Latest / Automated with Brian Heater / Nic Radford on Humanoid Robots Beyond the Backflip

Transcript

Nic Radford on Humanoid Robots Beyond the Backflip · Automated with Brian Heater

• 0:00
  SPEAKER_00: I'm a deeply technical CEO.
• 0:03
  SPEAKER_00: And as much as I love the technology, my product is not the robot.
• 0:07
  SPEAKER_00: My product is the company.
• 0:10
  SPEAKER_00: We are just obsessed with the commercialization.
• 0:13
  SPEAKER_00: How to bring these machines to market.
• 0:15
  SPEAKER_00: What exact activities are they going to perform?
• 0:18
  SPEAKER_00: What are those needs?
• 0:19
  SPEAKER_00: What are those requirements?
• 0:20
  SPEAKER_00: Maybe we don't need to open the aperture so broad that we spend time trying to get the robot to do backflips.
• 0:27
  SPEAKER_01: It strikes me that if ROI is there and getting to market and getting a product in customers' hands as quickly as possible are some of the goals, then you want to just get this thing out there doing a couple of things in the real world.
• 0:39
  SPEAKER_00: Persona is obsessed with creating a business, not just technology demonstrations for technology demonstrations' sake.
• 0:48
  SPEAKER_00: Technology grabs headlines.
• 0:49
  SPEAKER_00: The backflips grab headlines.
• 0:51
  SPEAKER_00: They grab the clicks and the YouTube views.
• 0:54
  SPEAKER_00: But the places that are probably going to make money are far less sexy.
• 0:57
  SPEAKER_00: I've put robots in space, I've put them two miles underwater, building robots for hazardous and tough environments.
• 1:04
  SPEAKER_00: They are markets that are unsung.
• 1:07
  SPEAKER_00: The romantic sci-fi view is a robot in every home, irrespective of the fact that that might be a horrible market.
• 1:32
  SPEAKER_01: This week's conversation with Nicholas Radford was recorded just ahead of the fireside that we did at the Humanoid Robot Forum in Chicago.
• 1:41
  SPEAKER_01: It's kind of a combination prep call podcast episode with somebody who we've been trying to nail down for quite some time.
• 1:48
  SPEAKER_01: Nick discusses his work at NASA and the development of his current startup, Persona AI.
• 1:54
  SPEAKER_01: He also actually debuted the company's newest robot at the Humanoid Forum the following week.
• 2:00
  SPEAKER_01: So you'll get that in here as well.
• 2:02
  SPEAKER_01: It's a fun chat.
• 2:03
  SPEAKER_01: Thanks to Nick and Persona.
• 2:04
  SPEAKER_01: Thanks to you as always for tuning in.
• 2:05
  SPEAKER_01: Don't forget to like and subscribe and check out our newsletter over at automated.fm.
• 2:10
  SPEAKER_01: In the meantime, please enjoy this conversation with Nicholas Radford of Persona.
• 2:15
  SPEAKER_01: Vision and AI are changing what's possible in industrial automation from quality inspection and predictive maintenance to robotics, logistics, medical devices, aerospace, and manufacturing.
• 2:26
  SPEAKER_01: At the Advanced Vision and AI conference, you'll hear from industry leaders, see real-world AI applications in action, explore the latest in machine vision, computer vision, sensors, edge computing, and automation technology.
• 2:38
  SPEAKER_01: You'll also be able to connect with the people who are shaping what comes next.
• 2:43
  SPEAKER_01: Whether you're building, buying, or implementing these systems, this is a chance to stay ahead of the curve and see how vision and AI are moving from promise to practical impact.
• 2:53
  SPEAKER_01: Learn more and make sure to buy your ticket early at the Advanced Vision and AI conference September 23rd and 24th in Santa Clara.
• 3:04
  SPEAKER_01: Of all the sports and of all the track and field sports specifically, how does one end up doing the high jump?
• 3:15
  SPEAKER_00: You know, it's uh funny question.
• 3:18
  SPEAKER_00: Um I was like in the seventh grade, and we were as you do, they corral you on the track and they're like, what one of these events may you have any innate talent in?
• 3:33
  SPEAKER_00: And somehow, yeah, and somehow I just I uh just lined up for the bar and just jumped over it.
• 3:42
  SPEAKER_00: And that was just a very natural thing for me to do.
• 3:45
  SPEAKER_00: And the coaches were just like, oh my gosh, you know, you have to typically you have to like try to really work with a student to to do the high jump, and it just became so it was so natural to me, and then really defined my entire identity all through middle school and high school and college.
• 4:00
  SPEAKER_01: Yeah, it's it's an interesting one.
• 4:02
  SPEAKER_01: So so it was it was calling your name, you know, you were you were able to do it.
• 4:05
  SPEAKER_01: You felt like, I mean, your your form was just like pretty good from the outset.
• 4:10
  SPEAKER_00: Very naturally so.
• 4:12
  SPEAKER_00: I mean, there's uh I I can't really describe it more than more than that.
• 4:16
  SPEAKER_00: It's it was just uh like riding a bike.
• 4:18
  SPEAKER_00: It was very natural, very natural thing for me to do and realized I could jump and uh ended up jumping pretty high.
• 4:28
  SPEAKER_01: When you say that defines uh the era of your life, I mean, how in what sense?
• 4:33
  SPEAKER_00: Well, it's continued to define my life because uh turns out I've I've been stricken with been stricken with some very bad arthritis, and uh actually ended up destroying my ankle uh doing the high jump, had to have it replaced.
• 4:48
  SPEAKER_00: I've had about 10 ankle surgeries, a total ankle replacement.
• 4:52
  SPEAKER_00: I have a titanium ankle now.
• 4:54
  SPEAKER_00: And uh live in continuous pain, but C5, contusion, all sorts of stuff.
• 5:02
  SPEAKER_00: And um, a lot of it was brought on by the abuse I put my my body through in my track and field days.
• 5:08
  SPEAKER_01: Yeah, that's a not a very fun answer, but uh answer.
• 5:12
  SPEAKER_00: Yeah, don't want to bring don't want to bring the mood down.
• 5:14
  SPEAKER_00: Sorry about that.
• 5:16
  SPEAKER_01: No, it's fine.
• 5:16
  SPEAKER_01: No, so you're so you're I mean you're you're part robot, right?
• 5:19
  SPEAKER_00: I mean you have a metal Yeah, got a metal foot, got a metal, got a metal joint.
• 5:23
  SPEAKER_00: And listen, track and field, it was it was one of those amazing things.
• 5:28
  SPEAKER_00: Getting to compete in college uh for the with the decathlon was was an incredible experience.
• 5:35
  SPEAKER_00: But as I mentioned, got hurt and then decided that I was gonna have to graduate this engineering career and and figure out a way to to to get out.
• 5:45
  SPEAKER_01: So so it was it was sort of a question of like, hey, am I going to be able to do this track and field thing on a professional level, or do I have to settle for this whole, you know, like robotics thing?
• 5:55
  SPEAKER_00: No, that's actually uh an insightful question because when you're a track athlete, especially at the collegiate level, you're looking to make the Olympic team.
• 6:04
  SPEAKER_00: I mean, that's that's what you're trying to do.
• 6:06
  SPEAKER_00: Because that's pretty much that's it, right?
• 6:07
  SPEAKER_00: I mean, that's yeah, exactly.
• 6:09
  SPEAKER_00: You know, the weenie's box here.
• 6:10
  SPEAKER_00: We here we come.
• 6:11
  SPEAKER_00: But uh unfortunately I didn't even make the Olympic trials team, let alone to make the Olympic team.
• 6:17
  SPEAKER_00: So that would kind of be you know highlighted to me that that uh you know I wasn't gonna be a professional track star, which was completely fine.
• 6:25
  SPEAKER_00: And then injury illustrated that it was really not gonna happen.
• 6:29
  SPEAKER_00: And and uh but I had a love for engineering, and Purdue was one of the most amazing engineering schools you could go to for various different reasons, and uh got out and found my way down to NASA and the rest is quite history.
• 6:43
  SPEAKER_01: Am I sort of overstating the connection here as far as you know putting putting your body through like difficult tasks, injuring yourself and an interest in the human form and at several different points in your career ending up in and around humanoids?
• 6:59
  SPEAKER_00: When I got injured, it really drew my attention to the human body and the complexities of the human body and specifically the ankle.
• 7:06
  SPEAKER_00: In fact, if if you talk to any of the engineers I've worked with, I've been obsessed with the ankle for a long time.
• 7:12
  SPEAKER_00: Even on this humanoid, I had a lot of opinions on the ankle we were designing.
• 7:17
  SPEAKER_00: Um, I was really fortunate enough to be working in the powered orthotics and exoskeleton group for a time at NASA as well, where we were working with patients with paraplegia and other and other limb pathologies and learned a lot about designing wearable devices and again biomechanics and human mechanics that helped inform back how to design robots and specifically humanoids.
• 7:40
  SPEAKER_00: So it's really been a full circle thing.
• 7:42
  SPEAKER_00: But uh it's when you when you live with limb like impairments, and and I don't want to overstate this, like there are people with serious injuries, and uh I don't want to ever feel like you know, there there are there are um if you let me back up in working with that, you get to meet some people with more devastating injuries than you have, and and therefore I'd always told myself I never I I don't ever want to complain about what I what I go through, even though it is a big deal to me, they're people much worse off.
• 8:17
  SPEAKER_00: But a long-winded way to say it is it has been one of my deep interests in human motion.
• 8:24
  SPEAKER_00: And in robotics, the humanoid is really the the holy grail expression of that.
• 8:29
  SPEAKER_01: Talk to me about the the ankle.
• 8:31
  SPEAKER_01: We talk about the the hand all the time, obviously.
• 8:33
  SPEAKER_01: You know, that's like the big the big problem everybody's trying to to crack right now.
• 8:38
  SPEAKER_01: Um why is the I mean, you know, for obvious reasons there's your own experience, but beyond that, is there a reason why the ankle has been such a source of fascination for you?
• 8:49
  SPEAKER_00: Uh yeah, it's extremely complicated.
• 8:52
  SPEAKER_00: Uh the the joint itself is is is a very complicated joint.
• 8:57
  SPEAKER_00: We we approximate it poorly, even in today's advanced robotic standards.
• 9:02
  SPEAKER_00: We approximate all the the true degrees of freedom of the ankle, the human ankle, really poorly.
• 9:07
  SPEAKER_00: There's a lot of subtlety to it.
• 9:09
  SPEAKER_00: There's a lot of uh off-axis motion that it's able to produce that isn't just easy to do in robotics, you know.
• 9:16
  SPEAKER_00: So the dorsal flexion and the planar flexion, we understand that pretty well, right?
• 9:20
  SPEAKER_00: We can get joints to do this.
• 9:21
  SPEAKER_00: The inversion and the eversion of the ankle, yeah, we can do that pretty well as well.
• 9:26
  SPEAKER_00: But the ankle actually has a significant amount of yaw to it, and there's a lot of compliance to the ankle so that when you uh there's a lot of form in the foot, the med, the mid-tarsal, uh, the the metatarsal area that that breaks people that that that bends on the metatarsals, people don't approximate that even at all sometimes.
• 9:46
  SPEAKER_00: And uh those all those things matter.
• 9:48
  SPEAKER_00: And in my my ankle specifically, I don't have I don't have dorsal flexion, I don't have anything.
• 9:54
  SPEAKER_00: Um, it's a fixed, it's a fixed titanium joint.
• 9:58
  SPEAKER_00: So uh that's if you notice, that's when I walk, I I limp.
• 10:02
  SPEAKER_00: And and uh some people that don't know me, they're like, oh, are you okay?
• 10:05
  SPEAKER_00: And I'm like, no, yeah, it's fine.
• 10:07
  SPEAKER_00: But but I have a limp to the way I walk because my biomechanics range in motion is it's very it's very asymmetric because I can't, I can't move the ankle through all of its any really of its range in motion.
• 10:20
  SPEAKER_00: But I do have, but my metators, they still flex a little bit.
• 10:24
  SPEAKER_00: And so you can recover some of that that motion back.
• 10:28
  SPEAKER_00: But if the human body adapts and you can still ambulate, you can still walk.
• 10:32
  SPEAKER_00: Uh but in robotics, the ankle is a very difficult joint to build, primarily because you don't want to put a lot of weight at the bottom of your leg, right?
• 10:44
  SPEAKER_00: You want to minimize the distal mass.
• 10:46
  SPEAKER_00: But in order to get all those degrees of freedom, you know, you need a bunch of motors, or you need to situate your motors and actuators proximal of that ankle joint and then run a bunch of cables, which is then complex.
• 11:01
  SPEAKER_00: Robotics doesn't have the the same sort of human muscle equivalents.
• 11:07
  SPEAKER_00: I mean, companies like clone are looking at it, yeah.
• 11:10
  SPEAKER_00: That Polish company, but in general, our actuators don't stretch and compress in order to create motion around a pivot point like the human body does.
• 11:21
  SPEAKER_00: So in order to get an ankle joint to flex or to extend or to rotate, you got to put big bulky motors down there.
• 11:31
  SPEAKER_00: Big bulky motors are heavy.
• 11:33
  SPEAKER_00: And then when you maximize, or whether when you have a lot of distal mass, now you walk really energy intense, you drain your battery faster.
• 11:42
  SPEAKER_00: And so most people start just approximating what the ankle does.
• 11:47
  SPEAKER_00: They approximate it, but it's also a very strong joint.
• 11:50
  SPEAKER_00: When you jump, there's an enormous amount of power that comes from the ankle joint.
• 11:56
  SPEAKER_00: And that's one of the reasons I was looking into it.
• 11:59
  SPEAKER_00: And the ankle joint can almost be on par with the hip joint.
• 12:02
  SPEAKER_00: That's how fascinating the ankle joint is.
• 12:05
  SPEAKER_00: It's an exceptionally strong joint, and it's really highlighted by your calf muscles, right?
• 12:10
  SPEAKER_00: Your calf muscles are just like pushing all that power into the ankle joint that stiffens up when you jump.
• 12:17
  SPEAKER_00: And uh, it's fascinating.
• 12:19
  SPEAKER_00: But robots, we suck at building ankles.
• 12:22
  SPEAKER_00: Um, yeah, I've followed a lot of Harris work over the years, and you know, somebody that I remain in contact with.
• 12:30
  SPEAKER_00: And uh, you know, he knows more about the ankle, forget more about the ankle than I'll ever know.
• 12:36
  SPEAKER_00: But uh he's got some insane, incredible academic research in the biomechanics of the ankle and trying to approximate it with electromechanical hardware.
• 12:45
  SPEAKER_00: And uh it's super complicated.
• 12:47
  SPEAKER_00: It's a really, really complicated joint.
• 12:49
  SPEAKER_00: A new joint.
• 12:50
  SPEAKER_00: It's easy, right?
• 12:50
  SPEAKER_00: Just a little little now, you know, there's there's a there's a little bit of there's a little bit of weird motion in a knee.
• 12:58
  SPEAKER_00: Most sort of, you know, it's mostly just a pen joint.
• 13:01
  SPEAKER_01: Yeah, so so uh, you know, again, obviously there there's a there's this piece and and there's this power required to to uh to to jump to propel somebody up, and then you know, obviously like to be able to land and have uh the pieces not break up.
• 13:16
  SPEAKER_01: And we've seen an actual example of that, but and and and to to land on two feet.
• 13:20
  SPEAKER_01: But you said in a recent interview, effectively, you know, I'm I'm trying to put a robot to work and not to get it to do backflips, you know, in reference against all these like kind of show-offy videos out there.
• 13:30
  SPEAKER_01: Um so when it comes to something like the ankle or any other part, I mean, are you risking over-engineering it if you're just trying to build a robot that does, at least for now, a certain set of things in a certain environment?
• 13:47
  SPEAKER_00: I think you can take it too far.
• 13:49
  SPEAKER_00: Um I I feel that uh our design philosophy and our design methodology has really found a good compromise.
• 13:59
  SPEAKER_00: I'm actually watching the robot right now walk in our lab.
• 14:02
  SPEAKER_00: Um, just that's why I glanced over here where I'm gonna do that.
• 14:05
  SPEAKER_01: I was wondering what you were distracted by and like that's the one thing.
• 14:07
  SPEAKER_01: That's the one hands except I just I keep watching it.
• 14:11
  SPEAKER_00: I mean, I I should probably adjust the camera so you can see.
• 14:15
  SPEAKER_01: Yeah, yeah.
• 14:15
  SPEAKER_01: If you're like a game was on, then I, you know, I might be offended, but like the robot walking by, yeah, it's fine.
• 14:20
  SPEAKER_00: Yeah.
• 14:21
  SPEAKER_00: But no, you're not wrong.
• 14:23
  SPEAKER_00: I think there are certain ways in which we can take it too far.
• 14:27
  SPEAKER_00: And then there are ways in which we don't go far enough.
• 14:29
  SPEAKER_00: So it it really depends on on the use case.
• 14:33
  SPEAKER_00: Uh it depends on really what your requirements are.
• 14:36
  SPEAKER_00: And um but you know, that comment I made was had more to do with where we think the real value in the market really is.
• 14:47
  SPEAKER_00: Yeah, right.
• 14:47
  SPEAKER_00: Of course it it's it's like we at Persona specifically, we are just obsessed with the commercialization, how to bring these machines to market.
• 14:59
  SPEAKER_00: What exact activities are they going to perform?
• 15:02
  SPEAKER_00: What are those needs?
• 15:03
  SPEAKER_00: What are those requirements?
• 15:04
  SPEAKER_00: We don't maybe we don't need to open the aperture so broad that we spend time trying to get the robot to do backflips.
• 15:12
  SPEAKER_01: Well, that that's that's my that's kind of my point, right?
• 15:15
  SPEAKER_01: Is that is that you know, when building a robot at this stage, is that and and when creating the hardware is that you don't have to create hardware that is capable of doing backflips if backflips aren't on your roadmap.
• 15:29
  SPEAKER_00: The backflips are useful from a capabilities demonstration, right?
• 15:37
  SPEAKER_00: It's like look how powerful the hardware is.
• 15:39
  SPEAKER_00: And let's be honest with ourselves, that's it's very impressive.
• 15:41
  SPEAKER_00: Yeah, it's cool.
• 15:42
  SPEAKER_00: It's extremely impressive.
• 15:43
  SPEAKER_00: It's it's cool.
• 15:44
  SPEAKER_00: It's it's absolutely cool and very impressive.
• 15:46
  SPEAKER_00: It's a good checkpoint.
• 15:47
  SPEAKER_00: Like, hey, check it out.
• 15:49
  SPEAKER_00: From a control systems standpoint, from the way we taught it to do that, from the from the mimic, the behavior mimic, no doubt human motion capture was used, retargeted in simulation, control policies were developed, you know, likely using a bunch of ML techniques, pushed out to the robot, and it created it and then had to fine-tune it for the physics differences between the real world and how they trained it in simulation.
• 16:13
  SPEAKER_00: And voila, they got something to work.
• 16:16
  SPEAKER_00: So, from that standpoint, it's very impressive.
• 16:20
  SPEAKER_00: We probably could have used that time a little better to create a robot that or to create a policy that had some economic value if you weren't just trying to be, you know, a fun toy.
• 16:32
  SPEAKER_00: Um and and but there are places for that.
• 16:35
  SPEAKER_00: Let's not forget ourselves, entertainment's a big deal.
• 16:38
  SPEAKER_00: And and these platforms could have some value in entertainment.
• 16:42
  SPEAKER_00: And and I'm not I'm not criticizing that.
• 16:44
  SPEAKER_00: What I'm criticizing is um companies that have this excessive, excess almost like this obsessive view on how to on building a business.
• 16:56
  SPEAKER_00: Easy to form a company, right?
• 16:58
  SPEAKER_00: We just sign a piece of paper, you have shares, I have shares, boom, we formed a company.
• 17:02
  SPEAKER_00: But a business is a cash-flowing monster.
• 17:05
  SPEAKER_00: It's a cash flowing machine where you are providing something to the market of such value that participants in the market will exchange, exchange that value for dollars.
• 17:17
  SPEAKER_00: Right.
• 17:18
  SPEAKER_00: That's a hard, it's a tough, it's very tough to do.
• 17:20
  SPEAKER_00: They're self-sustaining things.
• 17:22
  SPEAKER_00: That's a business.
• 17:24
  SPEAKER_00: And persona is obsessed with creating a business, not just technology demonstrations for technology demonstration's sake.
• 17:33
  SPEAKER_01: Yeah, well, I mean that brings up an interesting point because another thing that I've seen a lot of humanoid companies uh in particular obsessed with is speed.
• 17:43
  SPEAKER_01: But when we talk about speed, I'm not talking about speed to market.
• 17:47
  SPEAKER_01: I'm talking about look at how quickly we developed this robot.
• 17:51
  SPEAKER_01: Look at how quickly we got to to to walk in, you know.
• 17:54
  SPEAKER_01: They and and again, it is impressive.
• 17:57
  SPEAKER_01: It's impressive where the technology is, it's impressive how quickly um this stuff is coming together now, but they don't seem as focused on that next important part, which is getting it to market and and you know, getting to ROI.
• 18:11
  SPEAKER_01: Is that do you feel like that is really your primary focus right now?
• 18:16
  SPEAKER_01: Is how quickly can we get this to a point where it is actually useful to people in the real world doing real world jobs?
• 18:25
  SPEAKER_00: 100%.
• 18:26
  SPEAKER_00: I mean, I'm gonna I'm gonna screw up the the saying, but if it's not correct, I'm just gonna invent a new one.
• 18:32
  SPEAKER_00: You know, a first-time founder talks about technology all the time.
• 18:36
  SPEAKER_00: Like, check it, check out this thing that our company made.
• 18:40
  SPEAKER_00: Look at the specs.
• 18:42
  SPEAKER_00: Check out what it does.
• 18:44
  SPEAKER_00: A second time founder talks about go to market.
• 18:47
  SPEAKER_00: Right?
• 18:48
  SPEAKER_00: The technology is like over there, right?
• 18:50
  SPEAKER_00: It's it's like there's there's uh this person takes that as an assumption.
• 18:55
  SPEAKER_00: Yes, there's obviously there's risk, there's technical risk.
• 19:00
  SPEAKER_00: But but this but in my view, a second time or a a multi-time founder understands that that is that's that's that's important, but it is such a small, narrow part of the equation.
• 19:14
  SPEAKER_00: Product market fit, which markets?
• 19:17
  SPEAKER_00: What's the business model?
• 19:18
  SPEAKER_00: Who's your customer?
• 19:19
  SPEAKER_00: What's the pipeline?
• 19:21
  SPEAKER_00: What's execution look like?
• 19:23
  SPEAKER_00: What's the long tail market risk that I'm taking?
• 19:25
  SPEAKER_00: What's adoption curves look like?
• 19:27
  SPEAKER_00: I mean, like, that's all we're obsessed about at persona right now.
• 19:30
  SPEAKER_00: Like that's like literally, that's what we're thinking about nonstop.
• 19:33
  SPEAKER_00: The technology is like, yeah, you put enough smart people in a room and give them enough money and point them in the right direction, you're gonna, you're gonna have something to work with.
• 19:42
  SPEAKER_00: But if that hits a dead end because you didn't think about the rest of it, well then you equally screwed up by not getting the technology right either.
• 19:51
  SPEAKER_00: So uh we my my job more specifically.
• 19:54
  SPEAKER_00: I mean, I'm I'm a humanoid roboticist.
• 19:57
  SPEAKER_00: This is literally all I've ever done in my life.
• 19:59
  SPEAKER_00: I got out of university when I finally graduated.
• 20:04
  SPEAKER_01: You did high jump for a period there.
• 20:05
  SPEAKER_01: Let's be totally clear.
• 20:06
  SPEAKER_01: So that's your thing.
• 20:08
  SPEAKER_00: Yeah, exactly.
• 20:09
  SPEAKER_00: But you know, I get out of college and I get to get to NASA and find my way into the Robonaut lab.
• 20:15
  SPEAKER_00: And until I left the government, that's where I worked.
• 20:18
  SPEAKER_00: I worked in the Robonaut lab, building a robotic astronaut, investigating, you know, starting out as a designer and then ending up as a PI and every job function in the put in between, you know, being a chief engineer all the way to a principal investigator.
• 20:32
  SPEAKER_00: You know, I'm a deeply technical CEO.
• 20:35
  SPEAKER_00: And as much as I love the technology, my product is not the robot, my product is the company.
• 20:42
  SPEAKER_00: And it's a difficult transition to make.
• 20:44
  SPEAKER_00: And when you think about your product as the company, you think about all the other ancillary structural things that make companies successful, to which a small component of it is technology.
• 20:56
  SPEAKER_01: Yeah, I I this this brings up another interesting element in all of this, and and this is something that is um it it's not uh it's not specific to these sorts of robots, but it definitely is like, you know, like fairly unique.
• 21:12
  SPEAKER_01: When we're talking about generalizing and general purpose uh robotics, it's the question of, you know, are we sitting here waiting until we create this model, this model that can do everything?
• 21:25
  SPEAKER_01: Or when we're talking about deployment, uh, do we want to just get these robots to a point where they can start to do a few things in the world, um, you know, start the flywheel up, start collecting that physical data.
• 21:37
  SPEAKER_01: Um and it strikes me that if like ROI is there and getting to market and getting a product in customers' hands as quickly as possible are are some of the goals or the early goals, then if you want to just get this thing out there doing a couple of things in the real world.
• 21:54
  SPEAKER_00: That's correct.
• 21:54
  SPEAKER_00: I I that is more our view.
• 21:58
  SPEAKER_00: And we are We are obviously our pursuit is to create a general purpose humanoid someday.
• 22:07
  SPEAKER_00: It's to enable it with the skills to do that.
• 22:10
  SPEAKER_00: We just differ with several companies on how to go about it.
• 22:13
  SPEAKER_00: We are more targeted in our go-to-market.
• 22:16
  SPEAKER_00: We're targeted more in the economics, which which use cases have the right economics, the life, the right labor scarcity numbers, the right environmental considerations, the right task complexity, the right technology feasibility costs, like the right purchasing power of the client, the right willingness to adopt by the party.
• 22:37
  SPEAKER_00: And we are targeting more specific tasks now.
• 22:42
  SPEAKER_00: Instead of keeping the aperture so broad that we're gonna show the robot folding laundry, then maybe moving sheet metal and maybe moving a box and doing a sorting activity, then coming back over here and and showing the robot in some other config, it's just all over the place.
• 23:03
  SPEAKER_00: Now, sure, you know, it's got some usefulness to it, but in order like the the titration time to get a general purpose humanoid done and to market is going to be lengthy.
• 23:14
  SPEAKER_00: And maybe alternatively, you could do that in a bit of a more of a piecewise approach.
• 23:20
  SPEAKER_00: Take a page out of Peter Thiel's Zero the One book.
• 23:23
  SPEAKER_00: Don't be so concerned about the size of your first market.
• 23:26
  SPEAKER_00: Just go in and dominate it and then start moving adjacently and creeping adjacently.
• 23:31
  SPEAKER_00: So we've picked welding and shipbuilding.
• 23:34
  SPEAKER_00: That's our first market.
• 23:35
  SPEAKER_00: We're partnered with HT Hyundai.
• 23:37
  SPEAKER_00: We're delivering robots to them at the end of this year.
• 23:40
  SPEAKER_00: We're on our third generation of our system.
• 23:43
  SPEAKER_00: We've built a Gen 1 platform, a Gen 2 platform, and we're now kicking off the production on our Gen 3 platform.
• 23:50
  SPEAKER_00: All that done in 18 months.
• 23:54
  SPEAKER_00: And we're making that delivery for our pilots in November of this year.
• 24:01
  SPEAKER_00: That'll be 22 months of active engineering time.
• 24:06
  SPEAKER_00: So we think that's a beachhead market.
• 24:08
  SPEAKER_00: And then move more broadly through that with other shipbuilders, and then go uh slightly more broad in energy infrastructure, doing fabrication activities, grinding, welding, painting, assembly components, more of a heavy, heavier side.
• 24:25
  SPEAKER_00: You know, we're not talking iPhones.
• 24:28
  SPEAKER_00: But um but the heavy industrial segment is the market that we think has the right makeup.
• 24:33
  SPEAKER_00: You know, whenever you use tools, you have much more higher value add.
• 24:39
  SPEAKER_00: So tool usage is really what we've, you know, if you want to plant the flag on what persona does, we think we've got the right dexterity approach to use an array of tools.
• 24:50
  SPEAKER_00: We have the right data collect approach to train the skills for the robot, and we have the right customer base to deploy those robots into the right unit economics.
• 25:00
  SPEAKER_00: Go dominate that, and then we'll worry about laundry in a few years.
• 25:06
  SPEAKER_01: Yeah, I wonder how much of uh the fact that none of these companies early on, like this has changed a bit since, but none of none of these humanoid companies early on were going after these sorts of markets was a result of I what I would assume is an added level of difficulty creating a robot that can also withstand the elements.
• 25:31
  SPEAKER_00: Yeah.
• 25:32
  SPEAKER_00: I I like I like our first choice.
• 25:35
  SPEAKER_00: I like where we found traction.
• 25:37
  SPEAKER_00: Um it's it's calling upon a lot of our expertise.
• 25:41
  SPEAKER_00: I've put robots in space, I've put them two miles underwater.
• 25:44
  SPEAKER_00: Building robots for hazardous and tough environments is something that I understand and understand how to organize the engineering efforts around and qualify those systems for those areas.
• 25:56
  SPEAKER_00: And so I like I like those markets.
• 26:00
  SPEAKER_00: Uh, but but to your point, Brian, uh, they are markets that are unsung.
• 26:06
  SPEAKER_00: The romantic sci-fi view is a robot in every home, irrespective of the fact that that might be a horrible market for a long time.
• 26:16
  SPEAKER_00: And uh, I don't know, safety anyone?
• 26:21
  SPEAKER_00: Like, yeah, I'm could you even imagine a robot kind of roaming around your house with your kids and your dog?
• 26:29
  SPEAKER_00: And it just doesn't make any sense right now.
• 26:32
  SPEAKER_01: Yeah, the speci so you worked on uh Robonaut and Valkyrie specifically when designing robot for space, like what are some of the things that you have to deal with that you don't have to deal with on Earth?
• 26:45
  SPEAKER_00: Yeah.
• 26:46
  SPEAKER_00: Uh well, there's a lot of radiation in space.
• 26:48
  SPEAKER_00: It's probably number one.
• 26:50
  SPEAKER_00: So if you are inside the space station, the atmosphere is the low, the pressure is not exactly the same as it is down here.
• 26:57
  SPEAKER_00: I don't remember the numbers exactly, but if we live in 14-ish, 14.7-ish PSI here uh in space, it's more like 10, something on that order, maybe 11.
• 27:11
  SPEAKER_00: Um, I don't remember precisely, but it's it's not exactly 14.
• 27:15
  SPEAKER_00: I don't I don't think, but it's like 10 to 12, something like that.
• 27:20
  SPEAKER_00: Uh and so the environment's not that much different than here, other than the fact it's in microgravity.
• 27:28
  SPEAKER_00: But the radiation is very different.
• 27:32
  SPEAKER_00: And so the radiation uh really disrupts electronics.
• 27:37
  SPEAKER_00: There's a lot of types, there's a lot of different types of radiation, and there's ways to shield it.
• 27:43
  SPEAKER_00: This the space station structure does provide some natural shielding to it, but higher energy particles still get through, and they can really disrupt electronics.
• 27:50
  SPEAKER_00: And so the more, the more advanced the processor, where this the smaller that we've had to cut this silicon for the higher speed processors, the more easily disrupted they are by this ion ionizing radiation.
• 28:04
  SPEAKER_00: And it's really bad if you go outside, and then it's really bad if you get outside the banana belts, and then it's really bad if you're on the moon in a in a solar storm.
• 28:14
  SPEAKER_00: This stuff uh can really just destroy things.
• 28:17
  SPEAKER_00: So you have to think about that.
• 28:19
  SPEAKER_00: Uh with Robina, when we flew into the space station, we used this triplication strategy where each of the robot actuators had literally three processors that were called TMR'd, uh triple triplicated mode redundancy.
• 28:35
  SPEAKER_00: And so each command, each cycle, there were three processors that would all vote in effect on what the right state was, and they were doing it thousands of times a second through all 44 degrees of freedom of the robot.
• 28:51
  SPEAKER_01: This is like the because the physics are especially challenging that the you need a consensus.
• 28:58
  SPEAKER_00: Yeah, because the the radiation flips bits.
• 29:02
  SPEAKER_00: It like, you know, think about think about radiation, this this sort of these high-energy particles coming from deep space and and disrupting the electronics, and they change their, you know, they change their bits in their machine code.
• 29:18
  SPEAKER_00: It kind of literally flips bits around.
• 29:20
  SPEAKER_00: And it, you know, causes bugs and errors, and and so you want to check against that.
• 29:26
  SPEAKER_00: And but that's one thing.
• 29:27
  SPEAKER_00: That's a that's a major, that's a major pain point of flying robots in space is radiation.
• 29:32
  SPEAKER_00: Obviously, if you're outside the vacuum is is is a big deal.
• 29:36
  SPEAKER_00: But I would contend that it's actually much harder going underwater than it is flying.
• 29:42
  SPEAKER_00: It's harder to get to space, you know, you have to extend a serious amount of energy in a very short period of time to get to space.
• 29:49
  SPEAKER_00: But once you're there, you're kind of it's kind of chill.
• 29:52
  SPEAKER_00: Underwater is like, oh my gosh, you know, for every psi down there, yeah.
• 29:57
  SPEAKER_00: Yeah, so every every 10 meters you go underwater, it's another atmosphere, which is another 14 psi.
• 30:04
  SPEAKER_00: So if you go down, if you go down, you know, a hundred meters, that's 10 atmosphere.
• 30:10
  SPEAKER_00: If you go down a thousand meters, that's a hundred atmospheres of pressure.
• 30:16
  SPEAKER_00: That crushes stuff like like that, if you haven't designed it right.
• 30:22
  SPEAKER_00: Well, you know, wait till you go down 3,000 meters, 5,000 meters.
• 30:28
  SPEAKER_00: That's 500 times the pressure you feel on land underwater, 500 times.
• 30:37
  SPEAKER_00: And then if you think about 1961 with Walsh and Picard, when they went down uh to Challenger Deep in the tr in the Trieste, they were at 11,000 meters.
• 30:50
  SPEAKER_00: Fascinating.
• 30:51
  SPEAKER_01: Yeah, the uh one of the things that that I I've been talking to folks quite a lot about lately are the DARPA challenges.
• 30:58
  SPEAKER_01: They really have kind of created the modern robotics industry in the U.S., I would say, in a pre-without a doubt.
• 31:05
  SPEAKER_01: Bullprint is the godfather, and there's it's it's not even close.
• 31:08
  SPEAKER_01: For sure.
• 31:08
  SPEAKER_01: Yeah, like especially, you know, I mean they uh uh you know, humanoids, like that there's no there's no atlas, or at least we don't get atlas as quickly as we do without that.
• 31:17
  SPEAKER_01: I and I know that you were involved in in in those.
• 31:20
  SPEAKER_01: Obviously, you know, you were involved in um in NASA for a number of years.
• 31:25
  SPEAKER_01: Um as far as the the the work you were doing with with NASA um and and building humanoid robots, what's your what's your sense of what the end goals on a project like that ultimately are?
• 31:43
  SPEAKER_00: With NASA?
• 31:45
  SPEAKER_01: Yeah, we'll start with we'll start with NASA in terms of like all this you know funding that's going into what what at the time I'd say you know certainly like when the first robot was created, humanoids were a moonshot.
• 31:57
  SPEAKER_00: Oh what's worse than a moonshot?
• 32:00
  SPEAKER_00: Yeah, it was much more it was much worse than that.
• 32:02
  SPEAKER_00: Yeah.
• 32:03
  SPEAKER_00: Uh well, I mean, it's it's I can speak to what our ideas were when I was there.
• 32:08
  SPEAKER_00: I clearly can't speak to what NASA's thinking about about today.
• 32:11
  SPEAKER_00: Sure.
• 32:12
  SPEAKER_00: Yeah, of course.
• 32:13
  SPEAKER_00: But the uh the proposition was pretty uh was was a couple ideas.
• 32:17
  SPEAKER_00: Um first, uh, and this is what there's a lot of uh you can think about like an astronaut helper.
• 32:28
  SPEAKER_00: That was one one aspect we were thinking about, kind of like the way a nurse would aid a doctor in an operating room.
• 32:34
  SPEAKER_00: So you've got your very capable C3PO that kind of follows you around and and hangs out with you, and and that was one of the ideas that that Robinaut was.
• 32:43
  SPEAKER_00: The other thing was it's actually very difficult to get outside the space station.
• 32:48
  SPEAKER_00: So let's say you had an emergency.
• 32:50
  SPEAKER_00: Do you know how long it takes to get outside?
• 32:54
  SPEAKER_00: I mean, you gotta like don a spacesuit, pre-breathe the right partial pressure of oxygen because the spacesuit I think runs around four or five psi.
• 33:03
  SPEAKER_00: So if the space station's at 12 psi and your spacesuit's at 5 psi, well, just like a diver gets the bins, if you don't actually bring up your oxygen content in your bloodstream to the right partial pressure, you get the bins just like divers do.
• 33:19
  SPEAKER_00: So you got to pre-breathe the right partial pressure of oxygen for, I don't know, 20, 30 minutes.
• 33:25
  SPEAKER_00: It's on that order.
• 33:26
  SPEAKER_00: I don't remember the exact times, but it's like, it's like 30 minutes.
• 33:30
  SPEAKER_00: And then you got to get outside and then you gotta, then you gotta translate over to the area and the outside where you have a problem.
• 33:37
  SPEAKER_00: If it's an emergency, you're dead.
• 33:39
  SPEAKER_00: Like the space station's done, right?
• 33:41
  SPEAKER_00: So Robinaut could sit on the outside of the space station and have this Minuteman response aspect to it.
• 33:48
  SPEAKER_00: There's a problem, an astronaut could don some teleoperation gear, step into the robot, and be on the outside of the space station with an anthropomorphic cumanoid that can operate the same space tools, the same interfaces outside that were designed for the astronauts, and then possibly go fix the problem.
• 34:08
  SPEAKER_00: So that's the other aspect that we were thinking about it for.
• 34:12
  SPEAKER_00: And uh, but then you think longer range space missions, like we're gonna go to Mars.
• 34:17
  SPEAKER_00: Uh, it turns out it's gonna be really difficult to get there.
• 34:20
  SPEAKER_00: And so, likely we're gonna send some robots ahead of astronauts before we get to Mars.
• 34:26
  SPEAKER_00: They're gonna help organize some work sites.
• 34:29
  SPEAKER_00: They may even they might they might even like partially build some structures on Mars.
• 34:35
  SPEAKER_00: Because it's a nine-month journey to get to Mars.
• 34:38
  SPEAKER_00: Then you have to wait on the surface about two two years until the orbits reline up to where it's only a nine-month journey home.
• 34:45
  SPEAKER_00: And so you're gonna want astrona you're gonna want robots there before the astronauts get there.
• 34:51
  SPEAKER_00: Then the robots and astronauts hang out while they're on Mars together, and then the robots will stay back when the astronauts go home and sort of tend the facilities.
• 35:02
  SPEAKER_00: So there were multiple use cases that that NASA was thinking about.
• 35:05
  SPEAKER_00: But to your point earlier, Robonaut was started circa 1997, 1998.
• 35:14
  SPEAKER_00: It was in those little overhead projectors a few years before that, you know, the ones I'm talking about.
• 35:20
  SPEAKER_00: And and uh and there was always on the it was always on this this concept, right?
• 35:26
  SPEAKER_00: And then uh DARPA and NASA kind of got together and really DARPA put in the first real increment of funding.
• 35:34
  SPEAKER_00: That's when you see Robonaut 1, you see a big DARPA badge on it, because DARPA was one of the early funders of the Robonaut program at NASA JSC.
• 35:44
  SPEAKER_00: And it was it was one of the most gosh, I just remember the day that I first saw it.
• 35:53
  SPEAKER_00: It was, you know how you have those memories in your life where they're just so vivid.
• 36:00
  SPEAKER_00: I can remember the first day that I walked the hall in Building 9 in front of the Dexter's Robotics Lab, and I saw Robonaut sitting there, bolted to the bolted to the the the the wall, sitting there in its like ready pose, and like you just press your nose against the glass and you're like, what is this thing?
• 36:25
  SPEAKER_00: I didn't have any idea that this was here.
• 36:27
  SPEAKER_00: Who's building it, right?
• 36:28
  SPEAKER_00: Because I wasn't originally in that lab when I got to NASA.
• 36:33
  SPEAKER_00: And it was it's just it was just this most incredible experience.
• 36:37
  SPEAKER_00: And then the hand, right?
• 36:38
  SPEAKER_00: So the Robot One hand was exquisite.
• 36:42
  SPEAKER_00: It was just a marvel of engineering.
• 36:44
  SPEAKER_00: It had some of the most advanced electronics.
• 36:47
  SPEAKER_00: The the it was a tendon actuated hand.
• 36:49
  SPEAKER_00: It was only, it was only semi-antagonistic.
• 36:52
  SPEAKER_00: So it would pull, it would pull a tendon for the for the flexion, but on extension, it was a spring return.
• 36:58
  SPEAKER_00: The main designer of that was Chris Lovchak with Ron Diffler back in the day.
• 37:03
  SPEAKER_00: And gosh, just talking to them about the early designs of that hand are just amazing.
• 37:09
  SPEAKER_00: And then, and then the guys that did the electronics, the radiation tall art electronics, I mean, just an incredible hand, right?
• 37:16
  SPEAKER_00: And then uh, you know, it was absolutely a moonshot to your point.
• 37:21
  SPEAKER_00: But only NASA really had the kind of budget in the late 90s and the early 2000s to get something like a humanoid off the ground.
• 37:33
  SPEAKER_00: And that's not to say that there weren't research, there wasn't research happening in other labs, other universities.
• 37:38
  SPEAKER_00: I mean, humanoids have been coming together in some form or fashion for, you know, since the late 70s.
• 37:43
  SPEAKER_00: I mean, just look at at Walbot from Wasiti University and and then just all the progressions.
• 37:49
  SPEAKER_00: Yeah, exactly.
• 37:50
  SPEAKER_00: Just everything, right?
• 37:51
  SPEAKER_00: Just everything.
• 37:53
  SPEAKER_00: But uh NASA threw a lot of money at it and and really advanced the state of the art.
• 37:58
  SPEAKER_00: Very, very much advanced the state of the art.
• 38:00
  SPEAKER_00: And I'm very proud of the work that was accomplished there.
• 38:03
  SPEAKER_00: It was it was an incredible place to start a career.
• 38:06
  SPEAKER_00: The just the the people that you worked around were so brilliant, just so utterly brilliant.
• 38:12
  SPEAKER_00: It just made you want to be a better engineer.
• 38:14
  SPEAKER_00: Every day you went to work and you just wanted to be a better engineer because everyone around you was so talented.
• 38:20
  SPEAKER_01: Yeah, that that I that's part of the reason, too, why I brought up uh DARPA in in the context of NASA is I think that the the other um connected tissue here is when we're talking about moonshots, um we I think part of the way that we've really kind of uh contextualized moonshots is that um even if they don't hit that moonshot, even if it isn't accomplished, that there is still value in the pursuit because, you know, like in the case of this work, okay, we haven't gotten to the point where they're doing all of those things that you said that you had initially envisioned, but we're certainly not where we are right now in this industry, right now, without that work, without a doubt.
• 39:06
  SPEAKER_00: I like the way you characterized it.
• 39:08
  SPEAKER_00: Um, moonshots allow us to overreach that even if we were to miss what we accomplished along the way on that moonshot is still substantial.
• 39:19
  SPEAKER_00: Knowledge is a compounding sort of thing as well.
• 39:22
  SPEAKER_00: And clearly, you know, Newton said, if I've been able to see as far as I was because I stood on the shoulders of giants, right?
• 39:28
  SPEAKER_00: I mean, that was something that Newton said.
• 39:30
  SPEAKER_00: And there is 100% this compounding aspect that there's there's every generation the field gets incremented forward.
• 39:41
  SPEAKER_00: And um, it was just fascinating to work during that slice of time at NASA when Robonaut was went from a lab experiment to flying on the space station.
• 39:52
  SPEAKER_00: And I learned so much from the leaders that were part of that program, all the way to the engineers that came in behind me.
• 39:59
  SPEAKER_00: And it's it was a really inspiring program to work on.
• 40:03
  SPEAKER_00: And it was an honor to lead the team that put the robot on the space station.
• 40:08
  SPEAKER_00: And then the Robonaut 2 effort turned into Valkyrie, and and it was an absolute joy in my career to lead that team as well.
• 40:16
  SPEAKER_01: Yeah, and at the same time, you've got all of this work happening in in self-driving cars, and I I think it's like pretty clear why there are so many people who were working in autonomous driving are now in cumuloid robots.
• 40:30
  SPEAKER_01: I mean, granted, like we are finally getting to a point with Waymo where like it seems like, yeah, we are finally getting there, but um but all of the work and all of the companies that didn't quite make it, I think they still are contributing also to what we're seeing now just in robotics in general.
• 40:48
  SPEAKER_00: Yeah, I think that's right.
• 40:50
  SPEAKER_00: I think uh I think there's some good lessons there.
• 40:53
  SPEAKER_00: Um, you know, to hearken back to the point you made earlier, the self-driving car community.
• 40:59
  SPEAKER_00: Uh BMW had efforts before the DARPA Grand Challenge in 2004.
• 41:04
  SPEAKER_00: That's very clear.
• 41:04
  SPEAKER_00: I don't want to take anything away from from other entrants that were working on self-driving.
• 41:10
  SPEAKER_00: But without a doubt, DARPA in 2004 with the Grand Challenge and then 2005, and then with the urban challenge that followed after that.
• 41:18
  SPEAKER_00: Clearly, they instigated this concept of self-driving, making it mainstream.
• 41:24
  SPEAKER_00: But more than that, organizing an entire academic thrust around it to where it popularized it, it made funding available to it.
• 41:33
  SPEAKER_00: It helped, it helped distill some of the component technologies like the sick laser, which was probably one of the big success stories that no one really talks about that came out of uh the LIDAR that came out of self-driving and making a more economical LIDAR.
• 41:48
  SPEAKER_00: Early on in those days, those lasers were like 10 to 20 to 30,000, right?
• 41:55
  SPEAKER_00: And now you can buy them for hundreds, if not tens of dollars.
• 42:01
  SPEAKER_00: Like that's that's some of the lessons we cannot forget is how those enabling components that only the government could ever be a customer for because it's just so cost prohibitive.
• 42:12
  SPEAKER_00: But eventually the government will expose that there could be a market there.
• 42:17
  SPEAKER_00: And then they organize an academic community that helps create some of the entrepreneurs that go raise the capital to pursue those markets.
• 42:26
  SPEAKER_00: Obviously, not every company that's going to go after that's gonna make it.
• 42:30
  SPEAKER_00: It's there's literally this Darwinism survival of the fittest as you go through this.
• 42:35
  SPEAKER_00: And voila, 15 years later, oh my goodness, 15 years later, I mean, almost 20 now, uh, we're getting somewhere.
• 42:44
  SPEAKER_00: And there is a lesson in there, kids, about how long this transformation to humanoid robots could take.
• 42:51
  SPEAKER_01: As you noted earlier, humanoids have been a through line throughout not just your career, but your your research.
• 43:00
  SPEAKER_01: Um of course, you know, you you you launched a couple of companies in the meantime, you had Jacobi Motors, you had you had Nauticus, you you had some of the non-humanoid stuff that you were doing over at NASA.
• 43:11
  SPEAKER_01: What pieces had to come together in order for you and Jerry and your co-founders to say, all right, this is it, like this is the time that we're going to launch this thing.
• 43:25
  SPEAKER_00: Yeah, it was there was a lot of pieces.
• 43:27
  SPEAKER_00: Uh, from Jerry and I's point of view, you know, him and I have known each other 20 years, and we've always been kindred spirits when it came to the challenges of designing uh humanoid robotics.
• 43:40
  SPEAKER_00: Uh, we've had a lot of mutual interests and actuation and what what are the best actuators for walking and the torque speed envelopes and the dynamics of locomotion and all that kind of stuff.
• 43:53
  SPEAKER_00: Just stuff dudes talk about when they get together, you know.
• 43:56
  SPEAKER_00: Yeah, exactly.
• 43:57
  SPEAKER_00: Just just dudes talking, exactly.
• 44:00
  SPEAKER_00: Yeah, that's looking back on that now, that was pretty pretty nerdy.
• 44:04
  SPEAKER_00: Um, what I learned through Jacoby and Naunicus and a couple other ventures that I've started was this the commercial side, not to be overlooked.
• 44:14
  SPEAKER_00: Because it's it is as much as we love to the technology grabs headlines, right?
• 44:21
  SPEAKER_00: The backflips grab headlines, they grab the clicks and the YouTube views.
• 44:26
  SPEAKER_00: But the places that are probably gonna make money are are far less sexy.
• 44:31
  SPEAKER_00: And what we wanted to do was take a rather contrarian view of the market because we were we're not humanoid maximalists, we're actually rather skeptical that the the early market traction is gonna be in the popularized markets.
• 44:47
  SPEAKER_00: And so when we came together and we said, let's build persona, we had a pretty frank conversation with each other.
• 44:55
  SPEAKER_00: And we said, if we do this, we're not we're not building a lab.
• 44:57
  SPEAKER_00: This isn't a research institute.
• 44:59
  SPEAKER_00: Him and I could go work at nonprofits if we wanted and be researchers.
• 45:03
  SPEAKER_00: That's not what was interesting to us anymore.
• 45:06
  SPEAKER_00: It was actually building a robot that would make money, that that that someone else would purchase the service, the labor service that the robot would provide.
• 45:17
  SPEAKER_00: That was a more interesting problem.
• 45:20
  SPEAKER_00: And it required this crescendo from all of the things that we learned through not only the technology maturation, but also how technology is adopted and how you wrap that into a durable business model to where a customer will come back to you and ask for another one.
• 45:41
  SPEAKER_00: That was what was interesting to us.
• 45:43
  SPEAKER_01: Well, uh Nick, uh, thank you so much for taking the time.
• 45:46
  SPEAKER_00: Brian, it was such a pleasure.
• 45:48
  SPEAKER_00: And uh thank you.
• 45:50
  SPEAKER_00: Very insightful, very insightful conversation.
• 45:53
  SPEAKER_01: Thank you to Nick and Persona.
• 45:55
  SPEAKER_01: Uh, thanks to you as always for tuning in and making it very deep into this episode of the podcast.
• 46:00
  SPEAKER_01: If you enjoyed that experience, please like and subscribe and check out more content over at automated.fm.
• 46:07
  SPEAKER_01: And we will see you just about this time next week for another episode of Automated.