% of the local workforce and they can’t keep people in.
Of course, there’s all kinds of issues of the demands that they place on people there and the injury rates and all the things that are going on there. I actually know a person who was retired and just for a few bucks wanted to work in something, someone not in an Amazon delivery place, but he just couldn’t handle it … the demands were insane.
So that’s an interesting place to start, but there’s a lot of non-humanoid robots that are going and being built for those sorts of things. Why humanoid?
Jesse Coors-Blankenship:
I like this question because it’s come up a lot with the folks that invest with us. I think we certainly have very good robots right now in factories and warehouses that do specific tasks for automation. Those tasks are, there’s a number of them, but they usually are only good at one thing. That’s a very limited market.
For us, what’s so powerful about a humanoid robot is that our whole environment is so is built around our form factor, the way we are as humans. So if we have something that’s really close to approximating us, it can go wherever we go.
But it can actually do a bit more in more hazardous conditions as well. Many facilities are very cold or they’re very hot or they’re unsafe, but the work that we do is in physical labor is all shaped around us, like our form factor.
Gregg Hill:
Yeah, even the case of Amazon centers, I mean, they try to fully automate, and it just didn’t work. They still need humans. So what’s the closest thing to a human? It’s a humanoid.
And a humanoid can work 24 hours a day, a few charges like a Tesla car, work weekends, doesn’t get sick, right? And in the hazardous environments, it’s very… which I would say sometimes you could consider that hazardous environment. It’s really jobs that people don’t necessarily want to do … mundane tasks, not exactly a workforce coming up through the generations.
So if you think about five, 10 years from now, we’re going to have a 20 million job shortage. And how are you going to fill that? I mean, commerce has to continue on, and it can actually cripple our economy if we don’t have a humanoid.
So this is not really just showing up just to take people’s job is actually coming to, to kind of help, you know, help, uh, manufacturing jobs in the U S specifically.
John Koetsier:
You bring up an interesting point because you said nobody, you know, well, there’s going to be the shortage, but nobody in school is like, I want to be working in a warehouse. I want to pick product and pack it, right? Everybody wants to be an influencer or something like that. And hopefully we have engineers.
Gregg Hill:
Yeah, I think there’s some cooler jobs coming.
Imagine like understanding how to, you know, the complex [construction] of a humanoid robot. There’s a thousand parts, you know, that’s really interesting growing up and going to school and learning about this robot and how it how it functions and the technology.
That’s more exciting for the for the next generation, not working at a factory and picking 200 times and getting alerted and not even having a manager to talk to, because it’s just all electronic. pick this box up, take it over here. That’s a tough job to fill. And there’s not really enough money you can pay to keep people coming back. It’s more of a lifestyle.
John Koetsier:
It’s really interesting that you said not even having a human boss because there have been people allegedly fired by AI as well, fired by a smart system or an algorithm, right?
And honestly, we want jobs for everybody who wants to work and for that hopefully that’s most people, but we don’t, there are things that humans should not do, the repetitive stress type things, low skill type things, dead end jobs, there are things that should be replaced by robots.
We’ll get into some of the economic and financial and career job type implications in a bit, but let’s dive a little deeper into capability here. So it’s humanoid, we know it looks like a human. How many degrees of freedom? What are its hands like? How complex is the jointing there? Does it have a replaceable battery pack? Run down what you think the first version will look like.
Jesse Coors-Blankenship:
So the battery pack is chargeable, but not meant to be replaced on a day-to-day basis. It actually causes more problems than it solves to do that. So they will charge. I think the company is saying right now that they expect it to run for five hours before needing a recharge. But we actually imagine it could be quite a bit longer.
John Koetsier:
Interesting.
Jesse Coors-Blankenship:
The hands in particular are where a lot of the IP has gone into. And it’s, you could obviously make a robot with, you know, like claspers, but the reason we put so much into the hands is because there’s all of these different shapes and sizes of packages. And, and then as the robot moves out into the world more and more, you know, all types of different doors and door handles and, you know, the work to be done is really specific to having hands like ours.
Doesn’t have the same degrees of freedoms we do, but it’s very close.
Gregg Hill:
Yeah, I mean, you can pick up a glass of water with a few fingers, kind of like us, you know? That’s pretty amazing. So it’s not going to be as clunky as you think. It’s got a great gait. So when it walks, it actually moves like us with that freedom.
Not as good as us, but pretty impressive. It can pivot at the ankle. It can fall down and not break and get back up. Yeah, it’s pretty amazing how far we’ve come. you know, and with hardware and now hardware and software coming together.
John Koetsier:
When it falls does it fall like a log, or does it lift up its hands and brace?
Is it smart enough and fast enough for that?
Jesse Coors-Blankenship:
Yeah, yeah. So there’s a lot of, you know, work that goes into stochastics, you know, to, to fall gracefully, stuff like that. Obviously, everything we’re doing right now, it’s hardest anyway, but it’s trained to be able to, you know, adapt to its environment like that.
But what we really like about the warehouse specifically is that it’s not as complicated as to say, oh, we’re going into grocery stores, or we’re going into retail. It’s literally a fairly simple environment. It’s almost like self-driving cars, right? They’re not very good on the city streets yet. There’s just too many conditions that they have to be trained on and be aware of. Whereas in a warehouse, or I’m sorry, driving is actually really functioning well.
Gregg Hill:
It’s got a lot better in the street, but you’re right. It’s like, yeah, it’s got a lot better over the last 10 years. But Tesla, they went on the freeway first, right? And that’s a lot more predictable. And then they kept updating the software. And then Waymo went to the streets. And that was, yeah, it’s been quite a process, a lot of work arounds, right?
John Koetsier:
It’s pretty interesting with Waymo. I was in SF just recently and it seemed like there was a Waymo car on the street every time I stepped out of a building. They’re all over the place there.
Gregg Hill:
Yeah, and if you think, I would try to like look at human or robots and then also Waymo cars and taking your kid, your daughter or son, 11 year old to school, you know, just car driving it to school 15 years ago, imagine the human psyche saying there’s just no way. This is crazy.
Not going to happen. And now it’s just kind of… you know, something that’s sort of the norm. And I think that’s where humanoid robots are going. One day it’s science fiction. This is not going to be in our house. It’s not going to be doing task force. And then the next day, you know, kids are sending tweets. Did you see that cute bot?
You know, I just went over to Sarah’s house and they have the emojis, you know. And that’s really where we see the future going over the next decade.
John Koetsier:
Talking about that, there’s a viral video going around right now. Maybe I’m kind of usually last to see these, so maybe it’s been around for some time. But it’s Waymo and it’s two really old dudes who go on there. They look like old retirees in their 70s or something like that. And they go in and they just all they know is there’s a cab coming and it comes up and then it picks them up. Where’s the driver? Where’s the driver? And one is like terrified.
Jesse Coors-Blankenship:
It’s like Ghostbusters.
John Koetsier:
They go through it. Really cool. If you haven’t seen it, go check it out. It is very cool.
Let’s get back to the robot here. A really critical piece is learning slash teaching. And so the warehouse is a great place to start. Like you said, it’s a controlled environment. It’s got, there can be chaos in a warehouse, but it’s within certain parameters. And there aren’t, you know, 50 shoppers in front on the, on the way to the oranges, right? In the produce aisle.
So there’s greater simplicity, but the learning platform is really key because when you develop a general purpose humanoid robot, you want to use it for general purposes. You want to give it to old older people who can’t maybe maintain their house as much. You want to give it to somebody who’s, uh, super busy and they want something that will clean their home and maintain it, maybe even cook, maybe even bake, who knows what.
You want ones that can go out and be gardeners or work on a farm as a laborer or other things like that. Talk about the learning platform that Figure is building here.
Jesse Coors-Blankenship:
Yeah, so it’s… It’s largely proprietary. There is certain simulation, which is standard across all robotics. But I think that they did a great job at getting their product, you know, fully designed very quickly so that they could start simulating it and, you know, being able to simulate different gates in a virtual environment and optimize it there. Translate over to, you know, the build of the robot when they built up the first five robots. And that’s when. the next layer of training takes place is actually training in, you know, with the base of simulated training then go into real-world training. So it’s essentially like, you know, starting in imagination and bringing to reality is the way I look at it.
John Koetsier:
Yeah, there’s a lot of synthetic data being used for training environments right now and it makes a ton of sense. You got to think that something like an LLM or a GPT-4 could be super useful for something like this in terms of understanding the world and understanding things maybe to say as well. Thoughts about that?
Jesse Coors-Blankenship:
Yeah, the company is actively exploring that, but it’s pretty sensitive in terms of how they’re doing it.
But that’s, we’re really excited about, yeah, what’s coming there is what we can say.
John Koetsier:
So let’s talk about this on a broader sense then. You’ve said that you invest in products that are super innovative, that have deep tech, not just simple stuff. So that’s gonna have a longer time scale to go to market and to return on investment and other stuff like that. and they open up new spaces.
Well, there’s so much that you can imagine for a humanoid robot. What do you imagine the effects are on economies, global economy, but also regional, like richer countries, poorer countries, all that stuff. How do you imagine that’ll work out?
Jesse Coors-Blankenship:
Well, it’s gonna have a profound impact in many, many ways.
You can really just, you just have to use your imagination, but imagine a world where we could be one-to-one with humanoids pretty quickly, and maybe even more so, because some people are gonna want more than one, and… many corporations are going to want thousands.
So what does it mean to have means of productivity in physical labor where you have something that is. How do you put it? Doesn’t get tired, doesn’t have the same needs as people. So what does that do for production? It’s a complete game changer for productivity and our means of production.
Gregg Hill:
It could help bring manufacturing back to the U.S. potentially.
John Koetsier:
It’s interesting if you think about it in terms of, or make a kind of a correlation, you think about information and data and digital products and the arc that they’ve gone on as we’ve gone through the internet era, the digital era. And information wants to be free, right? The marginal cost of reproducing something approaches zero, right?
Does the marginal cost of physical labor approach zero?
Right? And if so, how does that work and how does that happen? Do we tax robots like Bill Gates has suggested so that we can provide universal basic income? Does it open up so many new possibilities for environmental remediation or other things like that are just too, in many cases, too expensive to throw people at right now?
The mind kind of boggles.
Jesse Coors-Blankenship:
Yeah, I think it’s all on the table. I think I just try to think about it simply. Like if I had one just in my own life, what does it mean for me to save time? It’s just imagine that all the little things that take up my day that I have to maintain, I didn’t have to. Like how does that free me up?
So it’s not just like leisure, but my own work. You know, like how much more time would I have to do the things that really propel innovation or whatever type of work you’re doing. So I like to think about the scalability of that. So I’m not worried about taxes, I’m not worried about… I look at it more like, what does it mean to have, as a society, just so much more time? You know, potentially. That’s a profound thing.
John Koetsier:
Potentially, potentially. It’s really interesting to look at it from another angle through a different lens. One of the things that we’ve seen AI doing is work that otherwise would not be done.
So for instance, this is literally from like 2016 or somewhere in the dim mists of AI time, right? The early epoch of AI, where I think it was Reuters or AP started. publishing AI driven articles of high school baseball games or small college baseball games.
And that was work that just wouldn’t have happened because it wasn’t a big enough audience, there was an audience of dozens or hundreds, not an audience of tens of thousands to make it worthwhile for a human to do. What does this open up in terms of work that’s just not being done right now, as well as, as you mentioned, increased leisure time if you can afford one?
Jesse Coors-Blankenship:
Yeah.
Gregg Hill:
Yeah, that’s a great question. Well, I’ll go further with AI, separate from the robots, just tell you how exciting we are about AI. AI is going to help third world countries. And one good, we actually have one of our companies, I want to throw this in as a fusion company, most powerful one or at least the one that’s raised the most money.
John Koetsier:
Interesting, hopefully not cold fusion.
Gregg Hill:
Yes, you remember The Saint, the movie at the end. We’re getting there.
But if you think about like a content like Africa, one of the biggest issues with Africa is infrastructure and getting like, you know, clean water and food, you know, there is a really huge area and it’s just too complex. And even the Gates Foundation, they really had a hard time because it’s just the infrastructure and it just, it would cost trillions and decades. And it’s just, it’s almost not possible to fix, honestly.
Think about fusion reactor being able to drop one in into Africa and light up miles in pure water without the infrastructure other than just a power supply.
And the reason why I say that is AI is the only reason why fusion is going to actually be realized. It’s because the reactor is so complex, AI, Google actually, AI is actually controlling that reactor, which is one of the most complex. piece of equipment in the world.
So you think about AI, what it’s doing over the next decade, we can really change the game in Africa. That’s the good side of it.
John Koetsier:
Turn off the reactor, Hal.
Gregg Hill:
Yeah, it won’t, it’s not gonna go, it’s not gonna explode or anything. But so we like that side and we love the fact that these are just completely new markets. I mean, if you think about it, IBM and all these companies, they’ve been around a long time. And now you have Sandbox, AQ and Figure and Optimus and TAE, and then these companies and these are completely new markets, completely new cells that the world’s never seen before.
That’s an exciting time to be a VC and to be investing in this fourth wave. So that’s what keeps us up at night.
John Koetsier:
It’s a very exciting time to be alive with the capabilities that we are reaching for. And I can only hope that we have the wisdom as a society as well to use these innovations well and hope that the benefits from them accrue to, well, VCs deserve some and inventors deserve some, but hopefully a broader spectrum of society as well. Super interesting, give us some timetables here.
What do you think is a timetable for figure one to do useful work? I’m not talking about a demo. I’m not talking about Spot, or Boston Dynamics doing a cool dance, which is neat and awesome. It’s amazing. It’s wonderful. But what’s the timeline to actually say, hey, you know what? We’re renting this guy out, SaaS model for warehouse labor, or we’re selling the dude or whatever the case is.
What’s that look like?
Jesse Coors-Blankenship:
Well, it’s very close. We’re looking at 2024.
John Koetsier:
Wow.
Gregg Hill:
Yes, very close.
John Koetsier:
2024, wow. I was expecting something like, well, we’re hoping to get something out there in 2025 or 2027 or something like that. You’re talking literally next year. So you’re talking maybe six, nine months or something like that, 12 months to get something that’s actually functioning in a customer’s warehouse.
Jesse Coors-Blankenship:
Yeah, selectively, right. I think the first handful of groups using it are going to be very close to it. And going out more broadly from there, more scale, it’ll be a little further out.
Gregg Hill:
Yeah,Figure’s built a factory within their office, basically simulating so we can do one person’s job throughout an entire day. And a really key executive at Geodes is there making sure that every aspect of that person’s job is done perfectly. So by the time it hits the floor, there’s going to be no guessing. It’s going to be ready to go. Really important.
John Koetsier:
Wow.
John Koetsier:
So what’s great about Brett Adcock, the founder, He’s a realist, he’s got a big vision, but he also understands what’s achievable. And once it hits the floor, once it’s ready, we’re gonna be, we think we’re gonna get there first. And that’s what’s really exciting.
John Koetsier:
Super interesting. Now you mentioned before you think it’s the best team building robots on the planet. Maybe you mentioned a couple names, maybe mention, a little bit about who they are, what they’ve done briefly.
Jesse Coors-Blankenship:
Well, the… are all very key in the industry. And there’s about 50 people now.
Brett Adcock is the founder and CEO, and this is his third venture. A few folks from his previous ventures who he’s worked with over the years have joined him. Brett’s a tremendous leader and a very successful entrepreneur, but also somebody who’s successfully built one of the world’s premier evital, electric vertical takeoff and land aircraft.
That was Archer Aviation and successfully brought it to market before FAA approval was ready from a business perspective. So the R&D in three years to be able to do that, you know, the guy’s a powerhouse.
But he’s also joined by CTO Jerry Pratt. Jerry Pratt was one of the leaders of the DARPA Humanoid Robotics Challenge and his robot, his humanoid was on the cover of Time Magazine. respected in the humanoid robotics space. And then many of the engineers came over from robotics companies that were specialized in this space in all the areas of IP that they were really focusing in on. So hands, actuators, control systems, everything you need to be able to successfully do this.
Gregg Hill:
Yeah, the way, yeah, Brett essentially looks at hiring like a football team, like he loves football. So Brett’s like hiring getting the quarterback, like the star quarterback, like a Tom Brady or somebody, somebody else. And then who’s the best, right receiver could be hands, who’s the best, you know, you know, AI and put together that team. And that’s what he did at Archer.
And that’s how he was able to, to do that in three years and achieve such, An amazing feat. I mean, it was the first ever to do that. And he hired every single person personally, and he did that as well. He sent out like a thousand emails to everybody and said, this is what I’m doing. Talked to every single person, brought them in. And now they’re just, you know, Balsam Dynamics, Google X. He’s got all the rock stars on the team. And it’s just really hard to create a team that can actually compete in this industry.
The entry is super hot. you got to be able to have decades of experience. You have to be able to raise a lot of money. So you have to have a super founder like Elon or Brett. So it’s not like there’s going to be a lot of competition, honestly, in this space. There’s a thousand parts to the human or robot. It’s very complicated. So there’s going to be a few players, and we love it. Few players and a huge market. The question is, obviously, can we do it?
And once we get past that gate and we can prove to the world or… the team can prove to the world that a humanoid can do one person’s job in a warehouse or a factory and can be dependable day in and day out. It’s not going to break down. The company can rely on it. At that point, it’s almost like when the first Tesla cars were coming out and hitting the road and you just can imagine what’s going to happen in the market. It’s just a ton of money is going to flow in. and then everyone’s getting really excited.
And then at that point, we’ll just keep updating the software and going into retail and like you said, home care. Who knows education. I mean …
John Koetsier:
Healthcare.
Gregg Hill:
… a humanoid teaching, it doesn’t have to be, it could be high schools, it could be colleges, it could be at home. There’s just so many applications. I mean, just a humanoid to sit there and make sure the porch pirates will take your packages all day. It’s pretty important. I mean, I don’t know. I’m just thinking these things that you think of, right? Or like right now I’m out of town and I’m worried about my packages in my hometown. And I would love to have my humor there alerting me.
So I just see there’s so much potential. We just got to get through that first gate. And once we get through the first gate, we’re all going to gain confidence as in the industry, Elon, figure, investors. And then it’s just like this wave, like this tidal wave after that. And that’s what we’re really excited about.
John Koetsier:
It’s the interesting thing to me is the pace of technological change is pretty insane right now and certainly past legislators ability to keep up … regulators ability to keep up. You mentioned aviation, of course, that’s a heavily regulated space.
Gregg Hill:
Right.
John Koetsier:
I think warehousing is probably a little less so, but there’s probably some hurdles you have to go over. But given the pace that you’re developing and you think 2024, you’ll have actual functioning product out there. Uh, there’s going to be some cycles. There’s going to be some, uh, iterations and some fixes. There’s going to be better software, but this could scale within, let’s say 2030, 2031 or something like that. If things progress as I’m guessing you hope, and as I’m guessing the team hopes as well, and that means the world changes very, very significantly, very, very soon. It’s. It’s been a pleasure chatting with you guys. Thank you so much for your time.
Jesse Coors-Blankenship:
Thank you, John.
Gregg Hill:
Thank you.
TechFirst is about smart matter … drones, AI, robots, and other cutting-edge tech
Made it all the way down here? Wow!
The TechFirst with John Koetsier podcast is about tech that is changing the world, including wearable tech, and innovators who are shaping the future. Guests include former Apple CEO John Scully. The head of Facebook gaming. Amazon’s head of robotics. GitHub’s CTO. Twitter’s chief information security officer, and much more. Scientists inventing smart contact lenses. Startup entrepreneurs. Google executives. Former Microsoft CTO Nathan Myhrvold. And much, much more.
Subscribe to my YouTube channel, and connect on your podcast platform of choice: