The Robotics Stack
00:14:59 · Oskar, Xara
[00:14:59] S0: As I said, $10,000,000,000 market. I mean, if you compare this to some of these other companies that do all sorts of models, but have yet to shown a compelling use case, then this is definitely much closer to a real market than, let's say, as you said, a robot that can autonomously do all sorts of tasks, like a generalist approach. [00:15:19] S2: Yeah. Yes. So so we wanna focus on vertical, but eventually, maybe it will... We'll go the general direction way. But generalist robots require some time. Right? So far, you guys probably have talked to a lot of robotic startups and founders. We believe there isn't really a chatty moments in robotics yet. There isn't really a single robotic model that can unify all of robots and make them intelligent enough to do a task in all kinds of scenarios. But doing this kind of vertical things like intimacy, we believe is a area or it's a type of business can actually go in full cycle. It can actually become a profitable... It's a sustainable business. Even though there's so much interest in robotics right now, we haven't really seen a robotic company that's really, really profitable. So we wanna become one of the financially healthy robotic company that actually have a sustainable business model, and that's precisely the reason why we choose Intimacy.
00:15:19 · Xara, Oskar
[00:15:19] S2: Yeah. Yes. So so we wanna focus on vertical, but eventually, maybe it will... We'll go the general direction way. But generalist robots require some time. Right? So far, you guys probably have talked to a lot of robotic startups and founders. We believe there isn't really a chatty moments in robotics yet. There isn't really a single robotic model that can unify all of robots and make them intelligent enough to do a task in all kinds of scenarios. But doing this kind of vertical things like intimacy, we believe is a area or it's a type of business can actually go in full cycle. It can actually become a profitable... It's a sustainable business. Even though there's so much interest in robotics right now, we haven't really seen a robotic company that's really, really profitable. So we wanna become one of the financially healthy robotic company that actually have a sustainable business model, and that's precisely the reason why we choose Intimacy. [00:16:05] S0: It's actually not such a new business. I mean, if you combine it with robotics, yes. But if you think about Realbotix or a lot of these other companies, RealDoll, these actually go back, like, I don't know, almost ten years, twenty years sometimes, you know. This has been around for a while, so it's not been, like, having this huge curve yet, but I would agree that if you enhance this kind of technology with the robotics capabilities that we see now, this can be an interesting add on, basically.
00:16:32 · Xara
[00:16:32] S2: I totally agree. So Realbotix is one of the OG, one of the oldest company in the space, and they've around for, like, I think fifteen, twenty years. The technology stack today is very different from fifteen, twenty years ago. So from my understanding when I was studying robotics is that it's actually a Canadian company. I always had the fortune of visiting company at some point in the past. And the robots only have a moving robotic face, but it does not have a moving body. So you you see on exhibition is moving his hands, but it does not have whole body control. If a robot needs to be whole body control, it can switch from any intimacy positions to any other positions. The most important thing is to do reinforced mirror learning. It has to be powered by AI. If it's not powered by AI, it's impossible to do very complex move. Because when you look at our robots, its body has roughly 25 different motors, QDD motors. So if if you want to make the robots intelligently move in any kind of direction or form, it needs to be controlled by AI. So I think one of the biggest difference being now versus robotics back then is that, back ten, fifteen years old, there really isn't any kind of AI that's controlling robots. Right? It was probably more traditional control. And second thing is that there's actually breakthrough in materials that was not possible before. When you look at Realbotix, the whole thing is still silicon, but we barely use any kind of silicon. In fact, our whole robotic parts, the muscle is three d printed. So using using this three d printed muscle, we're able to make the robots very, very light. And also on top of this, let me show you the full stack of what's happening that makes this possible. So so when you look at it, the robot has three parts. The internal skeleton, that's in this middle is actually the muscle parts, the three d printed parts, and on top it is the skin. So the skin is actually very, very thin. The thing with traditional STIRCON is that STIRCON is very, very heavy. That's why the dolls are forty sixty kilograms. So using three d printing technology, essentially, the whole thing still feels like a muscle texture just like human, and the skin is actually a very, very thin layer. And that's why the whole robots, everything together is only 20 kilograms.
00:18:27 · Xara
Versus something that's 40 to 60 kilograms like a doll or a 40 kilogram Unitree, it's a huge improvement when it comes to weight and reductions. Because we are one of the first robots on the market that actually interact intimately with humans. We believe that making light is very very important. And I wanna add a note to this is that when we build our robots, we wanna make sure it's very distinctive from a dog. So when we look at a dog, obviously, it looks like a human. Right? But we actually don't want it to look too much like a human because we don't want us people to see it as a upgraded or a better version of a doll. In fact, we wanna build a new species, a new type of being. Right? Somewhere between a human and machine or a combination between the two. I'm a huge fan of a lot of different sci fi movie, you know, Blade Runner, and I play a lot of sci fi games like Cyberpunk twenty seventy seven, my favorite game. And we wanna combine a lot of robotic or sci fi elements with human elements. We have to realize it's actually a consumer electronics. So there's power button, and there's also ports where you can actually do charging and put in some kind of latest data and software. There's places where you can probably switch out a battery pack and so so forth. It's a balance between what's possible with engineering and what's possible with aesthetics. We need to make sure it looks good too. So the whole thing comes out in a way that we constantly remind user that it is a robot. It's not an upgraded version of a doll. It's distinctively a new species, a new type of being. And the reason why we have this kind of three layer architecture, let me share this with you, is that we wanna build the robots in a way so that the skin and muscle can be changed. So once in a while, maybe you wanna switch to a different face, different skin, you can do that. You want a bigger breast, you want a bigger butt, you can do that. So the internal skeleton is stable. You only need to buy one skeleton. But the muscle, the face, even the private parts can be changed to any kind of configurations that you want. You can customize it. Because at the end the day, it's a consumer product, and people all around the world has different preferences for beauty.
00:20:19 · Xara, Leo
No matter it's a beautiful robot in human, female, or male form, we wanna make sure we customize this to different area and regions that are preference for for beauty. [00:20:27] S1: It's pretty much like changing your phone's case or background. [00:20:31] S2: Exactly. So that's a good analogy. Yes. [00:20:33] S1: When talking about movies, this really makes me think about a combination of her and ex Makena, if you... I've I've seen both probably. [00:20:41] S2: I've seen both. I'm a huge fan of both movies. Incredible movies. A lot of inspiration in our company. Yes. [00:20:45] S1: So it's like the internal body skeletons of the robot is what you manufacture with all of metal actuators and so on. The muscle part is actually not exactly muscle in an anatomical sense. It's more like it will shape the body and then the skin will make the appearance. Right? The muscle doesn't have any impact of on movement.
00:21:04 · Xara
[00:21:04] S2: I think that's a good way to describe it. But let let me add more notes to that. So in order to create the most human like muscle, we have to study human anatomy because humans, we are a very complicated species. There's over 206 bones. We have 652 up to 800 different kind of muscle groups, depending on definitions. Just on our face, our face have 40 different muscles... Muscles. Right? So in order to create the most human like robots, we we try to three d print every single one of the muscle parts in the shape and form of the real human muscle. We try to do a lot of simulations and physics on the muscle parts. So humans are incredibly complex with so many different parts. When you look at us in detail, we we have over we have over, you know, 26 bones, 600, up to 800 muscle different definitions. Right? Very, very complicated. So in order for us to do this whole entire thing, we put a lot of work into the muscle. So in order to understand that how to build the most human like robots, we actually study every single muscle on our body, and we try to create cell map, our air stream muscles. We we turn to a lattice structure so that later we can three d print the muscle. Then we do models to do physics simulations to try to make sure the physics of hobby is very similar to a real muscle, then we three d print the muscle. So to answer your question, we actually need to understand how every single one of our muscles works in conjunctions. Also, try to make sure that the the feeling when the user is feeling the muscle, it actually feels like a real muscle. And this was not possible before because this kind of very complicated muscle physics simulations was not possible with earlier technology. Right? Using, you know, the latest GPUs, using a lot of paracomputing solutions, we're able to understand the human muscles really well so that we can three d print it away. It feels like a real muscle.