Humanoid robots are finally leaving the lab. Neo from 1X will start shipping to home this year for just $20,000. I just had a chance to talk one-on-one with Dar Sleeper, the head of product and design at 1X, to get into all the details he can reveal about Neo.
(And a few that he couldn’t!)
Dar is one of the few people in the world who live with a humanoid robot, which gives him unique insight into what’s going to change when we all (maybe?) have humanoids. He explains why most people misunderstand “degrees of freedom,” reveals why Neo’s new 22-degree-of-actuated-freedom hands are such a breakthrough, and shares how 1X’s vertically integrated approach helps the company hit aggressive pricing and manufacturing goals.
- Get the deepest insights concisely on the TechFirst Substack newsletter
- Subscribe to the TechFirst YouTube channel to never miss an episode
- Or listen on Spotify or Apple Podcasts
Watch our conversation here:
(And btw, I saw Neo’s NEW hands in a video Dar showed me privately before we started recording. They are literally AMAZING. Dar does say a few things about them which are pretty cool. These are going to be among the best robot hands on the planet.)
We dive deep into:
- Why humanoid robots need to be soft, quiet, and safe
- The hidden complexity of shipping a real robot product
- Why 1X chose the home before factories
- The psychology of living with a humanoid robot
- Designing Neo to feel natural instead of “sci-fi”
- Why tendon-driven robots have a huge advantage
- The long-tail problems nobody thinks about (including robot laundry)
Dar also shares stories from his Tesla experience, explains why most humanoids are too dangerous for homes, and discusses how robots could become as normal as smartphones and AI assistants. If you want a realistic look at the future of humanoid robotics … not hype, but actual product thinking and engineering tradeoffs … this is a must-watch conversation.
Full transcript: Neo humanoid robot
Dar Sleeper: Twenty-two degrees of actuated freedom. I think people don’t really understand the differences here. People hear degrees of freedom, and they just take that as a given.
They don’t understand the variance. So 22 degrees of actuated freedom means you can actually actuate every degree of freedom that is available.
John Koetsier: Hello and welcome to Humanoid Mega Hub Live. My name is John Koetsier. This is the podcast on humanoid robots. Today we’re chatting with somebody really cool. He’s the head of product and design at one of the most exciting humanoid robot companies in the world, talking about 1X. His name is Dar Sleeper.
He hasn’t been sleeping. He’s been busy. He’s former Tesla, reportedly redesigned Neo from the ground up in his first six weeks on the job. We’ll see about that. He’s riding the tiger of a rapidly expanding company that has a massive shipping deadline coming up very, very soon. Welcome, Dar. How you doing?
Dar Sleeper: Yeah, how are you, man? It’s good.
John Koetsier: So pumped to have this conversation. So many things to get into. We’ll jump into all of this stuff, but we have to hit this one. I mean, you have this shipping deadline coming up, that end of the year you’re supposed to start shipping your humanoid robot. How are you feeling about that?
Dar Sleeper: I’m feeling good. I think the thing that nobody understands outside of being inside a humanoid company is just the deep complexity, because you can be feeling 100% on certain vectors, and the other vectors are like, “Oh shit, we gotta go.”
But nonetheless, shipping is really the ultimate constraint to make sure that you take technology forward. So we’re super excited to start shipping this year.
John Koetsier: I mean, it is the ultimate thing, right? It’s kind of the point at which it’s real. It’s super real.
I’ve talked to so many people at humanoid robotics companies. I ask them, “What’s the hardest part?” and they say, “Everything,” which kind of rhymes with what you just said. There’s always another thing, right? It’s the hardware, but there are so many pieces of the hardware. It’s the software, which is so challenging, and it’s a difficult world.
Dar Sleeper: It is. It goes to everything from the product experience and the small little details, from the end-to-end customer experience from the second that you place a preorder, to the second you fulfill your order, to the second you take the order, and the second you’re done with your Neo, hopefully 10 to 100 years later when you’re tired of being a Neo user.
And there’s operations, there’s everything. So it’s all of the probably most intense parts of business, along with the most intense parts of shipping a hardware product. It’s just about as much fun as I think you can have in the world. But yeah, we’re super excited.
Getting the product out there is super important. Getting that feedback loop is super important. We gotta get going. The humanoid space needs good products.
John Koetsier: Love it. So I showed you a little bit off record and off camera of a venture that I’m working on that’s gonna release shortly.
A partner in that venture has ordered a Neo, so he’s super excited to receive it and get it. I mean, 20K, that’s gotta be a loss leader, right? That’s really aggressive pricing, especially for a made-in-USA product.
Dar Sleeper: Yeah. I can’t tell you anything specifically about our BOM costs in terms of the specific numbers.
I will say that we are able to achieve BOM costs that I don’t think anybody else in the US game can achieve. I can’t speak directly to how the Chinese BOM costs look, but it really comes down to the simplified architecture. And I’m not the engineer here, but Neo’s tendon system and the motors that we use to drive Neo’s tendon system reduce the number of total parts in the system because there’s compliance and safety built into the system.
So when it’s moving its limbs and it interfaces with the world, there’s much less energy in the system, so you don’t have to have springs or dampers at the joint level. It reduces the part count.
Because we’re so vertically integrated, we make so many things in-house and we do all of our manufacturing here in the US, we’re able to control our supply chain.
We also have 10-plus years now of innovation on motor design and manufacturing, tendon design and manufacturing, all this stuff. So we’ve been able to drive the BOM cost super low, and that’s really one of the reasons why we’re able to offer products for interesting prices.
John Koetsier: The vertical integration has to be one of the key stories about 1X. You released a detailed blog post about that a week ago, maybe two weeks ago, and for me it was mind-blowing. Literally mind-blowing. The amount of vertical integration you’re doing there is off the charts. It’s Apple-esque, except Apple never did that with hardware, right?
And you’re doing the hardware and the software, and you’re saying that’s what’s enabling you to keep your bill of materials cost really low. The other part is, we see right now with the Strait of Hormuz situation, the Iran situation, we saw during COVID, supply chains are weak, right? When something isn’t in a country’s best interest, they shut off supply.
So owning as much of that as possible is insurance against those sorts of things.
Dar Sleeper: Yeah. It’s totally a dance because there’s no situation where it’s all farm-to-table. It’s not all locally sourced materials. You still have to get certain materials from certain countries, and there are parts of your supply chain that’ll always be a little bit less vertically integrated than the others.
But yeah, tremendous shout out to Bernt Børnich. He’s always been a proponent of “we have to control our destiny,” whether it be whether or not we build the software integration in-house or use a third party, or whether we fully vertically integrate the supply chain to the fullest extent we can within the realm of all the different constraints of logic and cost.
He’s taking that super seriously. People don’t really understand this all started with Bernt in a barn making the motors from scratch. I think that little kernel extends all the way through the entire company’s history and all of the different decisions we’ve made over the last 10-plus years.
John Koetsier: It’s amazing how much a founding story can actually become the DNA of a company. That is absolutely really, really cool.
I gotta hit something since we’re talking about parts and hardware. You showed me something before I started recording that was off-the-charts interesting, exciting, amazing on hands.
I know you’re not revealing everything right now. Tell me what you can about the hands that you’re gonna be shipping on Neo.
From what I saw, what you showed me, wow.
Dar Sleeper: Yeah. We did announce in the product launch, when we made the preorder available to everyone, that we were gonna be shipping with 22-degree-of-freedom hands.
We didn’t get super deep into the details. We kind of left it a little bit ambiguous. We showed a limited amount of operation with those to start.
So much of the magic here I’ll save for when it hits the public and let Bernt tell the story because these are really a brainchild of both the hands team here, which I believe are some of the best engineers in the world.
Our hands team is the tightest-knit group of people. You can wave hi to the left of this wall, and they go hard. They’re in here seven days a week, and they’re around the clock.
And Bernt has made some crazy architectural decisions that have put us in a place where we have one of the most interesting hands in the world.
Twenty-two degrees of actuated freedom. I think people don’t really understand the differences here. People hear degrees of freedom, and they just take that as a given.
They don’t understand the variance. So 22 degrees of actuated freedom means you can actually actuate every degree of freedom that is available.
A lot of people say 22 degrees of freedom, and it’ll just be like these degrees of freedom are both able to move on the axes because they’ve enabled that in the hardware, but you can’t actually actuate them. So you can’t actually do anything with that degree of freedom. Sometimes that makes it worse because then you can’t control it.
So there’s 22 degrees of freedom versus 22 degrees of actuated freedom, and then there’s also 22 degrees of freedom open loop or closed loop versus not.
So it’s the ability to—most hands that are 22 degrees of actuated freedom, which I’m not even sure there are any, probably the Chinese players in the hand space, but at least not for American humanoid players—you can actuate the hand closed, but you can’t actually open it. It’s usually just spring-loaded.
So I’ll try to save as much as I can for the hand for when we do the 1X-style reveal and show our customers and show the world in the most exciting and fun way we can.
Neo’s hand is probably one of the most interesting hands in humanoid robotics.
John Koetsier: Well, what I saw and the speed that I saw was pretty mind-blowing.
One thing I’ll say as well, because you’re also the product guy and the design guy, it’d be really nice if humanoid hands, when they’re not actually in use, wouldn’t just be like this all the time, right? If the hands had sort of a natural curve to them.
I don’t know if that requires continuous application of force or power or anything. I’ll leave that with you.
Dar Sleeper: No, I’ll actually answer that. I’ll butcher the explanation of this, so all the engineers that end up watching this, please forgive me.
I actually asked Ralph, who leads software here, about that one of my first days on the job. I was like, “Dude, this is game-blowing. How many more days do we have to look at this?”
And he said it’s because of the fact that, when I said the hands just actuate closed and then they spring open, if you want them at a natural position, it’s a calibration issue. Unless you have the closed-loop actuation, you can’t actually have them at a natural position when they’re resting because then you can’t calibrate them.
Insert details on the engineering side here that make this a little more intelligent sounding.
But essentially, we finally now have natural-shaped hands in a resting state.
John Koetsier: There’s nothing like shouting in your face, “I am an artificial being,” more than hands that are like this, right? That’s really cool.
And it actually kind of segues nicely.
Dar Sleeper: Yeah, I was gonna say, people don’t actually understand how many of those little nodes exist across the entire system.
A lot of it is obvious because we’re humans and we kind of know. The best engineers and designers in my experience through humanoid robotics, as I’ve spent time in this space, have been people that in the middle of a conversation will just start looking at themselves.
So yeah, there are a lot of crazy scientists in the space, a lot of really wacky designers. Little notes like that exist across the entire system, from the head to the shape of the body to the way the feet move.
The amount of time that goes into fine-tuning these walks, these runs, all the different things Neo can do from a locomotion perspective. The amount of times I’ve personally put on a mocap suit and tried to create the perfect walking references to then simulate a bajillion times and train good Neo models on, and then start tweaking and fine-tuning the walking with the RL team that sits right behind me.
It’s amazing. People don’t understand the depth of each little touchpoint with the world.
John Koetsier: Nobody understands how deep you can go into something, almost anything, until you actually go into it and you want to make it amazing, full-featured, incredible. You can go deeper and deeper.
And of course the magic is having a Steve Jobs type—maybe Bernt is like that as well—who says, “Okay, that’s far enough, now we’re shipping,” right? The engineer always wants to optimize. There’s always another thing to do. At some point you need to ship.
It was a great segue because I wanted to get into design decisions that you made with Neo. That’s everything from how it looks, how it stands, how it holds its hands, other things like that. Talk about some of those key design decisions, clothing even.
Dar Sleeper: Yeah. I always like to start off the design conversation with engineering because to me we don’t get to have all the work we did on industrial design without it.
There are some delineated differences. There’s the industrial design, which is what it looks like to your eyes—the A-surfaces, the clothing, the head, how the head looks, the shape of the head, how the eyes look, what lenses we choose, all that stuff.
Then there’s the design engineering, which is getting down to the component level of building all of the internal structures and all of the drivetrains that move and actuate the robot, to the connectors, everything that goes into building Neo.
And then there’s quite a bit more, like chip design, but we’ll stay away from that.
It all starts with the engineering because the tendon drive is super interesting. It creates soft, quiet movements. When it moves, you can’t hear it.
It’s the faintest sound. I remember the first time we made a Neo video, we actually sampled the sound from Neo moving and lifted it in post when we were editing the video because we wanted people to hear a little bit of what a robot sounds like.
And this is starkly different than any other humanoid. I don’t know if you’ve ever been around one, but they make really, really loud noises usually. Even the small ones.
So it’s the soft movements, very fluid, very reminiscent of human motion because it’s tendons. It’s pretty close biomimicry of the human form.
And then it creates safety from the inside out, and it’s lightweight. We’re talking 60-plus pounds versus the 200-pound monsters that a lot of these robots are.
That informed the direction. It all started from there. This was well before my time, so I can’t take credit for any of that stuff.
It was actually Bernt and Bernt’s brother-in-law that obsessed over what is the right envelope for the humanoid robot in terms of the body shape and proportionality.
I guess I can’t give away their secret sauce, but they took a few different figures that all needed to be layered over each other and then sketched out the envelope from there.
That informed the basis of Neo’s shape.
Then what became my job was, okay, what does this look like to people?
So take it outside of just the form factor of the body. What’s the most optimal shape and height? To us that was informed by what would be the least creepy and fit the best into spaces and operate the best throughout spaces.
So being able to reach everything it needs to reach, being able to fit in between things. A big fat chubby robot can’t squeeze between tight spaces.
But after we got the utility out of the way, it really became about this: humanoid robots were going in a direction that, to me, felt like a lot of copy and paste.
This isn’t a take on the direction of other robots, but it was all some iteration of a Tron: Legacy sci-fi universe where you have shiny metals and LEDs that are futuristic.
And to me, the world doesn’t really look like that, and I don’t think everybody else wants the world to look like that.
I think there’s a big population of us in Silicon Valley that are super hyped on that.
Personally, I’ve just been in the headspace of the future actually looking a lot more similar to what we have today. In fact, I actually think it looks even a little bit more beautiful and natural.
Elon Musk actually had this whole idea of turning parking lots into parks and forests once FSD can keep cars from ever having to be stationary. I think that’s beautiful.
I think that’s the right direction for the world, one where we’re a lot more in touch with nature, our family, and all the things that we love.
So I wanted to create a shape and a form that looked a lot more organic and felt like it fit in. It didn’t feel too futuristic. It didn’t feel that far out.
It felt like the future had already arrived, and in fact it was familiar. It didn’t need to feel like it was dependent on something in the future happening. It was just right here in front of you. You can accept it.
And then the last thing that really guided every single one of the decisions—from what threads to choose for the knit, to what the shape of the head should look like—the head’s super simple. It’s just a face, a head, and an ear.
Basically, the reductive minimalism of all the different design pieces is very visually unstimulating. I kind of just want it to fade into the background.
Not in an uninteresting way, not in a way that becomes ugly and you don’t need to look at it. No. When you look closely, all the details are very beautiful.
The knit has a lot of small little details that you pick up on. The knit itself is an amazing feeling. You put your hand on it, and it feels very soft and cozy.
The head, again, the knit that we have wrapped around the head on the soft covers of the head, is very beautiful.
So all the little elements of it are very gorgeous, but when it’s in your house, it kind of fades out into the background.
Then when you pay attention, you say, “Oh, it’s a pretty good-looking robot,” and it’s kind of cute. The cuteness definitely needed to be a part of it.
We wanted it to be pretty non-assuming, like something that didn’t demand you look at it.
If we have these huge metal creatures in our homes that feel super futuristic, it’s gonna feel way too juxtaposed to the whole home environment.
And nothing in the world looks like steel walls and light strips.
John Koetsier: I really like that. To me it comes across as sort of Scandinavian design, a little bit of minimalism.
And I like that you’ve had the foresight to think about what does it feel like. Is my robot hard and cold, or is it kind of soft like furniture?
Because if you have one in your space, you’re gonna brush up against it. You’re gonna touch it on occasion. What’s that like to live with a humanoid robot? What did it feel like the first time, and what’s it feel like after a week or a month?
Dar Sleeper: Living with Neo is like…
Neo definitely inherently takes on a personality.
I don’t think people fully agree on this broadly across even 1X, let alone the audience of 1X. But for whatever reason, Neo shows up as this little guy, kind of a homie. He’s always the one that wants to give you a fist bump and a head nod.
I can’t get rid of that.
I would say the parts that are great are the fact that it doesn’t feel out of place. It feels like it fits. Some of this probably comes from the fact that Neo’s suit design kind of came directly off me ripping my couch apart, so it fits next to the couch very well.
It’s the little moments of understanding that the utility Neo can provide doesn’t always have to be these hyper-difficult tasks.
People jump directly to things like Ikea furniture and the end-to-end laundry sequence and how hard that is. And by the way, that is super hard.
If people tackle that, that’ll be on the list of full-self-driving-level timelines. That’s on that end of the spectrum in terms of difficulty.
But what people don’t understand is the sweatshirt that you left on your couch, the shoes that you left in a spot that isn’t exactly where they’re supposed to be next to the door, the blanket being left out in the living room and not in the wicker basket.
These little things actually really build up in your psyche and get really stuck in your head because you walk into the house and you see it and you kind of ignore it at first because you’ve had a long day and you just want to get on to the thing you like to do in your home, whether it’s hanging out with your wife, your girlfriend, your kids, or whatever.
But it’s the little things that I think are hyper-impactful.
And then the more funny things, I think these are more audience-friendly, are generally speaking that new people coming over is probably the weirdest thing you can have at your home for now.
I find that my friends’ girlfriends at first take the longest time to get used to it. That’s actually a little bit of design feedback in itself. We gotta figure out how to neutralize the acceptance of Neo across all parties.
But yeah, having a robot in your home is not normal. Not yet.
John Koetsier: Not yet. Does it get normal after a few months?
And what are those little things that really matter? Does it pick up that sweatshirt and move it to the bedroom? Does it pick up that blanket? Does it move those shoes?
Dar Sleeper: Yeah.
The time to normalize, believe it or not, is no more than a few minutes.
I think people don’t credit humans enough with how adaptable we are.
We are extremely adaptable.
I’ve noticed that throughout my entire life. Some sort of new norm arrives.
We’re talking to computers every day now, and this is happening. And then I’m saying this and it’s obvious. Yes, of course, we talk to our LLMs all the time.
But that’s an insane sentence if you go back five years ago.
John Koetsier: Three years ago.
Dar Sleeper: Yeah. It’s amazing.
The further you go back, the more insane the sentence sounds.
But even the moment that happened, we all accepted it pretty immediately.
It’s delayed because it’s not even really distributed yet. It’s now reaching the Midwest and the outskirts of the suburbs.
But it’s not that much of a different acceptance curve. It’s just a matter of when it hits their phone.
It’s gonna be the same with robots.
People will view it as, “No way. That’s insane. This isn’t real,” at first. And then, “Oh, that’s happening.” Then it’s in your front yard. You haven’t quite accepted it yet.
But once it creates value in your life and brings good to your world and ends up right there in front of you, it’ll be accepted at a very similar pace.
John Koetsier: What’s the maintenance stuff? Do you have to change its clothes every few days and launder them? Do its hands get dirty and do you need to wash those? What do you do?
Dar Sleeper: Yeah. This is touching on the long tail of really hard problems.
People all over Twitter are always like, “Where the hell is my robot?”
And my answer to that is always super genuine.
At 1X, we try to answer these things from a company level too. We send out messages from Bernt that he’s written for our customers, and we have blogs. The main narrative is trying to help people keep up to date with what’s the latest.
“Where’s my Neo?” is the ultimate question.
My answer is always pretty honest. “Hey guys.”
Because I was at Tesla before, and I was part of the launch and preorder and eventual launch of the Cybertruck, which had a four-year air gap.
But that’s a car. The whole thing was made out of steel, and there are a lot of really impressive things about mass manufacturing. There’s nothing easy about what happened there.
But that was a car. People have made cars before, and that was a four-year gap between preorder and delivery.
That’s not our goal for Neo whatsoever.
But I tell people there’s such a long tail of hard problems to solve.
The truth is, until we have a great product in people’s homes—we have a great product from a fundamental standpoint. The core of the product is something we’re hyper-confident in. We’re excited to get it to people.
But it’s these long-tail problems like the one you discussed, which is Neo’s clothing.
Here’s one of the hard problems that we’re actively iterating on. We have iterations of this suit coming through every week. We’re iterating through multiple details on the suit to make it so Neo can take it off himself and clean it himself.
That’s the obvious thing. If you had to do your humanoid’s laundry, you’d be pretty pissed.
John Koetsier: My butler’s butler.
Dar Sleeper: Yeah, I’m a butler’s butler. It’s not where you want to be.
So the obvious answer is to make the best product experience, and these are the long-tail issues.
At some point you say, “Okay, we’re shipping.” There are some pretty high-bar requirements to make that happen with safety and reliability.
But as long as you have those fundamentals down, there are gonna be people who would accept some inconveniences.
If every time Neo’s suit needed to be washed, it needed to be serviced and powered down, or you had longer cycles between suit replacements and had to get it replaced during quarterly servicing or whatever the timeframe might be for an early customer, they would probably accept it.
These are things that really matter and make the product experience everything from 100 to somewhere lower on the scale.
John Koetsier: What I will say, having seen the hands that you showed me before we started recording, is it’s kind of comforting to know that the hardware capabilities are going to be off the charts.
As you get more out there, get more data in, add more training data, and your AI models get more sophisticated and your world models get more sophisticated, you’ll be able to, just like Tesla, ship new features over the air and create new capabilities.
All of a sudden, “Oh, my Neo can do this,” and then, “Oh, my Neo can do that.”
That gets exciting for people because it is a leap of faith to be one of the first on the planet to own and have a humanoid robot in your home.
That’s a big leap of faith. It’s not an inconsiderable expense. It’s not off the charts either, but it’s not inconsiderable.
It’s gonna be interesting what the day-one experience is like.
It’s gonna be interesting what the day-365 experience is like.
Dar Sleeper: Yeah, 100%.
I think the idea that your product gets better and better over time is so sticky.
There’s really only a few AI products in the world that compare to this, but we go through this with our LLMs getting upgraded frequently, and we go through this with our Teslas getting over-the-air updates and being better at driving themselves.
It is so pleasing.
Basically, as an early adopter, you feel hyper-gratified, and then you kind of get to tell your friends who thought you were crazy when you were the early adopter.
But then it broadens the distribution.
The second somebody gets in my car and I turn on self-driving and it does something better, or the second I show my mom a new feature on an LLM and say, “You can actually do this for you now, Mom,” all of a sudden you’ve created a customer out of that person.
I think it’s just right, especially for frontier technology, being able to ship a product that has hardware capabilities that take it to level 10, 100, one million, and then improve upon that and ship better and better features.
You don’t want everybody to have your product to start.
Of course it would be perfect if everything was fully ready for everyone on day one and it just worked. That’s never happened in the existence of humanity with any product.
Being in a position where you need trust from early users, and you need early users that want to provide lots of feedback, that’s what we want.
We want a design that we can improve and iterate on, ship new features, get feedback, iterate again, ship new features, and then expand that to AP as the experience becomes more palatable for a larger audience.
John Koetsier: There’s so much more we could talk about. We can’t talk forever.
Maybe we’ll close with this one.
You’re one of the few humanoid robot companies that are explicitly targeting the home.
There are some saying, “The home is way down there on my list of priorities. I’m looking at logistics facilities. I’m looking at warehouses. I’m looking at factories.”
That’s interesting in and of itself, and I’d love to hear your thoughts on that.
But also, are you targeting other verticals? Can you put Neo into different jobs and different places?
Dar Sleeper: Yeah. The short answer is yes.
Neo has always been for general labor.
As a spec’d-out piece of hardware, Neo is very capable of much more than just home chores.
Applications of Neo, in my opinion, will be as far and wide as you can imagine, anywhere people need a helping hand, whether it’s addressing labor shortages or aging populations in different countries.
There are so many important places where Neo can help the world.
The question mark around home versus industry is a hard story to tell if you’re anybody but 1X.
This isn’t me trying to punch down on competitors.
It’s just fundamentally true that if you have a harmonic-drive robot—insert pretty much every other humanoid competitor besides maybe the planetary gear-driven robots from China—if you have a harmonic-drive robot that’s 150 to 200-plus pounds, it’s very heavy.
It’s high-energy, very stiff, low back-drivability.
You can’t really be around people in a safe way.
John Koetsier: Yeah.
Dar Sleeper: It’s really risky, in short.
So it’s a hard story to swing, and the narrative comes downstream of the constraints.
People say, “Yeah, we could probably see them fitting into a factory.”
That’s not the only reason.
People also say it’s because consumers won’t really adopt the technology and it’ll be hard, so start where nobody cares.
I actually think that’s probably wrong because anywhere a human should fit, a humanoid should be able to interact among humans.
Safety among humans is very important.
If it’s a hyper-repeatable, high-precision, no-humans-around task, you might as well just use a KUKA arm or something.
But when it comes to the home, the home is actually interesting for data diversity purposes.
The distribution of data you get in the home is tremendous. It’s far greater than any other repeatable action you can do in a factory or logistics center.
So the data for making these models better will be better.
And this part isn’t talked about much, which is precision.
If you’re on a factory line, you’re chasing nines. Ninety-nine point nine percent uptime, ninety-nine point nine percent accuracy.
If you get a contract from a customer that wants to use your robot on the line, it’s gonna say, “I want your robot. It needs to be uptime this much, accuracy ninety-nine point nine percent.”
And don’t get me wrong, this is the goal for Neo.
The goal is to be able to perform any task in the world at a ninety-nine point nine percent success rate.
But to start with frontier technologies, that’s not the place to go.
You want to be in a place where if every once in a while you fail, it’s okay.
If my Neo goes to get me a beer from the fridge and misses the beer, gets me ketchup, or drops the beer every 50 or 100 times, I’m not gonna throw my Neo in the garbage.
I’m gonna say, “Hey Neo, thanks. Can you actually get me a beer this time?”
That’s a little different than a factory, where if you break down the line, it’s a huge problem.
So I think the home is actually a very interesting place.
It’s certainly a good place for many reasons, from data diversity to distribution.
If you think about PCs in their early infancy, going through the red tape of business-to-business contracts versus consumer adoption, consumer adoption is really what let PCs take off.
So I think there are so many reasons why consumer and home makes sense.
But in short, Neo’s not just for the home. Neo’s for everything.
John Koetsier: Dar, this has been fascinating.
I wanna thank you. I know you’re crazy busy. You’re probably working seven days a week and as many hours as you can fit in.
Wish you the best of success, and looking forward to the launch of Neo.
Dar Sleeper: Yep. Thank you. I appreciate it. Good talking.