Mat Vogels (00:15)
Welcome everybody to another episode of Pirates Only, a podcast where we bring together multiple founders all in a similar industry to talk about a similar topic. And today is a really, really cool topic. Personal to me, gonna be personal for a lot of folks here in the space of personal robotics. What is it like to have robots?
living in our homes, helping us with day-to-day tasks. We have three folks today that are deep into that space and can talk a lot about it. Let's hop into some introductions. Ben, why don't you kick us off with a little bit about what you're building.
Ben (00:51)
Yeah, I'm Ben. I'm the founder and CEO of K-Scale Labs. We're building humanoid robots and making them open source and available for everyone. think we're, I don't know, I don't want perjure myself, but I believe we're the best-selling American humanoid robot company, which is pretty good.
Mat Vogels (01:07)
I love it.
I was hoping for some of that spice and I think that's absolutely true. And we'll dive in more into the launch that you had recently as well. Ritwick, how about you?
Ritwik Pavan (01:17)
I'm, Britwick. I, currently lead partnerships at a stealth series B startup, have been in hardware for most of my life. Pre last two startups, heavily involved hardware, building cameras that we sold to the government to help with traffic congestion. And then, a more prop tech based startup, focused on accelerating the speed of construction. And then also I, share a lot of exciting consumer updates on X via hardware nation.
Mat Vogels (01:43)
which is a must subscribe. For anybody interested in anything related to consumer, would you say consumer hardware specifically or just hardware broad?
Ritwik Pavan (01:51)
Yeah,
I'd say like consumer hardware, robotics. ⁓ Yeah.
Mat Vogels (01:54)
Yeah,
love it. And last and certainly not least.
Mehul Nariyawala (02:00)
Hey, hi everyone, I'm Mehul. I'm a president and co-founder at Matic Robots. We're reimagining home robotics, starting with trying to build iPhone of the floor cleaning robots. We just started shipping in November of last year and got our first review in Wired, which was the 10 for 10, perfect 10, and first one in 17 years. So super excited about how we're changing that and trying to reinvent that entire space.
Ritwik Pavan (02:27)
It's a well-deserved 10 out of 10. use my Matic now. It's incredible and got to see the full in and working. So, Mayhull and the team, kudos.
Mat Vogels (02:29)
my goodness. It is the best.
Mehul Nariyawala (02:32)
Thank you.
Mat Vogels (02:37)
It's one of those products too that I think I'm excited to hopefully see over the coming decade. Because typically you see you have those magical moments and experiences with software products and we usually have those consumed via iPhone or personal computer.
The Matic, which I'm a huge fan of and I have two of them, it's a magical experience that you have in hardware, which I think is rare. And a lot of that comes into obviously what we've built into it. Maybe we even, let's stay with you on Matic. Can you give a brief background on how it started, why you chose to go into this space and maybe tackle this problem first?
Mehul Nariyawala (03:01)
Thank you.
Mm-hmm.
Great question. our background prior to this, we were at Nest.
I was a product lead for the Nest camera as my co-founder and CEO Numbnit was the computer vision lead there and we had been thinking about homes. So when we left, we were both home for the first time in our lives after having worked for about 18 years. And all of a sudden my wife's comments, because I told her I'll do everything, you know, I'll take care of all the chores. And after taking kids to school for three months and kicking them, making their breakfast and doing all the chores, I felt like I was losing brain cells by the day. So I told her that, okay, you know,
what you're doing is absolutely amazing and her comments that I feel like we're cleaning all the time or I just cleaned it started resonating and that piqued the curiosity for us because we're like, wait a minute, there are 200 plus self-driving car startups because back in 2017 that was all the rage. Equal amount of industrial robotics companies, but no one was tackling homes. So we just started digging into robotics and we really came to this conclusion that if you're going to have level five indoor robots, then you want to give them a sane perception system that we
do
and the way we thought about it is if level five cars means that cars drive like humans then level five robots must be that they behave or clean or do manipulation or any task on behalf of us like humans. So let's give them the same perception system and that means they should be able to go in any indoor environment, explore, build a map, go from point A to point B precisely 10 out of 10 times because unlike self-driving cars indoor roles did not have a Google Maps or the GPS and even self-driving cars if they don't have
that
those two technologies, wouldn't know where they were going. Like when we would be lost if we know where the road was going or where we were located on the road. So that's where really the challenge was. And that's how we started thinking about it. And floor cleaning was a perfect beachhead, perfect first space because what most people don't realize is that their iRobot has sold 50 million robots in their entire lifetime, which is 50 million Roombas. And I've never met anyone who says I loved my Roomba. So it's just kind of fascinating that, OK, here's
Mat Vogels (05:08)
crazy.
Mehul Nariyawala (05:14)
product that no one loves or no one has ever kind of been talked about yet they've sold 50 million so there's clearly a need and for us it always boils down to that that can we solve a problem can we can we do something along those lines
Mat Vogels (05:28)
Could you give maybe a differentiator for the folks that have not gone through and if you go to the website, the website's beautiful. It does a really good job of explaining how this works and differentiates it. You even have some great sections and that's how I sold it to my wife where it was, we had a Roomba. We also have two young boys in our house. is continually a disaster. It's like Legos everywhere and all this and it seemed like when it worked, the Roomba, was good. You'd walk downstairs in the morning. We'd have to like close all of our doors at night though because it was always so loud.
Mehul Nariyawala (05:36)
Mm-hmm.
Mm-hmm. ⁓
Mm-hmm. ⁓
Mat Vogels (05:55)
and there'd be times where we would like, I'd wake up when it's vacuuming up ⁓ or it's something we get stuck and you know, I would say that 70 % of the time it wouldn't complete the clean because it would get stuck on something. ⁓ Could you do, yeah, explain maybe for the folks that don't know what's the biggest differentiator between what you're building and maybe what Roomba's doing.
Mehul Nariyawala (05:59)
Yep.
Yeah.
Great question. So at a very high level, we just started from this point of view that if you were building robots in the world of AI.
what would it look like? And one of the first things we realized is that maybe it's not a circle, maybe it's not flight, maybe it's something different. So we started with this idea that if we build a truly fully autonomous robot, it has to be friendly, has to feel like it belongs in a home. But most importantly, as you alluded, it should definitely navigate any indoor world and indoor space without bumping into things. And in fact, if robot requires bumping, that itself is not a robot, in my opinion.
So
that was the first thing that can be built a robot that can avoid all kinds of obstacles and everything. And it really boiled down to as I was alluding to vision. And we had this simple idea that if it can see it, it can navigate it. If it can see it, it can avoid it. If it can see it, can clean it, manipulate it, all kinds of stuff. So it's really just started with vision. And along the way, we not only built probably the best indoor navigation system for robot in the world at the moment, but also it's quieter than a dishwasher at about 55 dBA because
We had this desire that if you have a noisy guest in your home, you probably want them to leave. So having a robot that really understands things is really cool. And then being able to differentiate between what floor type it is and what stain type it is and understand that rug with frills will probably get stuck. Or if it's a mopping, a wine stain on a hard surface, it requires mopping. So all those little things kind of came into the play and then we started building this product.
Mat Vogels (07:42)
And the one thing I'll say that I'm always impressed with, it seems like it gets better. I remember the first week or two weeks, it like cruises around the home and it's good. But now after months of being in the home, feels like it knows it so well. Like when it, when it's done cleaning and it has to go from one side of my house back to its docking station, it's like Mario Kart. just, it's straight there. It's like going through obstacles and underneath my table and it just goes direct and docks right back. Uh, the last thing I'll say, and then we'll hop over to you, Ben, but, uh, we call our
Mehul Nariyawala (07:47)
Mm-hmm.
Mm.
Yeah.
Mat Vogels (08:12)
Mo because it reminds me of the cleaning robot Mo from WALL-E like it's the same color scheme Kind of has like the same little front piece on there And our kids love that movie. So it's been it's been it's been great for the whole family ⁓ We'll certainly dive more into yeah, go ahead
Mehul Nariyawala (08:18)
That's correct.
Thank you. And
I was gonna say one more aspect that is near and dear to my heart is we always felt that you shouldn't have to jeopardize family's privacy or home privacy to actually have your floor clean or get any chores done. So being able to do the whole thing on the device, which also helps with the latency and the effectiveness of the navigation, but doing the whole thing on the device was really critical. So it's really the first vision only embodied AI robot in the world. That is shipping at least.
Mat Vogels (08:53)
And pet friendly
too. say pet our our chihuahua who's scared of the wind has no problem with the Matic and it's probably just the noise so I think that it's it's even dog approved.
Mehul Nariyawala (08:55)
and pet friendly.
Yeah.
Yeah, it's actually two things. one, part that I always talked about it is that if you kind of look at Roomba, the way they work is they just go until they see some sort of an obstacle or bump into it. If you're standing in the middle of the street and someone was walking straight at you, you'll probably say, guys, please move two inches this way or that way. So in the same way, if this robot sees them, because it behaves a little bit more like humanish in the way it sees obstacles and doesn't go straight at it, that actually helps quite a bit too.
Mat Vogels (09:30)
Yeah, I love it. We'll certainly dive more into the enigmatic side. Ben, I'd love to hear a little bit more about what you're building specifically. Maybe share with the audience what specifically you built. I'll definitely put a screen up that kind of highlights it. But maybe talk a little bit about what you're building in more detail and why you chose this particular problem to solve.
Mehul Nariyawala (09:34)
Mm-hmm.
Ben (09:50)
Yeah, so I was also motivated by my wife's complaints. I, uh...
Mat Vogels (09:54)
We've to start
a tracker to see if that's the reason for a lot of these startups to exist.
Ben (10:00)
I think that's basically why anyone does anything at some level. Most I was working at robotics research at Meta and before that I spent two years at Tesla on the autopilot team around when Optimus was getting started. I've been in software for machine learning and robotics for most of my career.
Mehul Nariyawala (10:01)
Dave.
Ha
Ritwik Pavan (10:06)
You
Ben (10:26)
I wrote some of the original CUDA kernels for the Voxel occupancy network when we retained it at Tesla. I think actually the Matic robots used some of the Voxel approach. It was really cool. Me and my friend Patrick actually kicked that project off at Tesla. I guess I'm pretty nerdy. I've kind of just been doing this kind of stuff. I think I was following the Optimus development and then that
Mat Vogels (10:36)
That's awesome.
Mehul Nariyawala (10:36)
That's exactly right. Yeah.
Very cool.
Mat Vogels (10:44)
Small world.
Ben (10:55)
you know, just some of the robots that you see in China in particular. And it felt like everyone sort of looking at humanoids from a very industrial lens. And I think a big part of that reason is because the price point is just very high for any kind of like humanoid that you can see today. But, you know, with startups, I guess you feel like you have to skate where the puck is going. And it's pretty clear to me that the bomb for a humanoid is going to come down quite quickly. this was, you know,
years ago that I was thinking about this when it was still 100k and now the bomb is closer to 5k. So we're really seeing like this new product category open up in front of our eyes. It's so early that no one really knows what it looks like, no one has any revenue, but there is just...
Mat Vogels (11:37)
very
few people even have pricing. Like that's the thing I think that's crazy too, is there's not even a price tag until really, I feel like you were the first robot company that I saw an actual price tag with. So that's how early it is. It's like, there's no even prices you can find.
Ben (11:40)
Yeah. ⁓
Yeah, I think the reason for that is because there's a general sense with humanoids that you have to build everything in-house. Like, Figure and 1X and Tesla and Aptron, I call these humanoid companies, mostly build all of their actuators, most of their stack kind of in-house. They have special hardware people for all of these things. But with the sort of supply chain maturity, actually you can get a lot more stuff off the shelf.
like, you know, consumer grade stuff, like all of our actuators we bought on Alibaba. And then I went and met with these suppliers in China and around Southeast Asia to kind of go to the source. But cameras compute, you know, and I think that's, that's where the puck is going on. Humanoids is like, like at the end of the day, even a very expensive humanoid is not particularly capable, but
A AI enabled humanoid robot could be the most valuable consumer product that has ever been created. And that's what I would like to sort of get an early start on creating. Yeah, there's been a lot of different humanoid companies already, but I didn't really see anyone doing that. that's why I decided that K-Scale should go in and take a shot at it.
Mat Vogels (13:04)
Maybe kind of you hinted on this. This is one of the questions I wanted to ask a little bit later, but I think it's a perfect leeway into it. Why do you think there's more of a focus so far on the enterprise side of it or the commercial side of it versus personal robotics, personal and humanoid? it a cost piece? And obviously businesses are more willing to pay to have a hundred thousand dollar six figure plus robot, even a million dollar robot doing things for them in a warehouse versus somebody at home
Is that the reason is it more of an acceptance piece is there a psychology piece where this is kind of things that hinted out earlier? Are people ready to have robots in their home is maybe the easy way to ask that or where do you feel that balances? As it relates to consumer versus commercial robots
Ben (13:46)
Yeah, I frankly think that the US is actually probably about six months behind China on this. China, the ecosystem really also started by focusing on these kind of B2B applications. A lot of people viewed humanoids as, I think Elon really kicked it off with Tesla building a humanoid and using humanoids for their internal assembly stuff, which is, in my opinion, was just a bad application, but it-
was sort of easy to map this onto your own business model if you wanted to start your own humanoid company. The thing that happened in China was basically a lot of companies raised a lot of money to go do this and had these kind of partnerships with these like auto manufacturers or other manufacturing companies and they just really didn't get any kind of sticky PMF. Like they would be able to put up a thousand robot order form and then all of those robots would be returned.
Yeah, like that that kind of hasn't caught up to the US market just because I think the US is sort of slower to ship. I think figure might be the only one that's really even kind of tested the waters here. But I suspect where it's going is going to be it'll be similar where I mean, even I think the figure BMW deal is not necessarily as stable as people might claim. And I think some of these other kind of contracts and deals are maybe going to be there's maybe going to
be some headwinds with these and I think that people will sort of... I was just watching the ecosystem unfold in China now everyone's going towards these kind of consumer routes and there's actually quite a bit more you know meat there like Unitree is doing very impressive sales numbers for the G1 just to sort of these kind of early consumer adopter type people and I think yeah Kscale sort of seen
similar type of person in the US, and it's pretty underserved. What I do think it's going to mean though is that the type of robot that these companies are building will have to change quite a bit. It's actually very difficult to build a expensive robot and then bring the hardware price down later. So you kind of have to start with a cheap robot and then have the software direction aligned so that you can make up for the fact that you're
you know, you're using commodity parts. So from a product development perspective, it's a totally different product. I am worried that this is gonna cause a lot of humanoid companies to go bankrupt. I want a humanoid, so yeah, figure we can fill the, yeah.
Mat Vogels (16:08)
You're like, we're gonna do it. And
the price tag, me I'm wrong, it was like just under 10 grand. Is that correct? Is that one that you're holding to and trying to sell through right now? I know that you paused orders, because you probably had too many of them, but.
Ben (16:21)
We paused
orders because people were buying it without any specs or any kind of details about what it could do and I was like, you know, we got to actually like put some real demos out before we like let people buy it. No, 9k is healthy margins for us. We are, we're basically, I we're not going to sell it at a loss. And I think for the foreseeable future.
Mat Vogels (16:27)
Yep.
That's awesome. Yeah.
Ben (16:47)
I mean, at least from a sort of business perspective, our goal really is to drive value with the software layer and sort of upsell on the software features. Very similar to Tesla with the FSD pricing model. no, we had positive unit economics at 9K. And we're also exploring a few kind of shorter form factor robots. And we'll launch kind of a couple different things.
Mat Vogels (17:10)
love it. We'll definitely dive more into those examples. Ritwik, I want to ask you, you've used Matic, you see probably more consumer hardware products than most people. When you think of personal robots, what we're going to have in our home, be it humanoid or other, what are some of the form factors that you think are going to come first, or maybe are already here today? But when you think of having personal robots in the home, what comes to mind? Kind of paint that picture for us.
Ritwik Pavan (17:40)
Yeah, it's a good question. think like the, you know, there's obviously the companies that are going for the straight shot of building humanoid robots. But when you ask any of these founders or actually go, you know, I've met with a lot of them. Most of them kind of say what Ben is saying, you know, where it's actually going to be a while before you see these like at mass scale or volume, because it's just not there yet. You know, you might build the hardware.
but the software are actually like the data to be able to train this on in a real life scenario, you know, just putting it into somebody's home today, it's likely not going to do all the things that it is, they're advertising to do. And so, you know, as, as you know, we've seen with mayhole, there's the robot vacuums that you're seeing, right? And then there's certain companies that I've been seeing that might be focusing on just like the washing and folding of laundry as the general use case in the beginning.
There are some, you know, robots that we're seeing that aren't quite humanoid robots, but kind of roll around and, you know, pick up things off the table and might like, you know, do a little bit here and there. and then you've got the ultimate form, you know, the ultimate, you know, ambitious goal of like being an optimist, you know, where it can just kind of be your assistant and, you know, do all sorts of things. and so, yeah, I mean, I personally have seen a lot more specialized robot use cases in the
b two b side which i find very interesting like i don't know if you guys have seen the palettes there's like these autonomous palettes that a company is doing where it just can help out with logistics and fulfillment in warehouses which i think is a very interesting use case my friends at borg robotics i think what they're doing is really cool on the warehouse side so i i i i kind of tend to think that
Mehul Nariyawala (19:19)
Yeah.
Ritwik Pavan (19:22)
You know, it's a very ambitious goal to get humanoid robots in the next few years into homes and actually have it operate as advertised. So I personally think it'll, it'll be more of a phased out approach. And I really like seeing the specialized use cases more in the manufacturing and warehouse setting, just because I do think that, you know, a company like Amazon is able to afford the cost at which it is today and still bake that into their unit economics and have it make sense as opposed to if you're buying that today for your home.
you know, paying $100,000, but it's not doing everything that you were initially promised to do.
Mat Vogels (19:54)
down.
And there, think they're willing to absorb some of the potential hazards that could come into that, like the learning curve that it would take for some of these where, uh, you know, there's the robots or memes that you see of, uh, even some of the palette robots where they, make a mistake and they trip somebody, they run right into somebody and it like goes through. I think of, you know, my, my house fed my two kids and I'm not as worried about the Matic running them over, but if it was a bigger robot that maybe has some weight behind it, you could, you could see that being an issue. So I think the tolerance level.
Mehul Nariyawala (20:21)
you
Ritwik Pavan (20:24)
Yeah.
Mat Vogels (20:25)
is gonna be harder on the consumer side too.
Ritwik Pavan (20:27)
Absolutely. I mean, these are not like robots, know, like if they fell, like they're pretty heavy on a kid, especially. you know, Ben had brought up figure and it's like, you know, yesterday, I think I saw a video out with the use case in the warehouse where it's like, you know, separating the mail packages, like different sorts of mail with the barcode facing down. And so it's very interesting to see this kind of, even with Optimus kind of being used within Tesla's factory first.
and then getting put out into homes. But I'm very interested in kind of seeing how it evolves because unlike, you know, software with chat, GBT or all these other AI platforms, where there's like infinite amount of data to pull from like the internet, I don't think the same necessarily applies with people's homes and real life scenarios. You know, so it's going to be interesting to see how hardware companies kind of evolve to be able to work in different homes.
Mehul Nariyawala (21:16)
you
Mat Vogels (21:18)
Maybe we kind of go back on the Matic side when you think of the hardware and the software all kind of working together. I'm kind of curious from a high level, which part is harder? Is it the hardware or is it the software? And which piece kind of has to work the best in order for you to create such a magical product?
Mehul Nariyawala (21:37)
I think it's all. the way I've always thought about the good example is really
iPhone versus Android, if you actually take a step back and think about it, neither of the devices have any unique feature left anymore. They both do exactly the same thing, yet experience is vastly different from what Android is and what iPhone is, and people prefer that experience. So product experiences are very hard to copy individual features you can. So it's really just boils down to how everything gets integrated together. But to be honest, one of the things we underestimated and, you know, unlike
self-driving cars, it's obvious that if you make a mistake, mistakes are fatal. In home, we thought that, okay, if you make a mistake, they are trivial. So maybe there is a little bit more tolerance for those mistakes. But turns out people actually, the difference between AI, current state of AI and robotics, and especially in home, what I always talk about now is that with AI, we are okay to collaborate. We're happy with the 90 % result and take it to the last 10%. With robotics, almost always you want to delegate and set and forget.
And that set and forget part is really critical. So bar for accuracy is much higher. The second reason is because we spend years learning how to craft a product or how to do machine learning engineering. So when AI does it, we get mesmerized. But we don't really go to school to figure out how to navigate without bumping or how to vacuum or how to pick up a glass. So when things that are trivial for us, if robot makes a mistake, we immediately think of them as stupid. If you had someone, another human being who kept walking and
home bumping all the time would be really frustrated as well. So it turns out that the bar for trivial mistakes, a bar for accuracy for trivial mistakes is much higher. And if you don't do that, people immediately assume it's a stupid robot. And that's really where we had to spend quite a bit of time making sure that we got to the end level of accuracy and then build what we thought is a graceful failure mechanisms when it does fail because AI isn't perfect. We still get tangled up into wires once in a while.
still run into accidents. So if humans make mistakes, AI will always make mistakes of one or the other sort. And then you have to sort of think about how to gracefully fail, how to get yourself out of it, and create that experience where it feels a little bit intelligent.
Ritwik Pavan (23:46)
Think
just to kind of tap in there if you don't mind that I was just gonna say like I think Mayho makes a great point because like one of the things is like with chat GBT right it messes up one time and I don't know about you guys, but the dashes. my god They always like you'll be like don't use dashes anymore, right? Like I used to use dashes I used to use dashes before chat GBT, but now it's like you have dashes It's like a known chat GBT if people think it's just chat GBT and you tell it a million times And it's still not gonna fix itself
Mat Vogels (23:48)
Yeah, please, please.
Mehul Nariyawala (23:49)
Yeah.
Mat Vogels (24:03)
Yeah, does it, yeah.
Ritwik Pavan (24:14)
And but in real life, it's like if you were to do the same thing with like a physical embodiment of AI, like Matic or a robot, people would just probably grab it and throw it in the trash. You know, it's like, so it's just very different, patients levels with like a real life scenario. And then just like, you're talking to a software, right? Like, I think the tolerance levels are a lot higher with like AI on a software level, even if it's like you're using cursor for code, right? Like it could make issues and you'll still work with it, but
If a robot in real life is making issues and keeps doing the same thing, you'd probably not be as okay with it.
Mat Vogels (24:46)
Do you think that helps? I feel like we're in this weird time right now where everything seems to be accelerating. hopefully the cost of manufacturing and material sciences, energy, all these things, batteries, is now kind of that perfect time where...
we can start building these more reliably, hopefully with less of the headache or hiccups, because I agree, I think that there's gonna be less of a consumer tolerance for some of those things, especially if the price point is very high. Is it just gonna continue to get better? Maybe Ben, you have a good kind of concept of this too, of what you're building. Could you have built this a year ago, two years ago even, and obviously you're still early.
But can you see the trend of all these things kind of coming together right now of everything getting cheaper, better, faster, hopefully being the perfect time to maybe start building for personal robotics?
Ben (25:35)
So I basically think for Kscale, our early adopters will not be interested in a robot that can do everything for them today, any more than the people that bought FSD were interested in a robot taxi that was available today. what I think is sort of, like what I've noticed, and just from talking to people that have ordered it, and it's kind of...
It's kind of beautiful, like frankly, is this mentality of there's people that want to see the future. They can see where the future is going to be and they want that future to be here earlier and they want to support that kind of coming. You know, I think like with, there was a lot of things I could work on right now. I could probably get funding for some B2B SaaS idea if I wanted to, but.
I wanted to kind of help the future come a little bit earlier and it's just a really difficult engineering problem but that's what I like doing anyway so yeah I would say that's kind of how things are. Certainly it would have been a lot harder to do this a couple years ago. We're definitely skating off of supply chain investments and there is a lot more general, there's like
better ideas how to build general purpose robots today than there were a couple years ago. Simulators are better, the methods are better, you can get some off-the-shelf DLA models that work reasonably well. But you know, mean, I'm not building something that's gonna be great today, I'm trying to build something that'll be great in three, four years.
Mat Vogels (27:04)
No.
So you think time horizon though, what is the time horizon do you think? Maybe whether it's your product or...
humanoid robots in the home. I'm trying to think of a scale that would make sense. Like a product today, maybe do you have an idea of like, what's a product today that is, it works really well, but few people have it obviously for maybe like a price point. What's the point in which maybe 10%, 5 % of folks that are interested in this product have one and it's working? Is that?
Ben (27:39)
Yeah.
Mat Vogels (27:40)
3,
5, 10, what does that kind of look like from your perspective?
Ritwik Pavan (27:44)
I mean, think I, yeah, sorry, go ahead, Ben.
Ben (27:45)
Yeah.
As I said, my mental model of robot adoption is going to be kind of like developer first. And the way that robot adoption will sort of look like is the type of person that can be a robot developer will just expand as the underlying ML capabilities get better. So today it starts with people that can train their own policies in SIEM and maybe understand something about fine tuning VLA model. But the goal is to have one model that can take in a command and just sort of perform that action in the
real world after you've got this sort of like fleet of robots that can collect data from all these different environments. So that's what robot adoption looks like to me is like, it's kind of like...
like cursor for robots, that's a very VC way to put it, but you're like sort of driving the adoption barrier down for people to become sort of robot developers. And at the end of the day, the robot requires very little intervention. It's almost fully autonomous. That's what I think coding will look like too in some sense. But the difference is you'll be able to do everything in your real world.
Mehul Nariyawala (28:32)
you
Mat Vogels (28:53)
Rewick, what do you think on like timeline? a lot of it, know, Matic's already here. So in some ways there's like, there are robots that are here and they're living among us. But when you think of timeline, yeah, what do you think there?
Ritwik Pavan (29:01)
of like
a fully functional, it's, you know, you're saying like, what is your, because I mean, like you said, Matic is already here, you know, and so I, you know, what are you thinking from Matic? Like, what are you thinking as the next horizon? Just so I can give you.
Mat Vogels (29:07)
Maybe. Yeah, to me it's one of those things. Yeah.
I guess I don't know because I
feel like the the it's similar to AGI, right? I know it's similar to AGI where people think like, when is AGI going to happen? Some people will say that you were already there to an extent and other people will say that, you know, we've already hit at least something that is amazing and we're already going to see the benefits of it. We don't need AGI and for it to do a lot of the things that it can do today. So that, know, there's there's pieces into that. So maybe take it to wherever wherever you think, but
Ritwik Pavan (29:23)
We're on the record here, so.
Mehul Nariyawala (29:39)
you
Mat Vogels (29:48)
Maybe let's go to the very end of something that is maybe not quite Westworld or iRobot, but very similar to what we see maybe in the forward thinking demo videos exactly from like the one extra figure in some of those.
Ritwik Pavan (29:56)
the videos today Got it. Okay
So so I think like in terms of like a full autonomous robot just being able to do everything I think we're probably like about I would say ambitiously probably three years out I would say like within the next three years if I had to make a bet probably optimists could make that three to four year timeline in terms of like
you know, having robots today where they're using tele ops and like not true, you know, like AI to like, or it's not fully autonomous. I think we could probably see that in like the next year or two actually, because I know some companies that are shipping these out already. But yeah, I would say probably like within the next three to five year horizon, we will see humanoid robots in homes that are actually doing tasks. It'll take like five years to get to volumes of scale where like people can afford it. I mean, it'll probably be in that 10k price point.
So, you know, unless Tesla or one of these companies tries to make it accessible via lease or rental or something of that sort. Yeah. Yeah.
Mat Vogels (30:56)
Yeah, which I think has got to happen, right? I mean, I feel like that's got to be the case because even a 10K
problem is more than most people would be able to afford. But if it's like a cheap car payment, ⁓ I think that's totally fine.
Ritwik Pavan (31:05)
Yeah. Yeah. It's like if you pay a thousand dollars a month
for the robot, maybe, you know, you're comparing that to a nanny or something. I don't know. Right.
Mehul Nariyawala (31:10)
It's.
Ben (31:13)
Yeah, think actually weirdly enough the consumer products sort of like BOGO stuff is probably going to be really key for adoption. Like with the 10k price point, like a 36 month payment, actually becomes like, yeah, the math starts to work out for like, oh, I have a nanny or I have a maid that cleans my house for $200 a month or something. And that's a payment for.
Mehul Nariyawala (31:14)
Yeah.
Mat Vogels (31:37)
That's what
you'd pay. And even then, mean, that's what we think with the MADEC too, where the first time we got it, I think we did the installment payments for it or whatnot. And that was the way we framed it, was it was like $100 a month or something like that. But to have a clean home every day, to walk downstairs and have it, to pay somebody to come into your home to clean two times a month would be more than $100. So that's how we were kind of thinking about it too.
Mehul Nariyawala (32:04)
and
What I would say is that sort of math or logic will have to be explained quite a bit. And the reason I talk about that is one of the, if I were to put my product hat on, for me, we always internally talk about that ultimately what people want is a solution to their problems. And robots are just the way we are solving a problem, whether it's clean floor or laundry folding or whatever that problem we're doing. And what's interesting is that we have yet to see any ubiquitous consumer electronic device that is priced
higher than $2,000. As soon as you go above $2,000, it's usually a professional purchase and even Mighty Apple sort of didn't make a dent with the Vision Pro. So there is a price point and in fact if you go back to history of Roomba, one of the most interesting insight for them was they priced it at $199, their initial product, because they thought that under $200 you don't have to ask your wife, which is...
which is all of us alluded initially. So that was actually pretty clever of them when they started at Roomba and kind of created that market. So price point is there. And then once you go about 2000, it is mobility. You guys refer to cars and scooters and things like that, that we spend time on, but that still remains a considered purchase. After a hundred years of proven utility, we don't just wake up and say, I'm going to buy a car.
It's still considered purchase. We still think through the loans, the payments, the plan discusses as a family. So to me, the other side of this is that what is the value?
And how, and this is about the ubiquity, which is, yes, you will have always a 1%, 1 % 2 % adapters who will take it for what it's there. But how do you actually keep adding that value and what is that price point? And then the product side of thing is a little bit trickier than even technology in my opinion, because if you kind of just dial back seven years, maybe even 10 years, it wasn't necessarily clear that robot taxi business model is the business model for self-driving cars. It was all kinds of stuff, right? But that's becoming more and more obvious than maybe that's how it's going to scale.
So what is that a value and maybe it is a subscription? But having tried I don't know I don't know if any of you know the history of Matic But we actually tried launching subscription on as a business model in 2023 and at least for us in this particular category It was like an organ rejection people just revolted and a lot of times it was more along the line that why is it a subscription? Because as it turns out and we may be wrong, but as it turns out that there is a strong psychological bias of
something that is inside your home.
Domestic help doesn't necessarily live with you. Like if you get cleaner, they don't live with you. But if it's there constantly, there is just this strong desire to own. Similarly, how lease has been very cool from a professional side of things, but personal side of things, people still just buy cars. And even if the lease might be more appropriate mathematically or sort of economically. So it's very interesting behavior that we thought. So we switched over to just selling it after realizing that people just don't want this as an subscription.
Mat Vogels (34:58)
Do feel like it's will robots be like a home asset? had somebody mentioned this the other day where he was envisioning that when you buy a home, you know, you have solar panels on your roof and it's one of those big purchases that you don't just wake up and think you want to do as well or a new air conditioning unit or maybe a pool stuff like that where they're big purchases. And typically it's like you look at it. I know that some people rationalize. I know I have with my own home of thinking like, ah, it's, $10,000, but the value of our home will increase along with it. Like, I wonder if robots.
Mehul Nariyawala (35:13)
Mm-hmm. Mm-hmm.
Mat Vogels (35:28)
maybe
have a similar piece. It's different than cars or maybe I'm trying to think of where that balance might be. Is it considered something that you would think of as a car where when you sell your home or if you leave or move you would take the car with you. Are robots going to be that way? I think they would. But I could also see it also playing into a value of you know sometimes you buy something and you keep it with your home if it helps sell the home maybe. I don't know. So I there's other interesting kind of piece there.
Mehul Nariyawala (35:41)
Hmm.
Yeah, just like a smart home tend to sell higher. know that.
Mat Vogels (35:56)
Exactly.
Mehul Nariyawala (35:56)
real estate build that definitely homes with robots will have that value. But I do think that if you really personalize the robot and as you know what we've observed and surprisingly the biggest thing that we got right is actually the stickers. It's kind of amazing how much people love the stickers. It's just a little delight in how much they personalize and as we're seeing this when people have any sort of, obviously with hardware you will have some of the reliability issues and once in a while you have to repair the robot and people
Mat Vogels (36:11)
We love the stickers.
Mehul Nariyawala (36:24)
get really, really attached. It's kind of fascinating how when we get support email, kind of like how you name yours Mo. We only, we never hear Matic, they always use the name for their robot when they talk about it, which is really fascinating. So there is an element of ownership, personalization, sort of belonging in the family that is coming with the robot as well. So I tend to think that yes, they will take it with them. That's my gut, but we'll see.
Mat Vogels (36:32)
Yeah.
Especially I think is the robot start to communicate. I have two young kids, one of them is six and uses, you know, chat UTP and talks to it like it's a real human. And it's sometimes I wonder, like, I wonder how.
Mehul Nariyawala (36:51)
Mm-hmm.
Yes.
Mat Vogels (37:00)
If he really thinks there's somebody real there where that connection is, because none of us grew up with that type of connection with an artificial being. I think when you add that element on top of a physical object, like if you were to add a communication piece on top of the Matic robots, it would take it even a step further. So I can't even imagine what it'll be like with more humanoid styled robots, even if they're smaller, but they still have that human
Mehul Nariyawala (37:06)
Sure.
Mm-hmm.
Mm-hmm.
Mat Vogels (37:27)
element to it and you can communicate with it, it's gonna feel like a part of the family. It's gonna be this whole other world there.
Mehul Nariyawala (37:35)
Yeah, and it's a very tricky balance to get right. I actually talk about, we talk about internally about this piece quite a bit. and I always mentioned Pixar because Pixar is fascinating because when you watch Finding Nemo or Cars, they kind of create these characters which you know are not as smart as humans, yet you know they're intelligent.
So you don't think of car as a stupid or Nemo as stupid, but you know it's still a fish or you know it's still a car and it does something funny or not. So in the same exact way, how do you balance this creator? If you're going to have a five-year-old out of 25, I have a robot that's inside of home. I really do not want her to think of as closely as another human being. think empathy for a human being has to be slightly higher.
Mat Vogels (38:00)
So true.
Mehul Nariyawala (38:17)
So then, as with the robot, what is that fine line that you create saying that, it is an intelligent robot, but it's not the same as having another human being, and how do you create that balance?
Mat Vogels (38:27)
That's so true. Something that we're all gonna have to think about over the next decade or so is drawing that line in between there. Breaking the fourth wall really quick. Do you guys have time? Do you guys have 10 more minutes of time? Are you guys good?
Mehul Nariyawala (38:33)
Mm-hmm.
I'm good, yeah.
Ben (38:40)
Yeah, I'm okay.
Mat Vogels (38:41)
Cool, Ben, I want to switch it, yeah. Yeah.
Ben (38:44)
I was just going to say something about that actually. I feel like the thing
that I like about doing hardware and there's an overlap between like hardware and AI, which is kind of this importance of this sense of taste. Where I think in a lot of kind of other types of businesses, it's almost like you're sort of following what's now possible and adopting that into some flow. Whereas hardware...
products and increasingly AI products are very opinionated. The thing that I love about Mid Journey is not that it's an AI image generator, it's this sense of taste that the Mid Journey team brings to the images that they create. I think that was why the intersection of these two things is just very, cool where like...
You know, a voice feature for a robot can be very unpleasant or very pleasant depending on the taste the person who's creating it brings to it. And it feels like a much more high agency thing to build. But then it creates quite a bit more risk because you're sort of betting on does your taste translate to something that people want to buy? But I think that's what's fun about it.
Mehul Nariyawala (39:37)
Mm-hmm.
It's actually 100 % correct, You what I find fascinating is that every single robot that talks in Hollywood movies, whether it's hell or even lards, they seem to be slightly more intimidating and scary.
And the robots that are really loved almost always have some version of a speech impediment like from Wally to E to R to D2 or even Baymax. So it's kind of interesting how that balance works out and how do you actually create that thing where you say, hey, this is friendly robot. And it has to be obvious the moment you take a look at it or the moment they start speaking. then that's really, at least from home perspective. So that's really the critical piece of the puzzle.
Mat Vogels (40:12)
So true.
Ritwik Pavan (40:32)
And I think like just one last point there is to Ben's point about taste, I think it's so important. I think in a lot of ways, we've kind of lost taste in a way to certain products. People just kind of take the same product and make replicates of it. I think it's largely due to even China. You look at the Roomba, then now you look at, think Mayho talks a lot about this, but it's like you look at Roomba, nothing has changed in the last 20 years. You have Roomba, you've got Eufy, you've got...
you've got dream you've got all these other robot companies they all look the exact same kind of largely operate mostly the same and start like it
Mehul Nariyawala (41:03)
and probably manufacturing
the same factory, sorry. I had to jump in, probably manufacturing the same factory too, sorry.
Ben (41:06)
Yeah.
Ritwik Pavan (41:07)
Yeah, probably right.
And so it's actually refreshing to see like, I got a shout out, Neo, which I'm like, it just like, you can see taste in their videos even and like the way the robot is designed. So I've got to give them credit, but, and, and it's, it's cool to kind of break out of the whole, here's the robot voice and treating it as a robot rather than like, this thing is actually going to be like a human assistant, you know, down the road.
Mehul Nariyawala (41:17)
Mm-hmm.
Ben (41:32)
Yeah, I was just gonna say that is like the ability for Silicon Valley hardware companies to compete with Chinese hardware companies is like, it's this thing that's way up in the air because the competitiveness in China is just so high. But the thing that I think, I think the thing that like I was really impressed about Matic and the hardware products that are starting to get footholds are able to just lean into that aspect of taste really aggressively.
Mat Vogels (41:32)
Yeah. ⁓
Ben (41:58)
And it just, yeah, I would say like my experience with this was when we first, cause we built everything in our garage basically. And we tried to be very opinionated about the way that our robots looked. We just used three printers. And then I had this experience of we tried to go and manufacture these in China and work with different OEMs and contract manufacturers. And it was like all of the soul kind of went out of the thing. And we made this prototype with, and we were with a partner and it was just like.
like crap, the same type of thing, but it went from something that I could show my friends and they were like, that's so cool to this thing that was like, there's no soul in it. And so like, it's kind of trite, but I feel like the design in California made in China dynamic, what that is trying to capture really is that there's this aspect of like taste and product love that is requires like a startup level of like focus and passion. And it's actually not.
easy to sort of offshore that. if products that can be yeah, yeah, don't just totally my experience with that too. So yeah, triggered.
Mat Vogels (42:56)
No, it's so true. There's a cultural thing too, yeah.
Well,
if you're learning anything with K-SKILL, you have to be able to put the googly, bobbly eyes on them so that your kids can decorate it. I will say though, the K-SKILL robot, is it the K-BOT, right? Yeah, the K-BOT, is that the one, the $10,000 or $9,000 one? I think it looks great. It walks that perfect balance of not being intimidating. You feel totally fine having it in your house.
Ben (43:19)
Yeah, they came by.
Mat Vogels (43:28)
And you know, I don't wanna talk bad about anybody or doing anything, but like the 1X robots or even the optimist ones, I...
Those ones, my kids would not like those in our house. They just would not like either of those in our house. My wife wouldn't like either of those in our house. And I think that when I was telling her about this, and she loves our manic, obviously, but then when I showed her some of the things like with One X or Optimus, she's like, we just wouldn't do that. It's just too scary. But I showed her the K-bot, think it's one where it's like in the kitchen and it's like making the bread or the toast.
And she was like, I would have that heartbeat. And it was hard to tell whether it was because it was like in a setting where she's like, I would love to have a robot like in the kitchen and helping me, or if it was just the way that it looked. But you guys are all exactly right. I think that there has to be a balance that makes it feel welcoming. And that's going to require a combination of, of taste design, even just interactivity. It's those little hidden moments. does Maddie, does the robot dance? feel like every now and then when it navigates
Mehul Nariyawala (44:25)
Yeah
Mat Vogels (44:26)
back
to the pod. It does like a spin like it like for no real like it'll it'll do a spin and every now and then I always think that it like in its brain it's like doing like la la la and just like going in like little things like that where I feel like you have to sprinkle those in to make them feel human.
Mehul Nariyawala (44:30)
Okay.
Yeah, I think it's the hesitations is the once in a while you make a mistake, something along those lines. But actually, you know, the analogy that I would use, so history doesn't repeat, but definitely rhymes. I think
Mat Vogels (44:50)
Hmm
Mehul Nariyawala (44:55)
Early 90s and arrival of cell phones and the entire iPhone from that journey of a first giant cell phones or Nokia phones to iPhone is really where I think Robots are which is it's amazing that we have so many experimentation and so many things going on But just like general magic sort of tried and failed at building iPhone in 95 in some ways going straight to humanoid without doing the expectation is is Especially the hundred thousand dollars once or anything. It's is
little bit like that. So I tend to believe in my opinion that we will see purpose-built robots and purpose-built use cases in whatever form that may be and similar to sort of like us getting PDAs and iPods and BlackBerrys and then eventually we'll see all these use cases getting combined into one devices where the value is obvious and the need is obvious. So there is an element of that and I think there is lot of... sorry, ahead.
Ritwik Pavan (45:47)
No, no, no, go ahead, Maywell.
Mehul Nariyawala (45:49)
I was just going to say there is an element of experimentation that is amazing because we saw tons of it in that late 60s, I mean sorry, late 90s, early 2000s before sort of iPhone came. So that's the place we are at, I think.
Ritwik Pavan (46:03)
Yeah, one thing that I'm curious to get your guys' take on, I know we've been talking about the robots a bunch, is on AI, like with consumer hardware. I always am interested to know, like, what do you guys think the final form factor is gonna be in terms of AI wearables? You know, like with with Mayhul kind of tapping into iPhone, you know, I think we could probably all agree that the phone is still gonna be there for some time, like, just like you have the phone and the computer, right? Like, you still use your computer on regular basis, but
Mat Vogels (46:16)
Great question.
Ritwik Pavan (46:31)
Nowadays you use a lot more time like you can respond to emails do everything on your phone So your phone has kind of substituted the computer like what do you guys think the next substitute for your phone is gonna be? You know, I've got my take but I'm curious to hear what you guys say about it Like you've seen the wearables, you know The necklaces the handheld the humane and then you know the glasses like what do think the final form factor is gonna be?
Ben (46:53)
I have a pretty strong opinion on that. The thing is, the smartphone is just kind the dominant platform in the internet era. And I think the reason for that is because it just provides so much value. So that a virtual reality headset or a pin or a smartwatch, it can never compete with that kind of standalone value that the smartphone has.
You know, this is my feeling at like with, with Tesla, like we're on the FSD. The reason I was so excited about FSD was because it's, it's like the sort of first kind of like, AI native hardware product that I could think of. And what AI native means is like AI model capability scale with the amount of compute that you put behind them. And a car that has this huge battery pack on it can just afford to run very like
highly capable AI models at the edge in a way that a smartphone can't. And that was broadly my original idea around what would a AI native hardware device look like, especially a consumer device. The key problem with a smartphone in terms of building very good AI experience is that you can't do anything locally because it has to fit in your pocket. And I think that's a big constraint for any kind of like
AI pin, a humane pin, anything that you wear around your neck, it's just going to be limited. And then on the other side of that is it doesn't actually take any actions in the world. AI models, I think, are just at their best when they're doing things in the world. They're kind the agentic idea of AI models. And if AI is the thing that drives consumer adoption for
you know experiences in the next five years then I think it will open up a huge amount of space to kind of replace the smartphone just because a smartphone is an internet era product. Yeah that's that was my view. I kind of settle on humanoids because you know I felt like the supply chain is getting very mature and it is going to be a growth product but I think that broadly I mean I think that yeah the Matic robot is
got the same exact kind of features where, you know, it's a big battery pack. Like the thing about Matic is cool. It's got a Jetson in it. It doesn't have this like rock chip or some low quality compute. So you can run real AI models at the edge and the types of products that you can, like the direction for the software for like, yeah, like Matic robot or, you know, these types of products that have like big battery packs, big compute at the edge. think the software future is just so much brighter.
Mat Vogels (49:11)
Mm-hmm.
Ben (49:25)
than any kind of product where you're like constrained on it has to be wearable or has to be like small and sort of energy efficient because it's just you're going to be fighting with like the AI capabilities you can put into it.
Mat Vogels (49:25)
Yeah.
Do feel like that changes though? we develop over the next few years the ability for some of those compute and ships as those get better and better?
If we assume that that's not an issue, do you still think that the form factor is better off into something that can participate in the physical world? The last thing I'll say is I imagine my future perfect robot would be the BD unit from Star Wars, where it's on my shoulder and can walk off and do stuff around my house type of thing, where it would almost be my companion device.
Mehul Nariyawala (50:00)
Hahaha
you
Ritwik Pavan (50:05)
That's that's
got to be the first I've ever heard that but it's it's kind of I like that I kind of like it Yeah, you got like a little owl sitting on you or something So I So I want to hear me I was taken then I'll give I'll tell you guys what I think
Mat Vogels (50:10)
Yeah. Yeah.
Mehul Nariyawala (50:11)
I ⁓
Mat Vogels (50:14)
Exactly, but it folds up nicely too until like I could fit it in my pocket if I needed to.
Mehul Nariyawala (50:19)
Bye.
No,
I'm still of a believer that Google got it right when they actually built glasses just 10 years, 15 years too early. tend to because I already wear glasses and I've seen actually one of the better rate products which people which unfortunately did not find.
Ritwik Pavan (50:31)
I am with you on that.
Mehul Nariyawala (50:40)
as ubiquity as it should have because it's Facebook and no one's heard of Facebook from privacy point of view. really, Meta Raven were really, really good in many, many sense. for me, I don't think it replaces phones. I don't think it replaces sort of a need for robots inside home to do chores and things along those behinds. But if I'm already wearing glasses and it happens to kill it up for me would be if I can go to different country and it just visually translates everything I'm looking at on the fly. Now that's really where it gets very interesting.
It translates everything on the fly and I'm just wearing it. It gets really helpful. the challenge for those form factors, in my opinion, is not about the form factors as much as we haven't found the killer app. What is the killer app?
Ritwik Pavan (51:19)
Yeah,
I think, so I'm with Mayhole and it's actually, believe Google is making a comeback with the Google glasses as well. I think they are kicking off the project again. ⁓ But maybe I have, you know, not sharing anything proprietary here, MetaOrions, which are coming out. I got to do a demo the first week that it came out. Shout out to Josh too. But.
Mehul Nariyawala (51:27)
Yeah.
Mat Vogels (51:28)
yeah, Google XR, yeah.
Ritwik Pavan (51:41)
I would say the MetaOrions is probably the most impressive piece of like AI wearables I have seen to date. I mean, they're far from, I think, getting to mass production, but it was, it felt like I stepped into the future when I wore them. And you can go online and see all the information about it. But like, just, I think it, I think it is going to be the next evolution of like substituting your phone because with Meta glasses itself today, you're able to like easily take photos and record videos and stuff. You know, a lot of people use phones.
primarily outside of calling and texting for taking photos. But like Mabel said, I think apps is the next side of this, right? Like if I could just have a little piece of text on the right side with a transcription, you no longer need a teleprompter for, ⁓ like when you're giving a speech, you can literally just wear your glasses and it's showing you what you've written up, right, as a transcript or translation too. And then down the road, navigation. So when I start to think of like the initial apps,
Mehul Nariyawala (52:24)
Okay.
Ritwik Pavan (52:36)
which you already use on your phone outside of like social media in the sense that we see it today, it's already there with glasses largely. So I don't think it takes too much more to then add apps to the mix where, you know, maybe you're now able to easily share and upload the photos you're taking off of your glasses to like an Instagram-esque app. I think social media as we see it today is going to change so much more once people are wearing these glasses and have a lot more accessibility.
without having to pull their phone and you're immersing into the real world without, know, like when you're on your phone, you're looking at your phone, you're distracted from like what's going on in the real world. When you're wearing the glasses, you're still seeing the real world and it's complimenting it in a way, right? So I would have to go with like, I'm very excited about the glasses as like the next big form factor. I think I, and you know, it's pretty cool to see like Zuck take a very hard stance on it. And he's basically like, I'm going to do anything to beat Apple on this. And Tim Cook seems to be.
Mehul Nariyawala (53:18)
Mm-hmm.
Ritwik Pavan (53:32)
Kind of saying the same, too.
Ben (53:34)
I have a weird anecdote about this. I have kind of a weird anecdote about this, but the first person that I showed any robot that I built to, like the first really, really crappy prototypes was Brendan Iribe, former Oculus CEO, because he started his own smart glasses company.
Mat Vogels (53:34)
We'll see who wins, but.
Ritwik Pavan (53:36)
We'll see. We'll see.
Ben (53:53)
And I was very, very strongly considering going and working on, he offered me that head of ML role there at one point to sort of do some of the voice stuff for this glasses. I think voice is the killer app for glasses, for what it's worth. The demo they put out was really impressive, yeah, was just like, very, very, I was like, I was very strongly considering it. And then I kind of decided to do humanoids, but yeah, I'm.
Mehul Nariyawala (53:54)
Mm-hmm.
Mat Vogels (54:06)
Yeah, it is.
Mehul Nariyawala (54:08)
for us.
Mat Vogels (54:20)
I'm glad you
did humanoids. Yeah, because no one else, don't think anybody else would have taken that, I the approach or at least maybe it would have stretched the timeline a little bit longer. My BD unit would be, you know, decade further away.
Ben (54:21)
with this.
Mehul Nariyawala (54:22)
Yeah.
Ben (54:31)
Yeah, I decided Humanoids because I felt that smart glasses seem like they're going to happen without any intervention. Like it just is a pretty awesome product and you can see where it's going. Whereas Humanoids, the industry felt kind of backwards and I wanted to sort of make a change in the industry. Yeah.
Mat Vogels (54:52)
I agree. I also agree on the glasses front. To me, I think it's actually like not even a question that the final form factor for what we're seeing on the AI side until we get to like brain implants, which is probably more of the like final form factor, which are probably sooner than we think too. I think the glasses are no question to me. You just can't convince me otherwise. And I'm, love, I have the Apple vision pro and I love VR and AR and everything there, but
Mehul Nariyawala (55:07)
You
Mat Vogels (55:17)
There's no other device I think that brings together so many senses all into one while still keeping you in the real world. The audio, the visual, I saw there was a company the other day that had pitched a black flag where they were producing glasses that had like all these different smells in it. So you could like put them on to watch movies and it would have like these little fans that would like trickle into your nose to do like smells and stuff. ⁓ So maybe it's eventually too.
Mehul Nariyawala (55:37)
Heh.
you
Ritwik Pavan (55:42)
It's the first time I've
heard of a product doing that. That's a new one. gotta share that company with me so I can.
Mat Vogels (55:45)
Hey, it's a,
Mehul Nariyawala (55:45)
Yeah.
Mat Vogels (55:47)
Yeah, yeah, they had like 50 it
was like different chemicals that make different types of smells and you could watch something and it would have and it was funny I was in Disneyland earlier this week and I think of like soaring over the world and they have this the smell things coming into that's what made me kind of think of it again, but I think there's a combination the other thing I think where it fits into personal robotics is if you do have glasses on like if I'm with my Matic or for example, you know similar to VR now I'd like have my menu system and I could tap on But then I can imagine like being in my home and like seeing the path
Ritwik Pavan (55:52)
Yeah.
Mehul Nariyawala (56:13)
Mm-hmm.
Mat Vogels (56:16)
that Mo's about to go on in my house and like it brings this like visual and ties the robotic piece in there I think even better and even language all these things I think that's just it's definitely the form factor that I think will happen and I'm the most excited about too. Well I know that we're going
Mehul Nariyawala (56:21)
All
it.
Mat Vogels (56:35)
A little bit over I want to end a little bit with maybe quickly around the horn. What besides what you're building what area of personal electronic robotic hardware will open it up to kind of anything and maybe besides that we talked about the the glasses side. Are you most excited about over let's assume the next 10 years so no barriers all these things are going to happen you know energy costs all these things.
what's the thing that you look forward to the most maybe in 10 years? Rewick, let's start with you on the thing that you're most excited about as it relates to personal robotics or if you want to stretch into personal hardware.
Ritwik Pavan (57:11)
Yeah, I mean I think I'm most excited about self-driving cars. know, the CyberCav I think is probably what I'm most excited about. I used to like driving, but now I kind of hate it. Once you live in a city, you realize all the traffic. So yeah, I'd have to go with self-driving or like just like autonomous vehicles in general.
Mehul Nariyawala (57:14)
Yeah.
Mat Vogels (57:27)
Awesome, and Rebecca, now you have to jump off. Quick CTA on where people can find you, what you're doing, just in case you have to hop off here.
Mehul Nariyawala (57:28)
I think... ⁓
Ritwik Pavan (57:35)
Yeah, you guys can find me at at Ritvik Pavan. It's R-I-T-W-I-K-P-A-V-A-N. And you can also find me at like Matt mentioned Hardware Nation, which is on my ex or YouTube or LinkedIn. And my newsletter hardwareherald.com. Thank you guys. Gotta hop. Great chatting with you all. Yep. Yep.
Mat Vogels (57:55)
Thank you. We'll chat soon.
Mehul Nariyawala (57:57)
I was gonna say I'm with Ritwik. I'm actually quite excited about self-driving cars as well, partly because this will be the first technology since the internet that is profound enough to change everything around us. I mean, maybe iPhone did that too. Sorry, I should say iPhone as well, iPhone chat, GBA, all that, but I feel like the physical change that we're gonna see.
Mat Vogels (58:09)
Hmm
Mehul Nariyawala (58:17)
that around us based on how self-driving cars get adopted and how taxes get it. We're not necessarily clear on what that's going to look like. So five, 10 years from now, especially with me having my five-year-old, I don't think she will ever drive. That's my gut, but we'll see. We'll see. So that's really interesting to see. And I think that will open up lot of imagination in terms of how new robots or humanoids or any other devices will scale as well. So I think I'm pretty excited about that.
Mat Vogels (58:43)
When it's fun, the tangent industries, like you were saying, and both of you having worked on the self-driving car stuff at Tesla, how much that technology is going to push forward other areas of AI and just visual interpretation of the world. So I agree, if you package it there, it's great. I think I wasn't assuming that you guys would say that one, but it is an interesting topic because I think for a lot of people, might think that, well, self-driving cars is kind of already here, but it's really not, at least to the capacity of what it's going to be.
Mehul Nariyawala (59:09)
That's not it.
Mat Vogels (59:11)
⁓ which is fascinating.
Mehul Nariyawala (59:12)
Yeah,
yeah, it's it's we haven't seen the impact of it yet. It will change us drastically, change the world drastically.
Mat Vogels (59:19)
And do
you think that happened? mean, I know that the self-driving with Tesla and the robo taxis and we think of freight and autonomous vehicles there is it, I mean, it's, guess it's everything you're right. It kind of, it's a whole systematic change across multiple industries that will happen.
Mehul Nariyawala (59:33)
Mm-hmm.
Mat Vogels (59:35)
Ben, how about you to finish us off here with what you're most excited about aside from maybe humanoid robotics.
Ben (59:42)
I'm most excited about brain implants. I think this is almost certainly going to be much much better in 10 years. Personally, my wife is a neurosurgeon and she does these brain implants some so I hope she has good prospects. it's like right now it's mostly medical but you're getting if you if you ever watch a brain
Mat Vogels (59:45)
Mmm.
Mehul Nariyawala (59:55)
That's awesome.
Mat Vogels (59:56)
Yeah.
Ben (1:00:03)
like an implanted device for treating like Parkinson's or Alzheimer's or any kind of like these are there's a whole swath of diseases that are impossible to treat any other way and the impact that you can have from this is just life-changing. It's probably going to be the most important sort of medical device in history and it's a lot closer than people would would think.
And I think that's just on the medical front, the opportunities for integration with robotics or with augmenting your intelligence is just, I think that's gonna be the thing that really changes the world, Like changes what it means.
Mat Vogels (1:00:41)
And that to me, like I mentioned.
It's the ultimate form factor, I think, for all these things that we just spoke about. It'll combine everything to be able to think of the speed of thought as we're controlling things around us. Or, you know, I'd be sitting on my couch instead of needing a visual. I'll just think about, I want the Matic to clean this part of my house and it'll happen. And then it's, it's, it's going to feel so bizarrely natural. I think it's probably, it's gotta be a couple of decades away. think of maybe my grandkids will be in a world where that's just such a normal thing to
do. It'll be fascinating. It'll be definitely be fascinating.
Ben (1:01:16)
I don't think it's more than a decade away. It really, I think...
Mat Vogels (1:01:19)
Really? be, maybe it's more of like a,
to be available from like a price standpoint or maybe like an accepted, like, does it take a generation to think of like, it's, normal to have something implanted into my head. Cause my guess is anybody over like 40 would be very hesitant unless there's like real medical reason. Maybe, you know, 2030s you're, you're probably leaning with it, but then certainly younger generations, they grew up with that. They'll, they'll be good.
Ben (1:01:45)
I think it's actually not a technical problem is my point. It's a go-to-market problem. there's going to be like people that can be creative about go-to-market problems will be able to push the future forward a bit. That's what I tell my wife at least.
Mat Vogels (1:01:58)
I
love it. I love it. That's an exciting one. I'm excited about it. All right. Thank you guys for hopping onto this. Can we wrap up a little bit with Ben, maybe where people can find more about you, about K-scale, what you're building, what you're looking for if you're hiring right now, any final CTA.
Ben (1:02:17)
Yeah, my website is kscale.dev. We're hiring firmware engineers, ML engineers, hackers. And we're raising money right now. Probably going to try to do some fundraising in the next few months.
Mat Vogels (1:02:30)
Fundraising is the hardest part about a startup, period, I think. Even when it goes well, it's hard. yeah, we'll try to help as best we can there. Mi'hul, what's the final piece? CTA, what's going on at MADEC? What do you need help with?
Mehul Nariyawala (1:02:45)
Yeah, getmatic atmaticrobots.com. It's there. We have 10 extover sales in last four weeks. challenges of growing and things beyond once you find a PMF is where we are at. And we're pretty much hiring in every single class set along the way. So it's really just an exciting time when you finally get a product out and it's just taking off. And that's really amazing. And you can all find me at
Mat Vogels (1:03:07)
So great, love it. Well thank you both. Yeah,
go ahead.
Mehul Nariyawala (1:03:10)
Sorry, I was just
going to you can find me at at mehool on X, just M-E-H-U-L.
Mat Vogels (1:03:15)
And a great follow-up, some of the demos and things that you do on there are fantastic. So it's another way to continue to follow along with the product too.
Awesome. Well, thank you both very much. Enjoy the rest of your Fridays as we're recording this and have a great weekend.
Mehul Nariyawala (1:03:30)
Thank you.
Ben (1:03:30)
Thanks.
Mat Vogels (1:03:31)
All right, bye.
Mehul Nariyawala (1:03:32)
See ya.
