 Hi. Hi. Nice to see you again and this time face-to-face. We did the live remote chat video a couple years ago, right? Yes, it was in 2021. And please introduce yourself. Okay, I'm Rong Zhongli and I'm the inventor of OpenCat and we have the product as Nebo Cat and Beto Dog. We have proud founded both of the products on the Kickstarter. This is the first prototype, the original OpenCat. Nice. So it's for people who don't want to have a real cat, I'm joking, but it's like it's like a fully articulated cat-dog device. Yes, I'm trying to reproduce all those complex motions. That means I have to have so many actuators, springs and sensors and chips. Are you trying to copy the real animal with all the articulation? It's like they've been designed by nature over millions and billions of years. So they must be optimal in the way the legs and everything is. Yes. So actually I'm taking some inspirations from nature and I believe since they have been optimized during evolution, I think if I follow some concepts, it will be an optimal solution in engineering. Nice. And you have a lot of fans around the world, no? Since you launched? Yes. And I think it was a fan that supported me so that I can make my dream come true. What is your background? I'm a physics PhD and a master in income design. And after my graduation, I also taught in my university for two years. So I know the so I was able to derive some basic principles for the robot, not quite from the aspect of robotics major, but from the very first principles of physics and math. And I also know like how to introduce those concepts to people who don't know because I have been a teacher. So that also defines the features of our product. First is a very cool robot. Second, it has functionality as a teaching tool and also a researching tool. There's one thing I always wonder, which I don't understand how it's possible, that biology in nature, sometimes you have you have like dogs, they run so crazy fast and I don't know where they get the energy from. They just run, run, run, run. And you have birds like flying thousands of kilometers. What's the kind of battery they have? Where is it coming from? OK, that's a different kind of question. But you're trying to emulate that because you don't want to use too much power when they move, right? It has to be optimal. Yeah, so every energy comes from the sound, basically. And we store the energy into the battery because the limitation of the battery, we cannot store more power in the battery. So we have to optimize the design. One very important factor is the weight of the robot. So that means I will optimize all the structures that it has enough strength, but not too much weight so that it will reduce its own burden. And also we need to optimize the position where the actuators are positioned to reduce the burden on each joint to reduce the rotational inertial on each on the joints. Nice. So can we walk into your office right here? Yes. So we're here. Actually, we are during the holiday, but you're always working? Yes, every day, seven days a week, 14 hours a day. 365 days? Almost. You go to hometown, no? I do hometown. It's about 10 hours train, high speed train. Is that also where you study and teach? Is it in the hometown? No, when I was in the USA. In the USA? Yes, I've been in the USA for nine years. And where was that? My university was called Wake Forest. I've been there for eight years and then I moved to Pittsburgh to start my company. Cool. Alright, so let's walk in. So obviously your colleagues are on holiday right now. Yes, everyone. Can you introduce a little bit your colleagues? Who could it be here? Okay. This guy is for the structural design. This one is the office manager. This seat is for the BD, business developer, development, basic sales, and this is our software engineer. This is our electronic engineer. Alright, I have a quick connection there. This is our media producer who takes pictures, videos, and edit. This is our workbench for the hardware. All these inventories, soldering irons, it's quite messy. And we also have our water tank. This is the mascot? No, you already have the mascot. This is the cat and the dog. Is there electronic in there? Is something happening? Yes, it's filtering the water. And we have some baby fishes. Oh, they have the baby fish area. Nice. 3D printers. There was one desk over there also? Yes. Did you talk about this one? Yes, it's a mechanical structure. Alright. And so where do you meet your colleagues? Where? Yeah. How long time you work with them? Well, actually for the first three years, I was a women company and I designed all the stuff at the factory area. It's not like the city. It's basically factory area. I just lived in the factory and so focused on my development. Then two years ago, I moved back to Shenzhen and tried to form the team to work with me. Where was the factory? It's in Dongguan. Dongguan. So it's like one hour or something from here? One hour and a half? If the traffic is good, then it's about one hour drive. Alright. And so there's still the factory over there? Yes. That's where you manufacture? Yeah. This office is basically our research center. Nice. For the production line, it takes much more space. We produce thousands of the robot. Is it okay to film some of the prototypes that are around? You can take it and later I can tell you which one cannot be released. Ah, okay. Alright. And right here, you are checking some testing a bunch of stuff. Yes. And this guy, this is the secret? No, it's some people, some other people's product. And here, this is a special workbench for the dog bench. Alright. And this is your office in there? Yeah. Alright. It's quite messy. Sometimes I sleep here. Cool. Yeah. In China, this is a tradition after lunch. Everybody sleep a little bit in the office. Is that also here? Yes. That's true. Alright. So let me just put this right here. Alright. So you mentioned business development before. What's the potential for business development? Because you did an amazing project in crowdfunding. And what could be the potential, you know, next level for your company? What could happen? So originally the project was so focused on the maker. Maker sector, like people who know Arduino, who know Raspberry Pi, and basically the DIYers who has the knowledge and patience to learn. We were successful in that little sector, in that niche market. And this nibble was the product for that stage. It's made by laser cut wood. But now we face the problem. That is, if the product is too complex, it increases the skill level. It increases the barrier for people to use it. Well, my goal to make this company to run this business is to introduce robotics to the general public. So although it's very challenging, but we are still trying hard to lower the barrier and make our robot more accessible to the general public. So our second generation robot is like this one. If you put them side to side, you can see this one is more polished. It has all these injection molded plastics. It's more durable and it's much easier to assemble. And during the development, we also improved the software a lot. Previously, we used this, we draw this PCB. It's basically an Arduino with many other assistive chips and sensors on it. Is it the same chipset that's on Arduino? Yes, it's compatible with Arduino Uno. And this was the only chip we used for this cat. If you want to program it, you need to learn the Arduino. Wasn't it ATML or what did it do? It's an ARM Cortex-M3. It's a very classical simple chip with very limited resources. But I managed to shrink all the code into this. Cortex-M, right? No, not yet. It's not even an ARM Cortex-M? So, to consider its parameters, it has 32 kB flash. 32 kB flash and a 2K S-ray. Oh, wow. People usually use it to just light up some LEDs. But we shrink all the motion functionality into this chip. And later, because to program this chip, people need to know Arduino, which is a text-based coding arrangement. A lot of people feel very uncomfortable with it. And sometimes they need to load libraries. It sounds familiar for programmers, but it's not for the general public. So we improved the workflow. For example, include the libraries into our release of the whole coding package. That's the first step. After they download our package, they can run the Arduino program directly. And second, we even skipped the Arduino environment. We wrote a wrapper using Python. And we wrote a Python graphical interface in Python that people can upload from where they can calibrate the robot using some sliders instead of remembering the numbers for offsets. And we also designed a tool that can visualize the postures of the robot and can design them just like designing some stop-motion animation. And we also developed the smartphone app that people can wireless connect to it. Before it was using this remote, it's one-to-many. If two people want to raise their robot, it's impossible because it's basically broadcasting. But using the Bluetooth smartphone app, it can be one-to-one controlled. And they can design their own functional buttons, customize the buttons in the app. Arduino is a big community, no? Yes. So are you able to bring in a lot of the stuff that's happening in Arduino? Yes. One benefit of using the Uno compatible board is because Uno has been around for about 10 years. For 10 years? Yes. Uno has been around for 10 years and people have generated a lot of content, tutorials to utilize this chip. And we have all these sockets that can plug different sensors. And we actually have been selected as the most popular Uno project during the 10 years. There were 12 of them and we are among the 12. Wow. That's cool. But one thing I wonder is one of the trends in the industry is computer vision, AI, self-aware little robot thing. Is it possible that you want to consider getting to a bigger chip? Yes. We solve this problem from two directions. One direction is to increase the processing power of the board. And we are now having another board called the Byboard. If you know, if you are familiar with this, it's an ESP32 board, a very popular board in the IoT. It has the built-in 1S connection, Bluetooth and Wi-Fi. And it also has much faster CPU and larger programming space. And we have already released this board. It's been mass produced and the firmware is also ready for people to use. But still, this board is only for motion. It can feature, oh, maybe you can get up. This is the board right here? Yes. ESP32, a famous board. It also has been around for more than five years. Right? Yeah. So this board is much faster and can feature smoother and faster motions. But to stance the environment, we need more sensors and processing power. So we don't try to put everything onto a single board, but try to separate them into different layers. It's quite bionic like it has the brain that can process all the sensors from the eyes, ears, touch. And then the instructions are sent to the spinal cord and the little brain for coordinate all these motions. Basically, this board is like the little brain. And for the brain, we have a separate processor. For example, we have an intelligent camera. By itself, it can recognize people, shapes, symbols, and it can return the conclusion through this cable. And our main board can process some feedback loop after receiving the conclusion from this camera. And also, for example, this one is a voice recognition module. It can recognize more than 40 voice commands and it has 10 customizable voice commands. All right. Yeah. I'm trying to focus right here. All these chips have some AI calculations on it and then send their conclusion to our main board. So for example, the computer vision one is just basically you put the processing at the level of the camera. Yes. This one. This one has our logo on it. This one has our logo on it. Yeah. This one is a new sample we are testing. So this could be its own chip going on here. Also expressive. And so the idea is modular processing where each organ or something has its own chip and then it speaks through communication. You can add processing that way. Yes. And it also brings the benefit that after like in the future if we upgrade the chips in any section, we don't need to worry about the compatibility because they share the same communication protocol. Is a new board compatible with the previous design? Yes. You can just swap it in. Yes. This you can see very clearly both are using the same frame and both are compatible. Nice. That's awesome. So your dog and your cat is your pet is future proof. Yes. And because we are mass producing them, we try to use the same component even for different models like they share the same parameter of the springs. This is a very tiny aspect, but it's also a workload for the inventory management. So we try to use the same component and also for this battery because we customize this battery, the MOQ, the minimum of order quantity is 5K. So we have to make good use of these batteries. It's a good battery? They share the battery. The chip iron and everything? Yeah. It's quick release. And do people buy a kit that can have extra battery? Yes. We offer the battery as an independent accessory. People can buy as many as possible. One battery can support for about one hour continuous working. And if it's sitting, it can be there for more than two or three hours. And charging takes about two hours. Nice. So a couple of years ago, I started seeing these Boston Dynamics. And they're so fascinating. I wonder, have you seen if people have done really useful things? Is the most useful example is a SpaceX. They send it in to check the launch pad after a launch to check there's no bad chemicals in the air before the humans come. But do you see any inspiration from that kind of area to bring something smaller that could do a bunch of useful tasks? So for the useful tasks like what has been advertised by Boston Dynamics, for example, patrolling or inspection at factory, or some things after the deletors, or some general industrial or transportation tools. All these applications, in my opinion, are a long future to come. Because all those demos they should currently they always have an operator behind it. It takes at least maybe 10 years. At least 10 years so they can be fully autonomous in these applications. So I think I consider usefulness from a different way. Both of Boston Dynamics and me are making robotic animals, robotic dogs. Why is it not a wheeled car? Because we both see the value in its animal shape. Animal shape means it has legs that can work through different difficult terrains like stairs or even spaces. And it can roll recover, it falls back. That's the functionalities that a regular car doesn't have. But besides that, I also see the real meaning of animal in our daily life. Just consider the relationship between humans and animals during the long evolution into the civilization area. How we interact with the home animals, how we collaborate in our household. So there are basically two types of animals. One animal that can be the labor for humans like horse, cows, that can be the engine in production. Yeah, donkey or transportation to like horse or dogs who can be the guard for the home. That's the real helpful animals. But there are other kind of animal, we call it pet. The pet means it may be not doing any meaningless things. For example, cat is basically eating your stuff but not giving you respect. But people just cannot get rid of them because we humans are emotional. We have the need to interact with another form of creature. We are social animals, we talk to people. But sometimes we just don't need another people but some form of intelligence that can communicate with us but not as smart as us. So that we can feel less lonely. I think my little robots are in that application. That is a little moving robot on your table. And you can interact with it in some forms like programming. That's for technicians who can define all the behaviors of the robot. Or it can be just plain cute using some others already written program just to make you happy. Because there is this role that I think dogs, for example, bring better health to their owners. Especially when they have to walk with the dog for hours and hours, nearly every day. But also it triggers something in the psycho-conscious, in the brain and stuff. Psychology, philosophy, there is some... a dog is never going to be against you. It's always supporting you. Potentially you could have small functionalities like I'd like my pet to say give me notifications when I have new email. Or when I have like there's a new show on Netflix or some kind of like if this then that kind of things. Potentially there would be a whole range of useful cases. Maybe it starts barking when my wife is approaching the house. So to warn me a little bit and say I need to finish cooking or whatever. Or wake me up in the morning at an alarm clock. Is there any consideration in adding those kind of functionalities like a smart assistant kind of thing? Yeah, of course. As long as the robot can handle itself. Meaning it can move on its own without the help of people. This cannot be done by some regular cars like if the car flips it cannot roll back and it will cry for help. But for our robot if it falls over it knows it's a beauty instinct it will roll back. And it should have a very good understanding of the environment. That means it will have a bunch of sensors like we are making this prototype just to integrate all the sensors into one big head. What sensors are you talking about there? Like the light, distance, sound, heat and so on. It's also modular we can add more in the future. Where do you add the extra sensors on the robot? Head. You just go on the head? Yeah, of course not this little head but a much larger head. Bigger head? Yeah. For example when you showed this before the camera module it would also go somewhere up there? Yeah, before when we designed this robot is still a bridge between the programmer and the general public. So we don't consider too much about the integrity. We basically let the dog to bite on the sensor. And for teaching or researching it's enough because every time people use it they just focus on one module. Like with the camera it makes one class with another sensor it makes another class. But if we want to let more people to enjoy the robot we have to consider about the integrity. That's why we need to put all these sensors into one so that it's almost like a smartphone. You don't need to consider like I need to plug in the camera today. I need to plug in the microphone today. It's just a built-in and after this product is built people only need to work on the software side. And you said it has to work by itself. It'd be nice if you can plug itself into charge. That's one feature feature. To make that happen it will need to have the planning. First the understanding of the floor plan of your house. And then the planning how to return to the charging station. And we don't think this technology to be too difficult. There are so many projects already around just controlling a car. And using the lighters or double camera for special slam. So for us because we are a small team we don't expect to build everything by ourselves. Our key feature is the motion ability on the legs. So we are now more focused on making this platform stable. And for the other features such as the human recognition or the slam we can just use in some existing modules. I like listening to Elon Musk interview sometimes. And he said that in the industry there was some issue with actuators. That nobody was really optimizing very powerful actuators to do his humanoid that he wants to do and stuff like that. What's the status in Shenzhen where you source all these components? Like new generations all the time and the latest one is really good at doing exactly what you need for the weight and everything? So when it comes to robotics there is no correct answer. Or all in one solution. It all depends on how you design the full system. It's really a search and experiment to see which servo or which motor fit in your robot. For us our strategy is to use some traditional solutions but with optimized parameters to maximize its performance. And also on our side we optimize the body frame to make it light. And to optimize the algorithm to reduce the computational power to reduce the need for batteries and so on. So that whole system is so compact that it can be holding on your hand. And for those larger robots their strategy may be different. Sometimes you have to partner with some top tire of the manufacturers who are always exploring how like the best performance of structure. But in that sense that means it will invest a lot of research and development funding into maybe just 100 units of a certain structure. They are very cool to look at and let you know how far human technology can go. But between that technology to our daily life there will be a very long time to come. Would you say how many fans and how many users you have in the world? So far we have sold more than 10k units worldwide. It's quite a lot no? Sometimes I guess one customer is like a whole class because it could be the teacher. It's actually times 30 could be 300k. But it's not how many of those go into education? How many of those are the parents, geeky dads giving their kids or playing it with themselves? Who do you think your customers are? It's a mixture and sometimes it's hard to give a very accurate ratio of those. But yes the kind of users you just mentioned are the typical users. For example a teacher or professor who may put an order for 10 units for her class. And for some engineers, technical guys who buy the robot. And they may explain to their wife it's for their kids to know programming. But maybe it's just for him who really loves it. Country these are the two major users. One is from the educational researching aspect. One is for the individuals who likes programming and robotics in general. Where is your community kind of like forum and a host where people can interact with other users and ask how to fix something or explore new ideas and everything? We are on the social media. We are on YouTube, on Twitter and Instagram. And we also host a user forum that's hosted on our official site. It's called patoy.camp, P-E-T-U-I-DOT-C-A-M-P. If you enter that address you can directly go to the forum. So would you say you might be one of the more successful robotics education projects in the world? Or are there some that I'm missing here that are huge huge? Like are there some education robots for example or that are selling in millions or it doesn't really happen? It's not happening yet. For us we are still during the stage to transform between a hobbyist project that meets the needs for geekers, makers to some industrial or educational applications. So for us if we can expand the applications in education that will already give us maybe 10 times or 100 times of growth. So this is also our current focus to make it more general to meet the needs for people who has the need to learn, who has the patience and the funding to learn. And after that if we have reached the ceiling of that application we may explore some higher or more general applications. For example some programming tools, even replacement or alternative of pets. But what's the biggest project in the industry in your market? The biggest industry. The biggest, most famous kind of like competitor to you. Do you want to talk about it? Like are there any ones much bigger than 10,000 people or not so many? It doesn't really, you're one of the bigger ones. Well in terms of units sold, I think Sony Ivo has more than maybe 20 or 30k units sold worldwide. But that was since 1990s. Yeah that was an amazing product for this time right? Yeah and it was also an inspiration for my project. One very interesting thing I discovered was after I returned to China when I was cleaning up my room I found my old diary. And on one page that was on July 5th about 20 years ago. I wrote, oh today I saw a robot on television. It's a robot dog. It can really walk. That's amazing. 20 years ago. And my company happens to be founded on the same day July 5th in 2017. Yeah that's a very great coincidence. I didn't know that. One thing I heard is that the Ibo's, some reason Sony didn't want to make millions of them. And so they stopped making them. But like the old people's homes in Japan, they look for them everywhere in the second hand market. And they've been paying a lot of money to buy them even though they're several years old. Because these people in these elderly care homes, they just love having that company to have the little Ibo dog walk around and do little noises. That's another evidence to show the companion robot is the real market. Maybe it's not for everyone but it has the user base. For example the lonely elderly or lonely young people who live in the city area who doesn't have time and space to take care of their pets. And there's a good news. Sony has resumed the Ibo product about five years ago already. They have their new Ibo model and it's currently being sold. Brand new design. It's cost a few thousand dollars right? Three thousand. Three thousand US dollars. And with subscriptions, you have to pay subscriptions to keep it active, to keep it intelligent. That's a lot. That's funny. But how many do you think they make of those? How many? Is it the small quantity? I think it will also be counted in tens of thousands of units. All right. So that's in the dog cat-like style. But how about the other style of the educational robots? Would you say what is the most successful in the whole industry? How far are you from them? Which one could it be? Is it like Lego Technics? Yeah. That kind of stuff? Yes. This is maybe the top most popular platform? Yeah. But I also heard Lego has stopped the Minestorm product line. Maybe they are not making enough money for that, even for Lego. Yeah. I need to call them up over there in Denmark and say, hey, what's going on? But of course the big challenge is to bring this, as you were saying, to millions and millions of people. And to make a platform, if you can do that transition, your company is completely different, right? Yes. And so this is what you're working on, to try to see how you bring this to the next level. And who knows? You say it's maybe not for everyone, but everybody I think gets notifications on their phone. They spend way too much time on their phone. And they probably need something a little bit nicer at home. Different shape. It's more interesting than the phone. And the key factor, of course, is still the price, right? If it doesn't cost any money, it would be cool to have a working gadget on your desktop. So our problem to solve is how to zip all those core functionalities into your platform that's affordable to everyone. And easy to use for everyone. Because when we talk about the Boston Dynamics, isn't it 80,000 or 100,000 dollars? Maybe, maybe even 140, I don't know how much, but it's quite expensive. And then you are doing something, what is the retail price? It's about less than 300 US dollars. So that's like, if it's become very, very useful for many, many people, it will be no-brainer. Yeah, it's already been kind of popular gift during the holiday season. We see the increase of sales because people are searching for robot dog toys during that holiday season. The problem for us is still like how to make it very convenient, easy to use. For people who are paying for 300 USD, if they cannot fully explore its features, I mean all these open source programming features, it's still too much for them. But if they can enjoy the programming possibilities, then I would say it's the best option they can find on the market. Because we are not making a lot of money over the boom, the build materials, but really trying to promote this kind of robot. So what I'd like to see is, for example, when I'm sleeping, it should go and sit and wait in a bedroom and wake me up when I need. And then it should be very smart at staying out of the way. Like that's one thing that I'm amazed with cats and dogs. They have such ability of, you know, I'm always worried that people step on their dogs, but somehow the dog is always staying away and they know how to stay away from these huge humans. So it needs to have that capability. And then when you go to the living room and you sit and watch TV or something, it should be following you over there and do some smart functionality for you over there. If you go to the kitchen, it follows you over there. It should be always kind of like following you around and always ready to help with something. Or just be there for you like a little friend. How far are you from potentially doing that? Because do you need to have the full computer vision? It's just going to be on the little module for the camera like this. Everything will run on here. Or could you potentially have a cloud processing that helps you in real time, you know, avoid obstacles, step over something and go a little bit to the side, avoid the human and then always know where they are in the house. We try to avoid cloud computation for now because one thing is the latency of the network connection. And second is we are building kind of digital life creature. And for creature, the one key feature is the independence of its own. Like as long as we can find food, we can live on our own. It's the same concept. If we have enough power in the battery, the robot should be able to run it on its own. Even when there is an outrage of the whole city's network, it should still be working. So we try to make all the computational locally. And for the need, you mentioned like how the robot can follow you to be your companion or to provide help. If we just develop this functionality as a demo, it could be accomplished in a few months. But our problem is just making a cool demo is not really making values to the users. We have to make some completed functionality integrated into a product. That would be a very hard problem to solve. There will be a lot of edge cases. Just to give you an example, like I made this prototype of the robot in about two or three months. It has all the four functionalities of working around. But it took me five years to polish it from this version to this version. And to optimize the software, the user interface, user experiences, and so on. It's very different between making research in university, which I was familiar with. You just write papers saying it could be used for what, what, what. And it will be accomplished by some magic magicians, like people who happens to have the ability to turn them into the product. But the fact is, although we have made so many statements in papers, in academic, but very few of them can really be turned into a really useful thing for the general public. And since I'm so into this technology, to this robot, I feel the responsibility that I really dive into the industry and to polish every aspect of the technology as a product that's easy to use and affordable for everyone. And it can be nice if you have 10,000 people working for you, like all your users who want to upstream improvements, like software capabilities, and they could be contributing back to the ecosystem and have kind of like applications. Do you have something like that, applications, or how do you add, I want this module, this, and just click and add, add, add, add this functionality. And then, oh, this thing has bugs and some other guy somewhere else, like in Italy or in Brazil. And it just goes in and fixes that and, you know, submits it back. Is that the vision? Yes. So by the framework of this software, it's already possible because we are open sourcing the software on GitHub. It's open source for basically for everyone who uses this robot. They have access to the code. They can change any bit of it. And some of them will submit changes, submit bug reports to it. And some more general public users, they will write to our support email. And we can recognize the bug and we can provide the first we direct them to the FAQ where we have a detailed debugging routing to locate the problem. Then we either to give them a remote session to fix it or we provide replacement for them. So far, although we hope that the technical users can submit modes or packages, but still I think it's based on the probability and the people skill set, like people who are able to understand the whole code, maybe just a little portion of the whole user. And the total user base is not as large as the smartphone. It's just 10K. That's it. So maybe we only receive one or two uploads from the community every month. What if there's a big Sony competitor or something like that who thinks, Oh, I really would like to launch in partnership with you something in a bigger scale. You open to that potential? I actually have received a few tycoons in both China and the US who said, Well, I would pay you a certain amount of money and you build a robot that meets our definition and will have our brand on it. That means we will be the OEM manufacturer or ODM manufacturer for them. Every time I would consider their offers and their requirements, whether it's applicable and whether it can help us to grow. So far, the featuring posters or featuring motions are kind of, how to say, symbolic features of our product. But if we make this available on those bigger brands, then they will have the benefit of their fame. They have their marketing channels that everyone would think, Oh, it's made by those big companies and later it will really harm our own brand. So I'm not very open to that opportunity. Because what I wonder is, it seems you have an amazing office with some, I haven't met them because they're on holiday, but nice guys that potentially you can do a lot and a small team. But sometimes I wonder if suddenly somebody comes and says, I will add 100 engineers with your project, but maybe that could be like one way to get to the next level somehow faster. I guess sometimes if you are a bigger team, it's hard to have everybody as efficient and then the efficiency turns around and suddenly that's why I think many big companies, they don't do much anymore. It's weird and it's hard to be innovative for them. And an example is once a very large company in China who is in robotics, asked me to make a robot product for themselves. The funny thing is they have thousands of engineers and maybe half of them would be PhDs. But the thing just happened, they cannot handle these many people to push a project forward. There are so many opinions around and the project, there will be so many conflicts. Just taking example, regardless of politics, regardless of the relationship between people, just think that you have five good engineers and each one is handling one aspect of it, structure, electronics, software, industrial design and the general functionality definition. Everyone, they may be the expert in their area. Say everyone is trying to make the best out of their skill set but there will be conflict of some interfaces or something. They may not be that willing to compromise for the benefit as the whole. When I was designing this robot as a one man team, it's very fine for myself. I may have already made seven iterations of the frame but later when I realized it cannot fit, I can just abandon all those previous work. And that would be unacceptable if it was done by one of the top engineers in that team. So in that sense, if you have a smaller team, you can have better communication and understanding of each other. And that can be more efficient than a huge team. When I see your product, I'm just thinking, oh my God, there's so many potential things that could be added. And I'm thinking it's probably hard for even if you guys are the best engineers in all of Shenzhen, it's probably hard to do all these ideas. Because you want to have, for example, perfect microphone array. You want to capture the sound like these Google speakers and everything. You want to capture the sound from the other room. It needs to be clear all the commands. And you want to have decent speakers so you can speak back, at least bark back something meaningful. And you want to have computer vision with two cameras. They both need to be 4K, 60 or whatever. And then you need a huge ARM CPU and a powerful GPU and a neural network, everything on the dog. That's kind of like... That's a huge amount of work. That's far away from where you are, no? Yeah, that requires a lot of engineering and R&D work. So I just said that from our current stage, it's the best option just to increase the efficiency and the communication between team members. But if we got the chance to work with more people, better people, and I would be also happy. So for now, because we don't have so many people, we have to make choices. Like vision is cool, voice is cool, blah blah blah is cool. But what are our core values? We have to make that choice, that decision to focus our limited resources into one that can make us unique and to make this product to be sold. We have the chances of getting investment and we are also opening to it. But I think besides telling cool stories that can raise funds, I think more real value comes from the market feedback. Like I quite enjoy getting the money from our real customers who are willing to buy out from their own pocket which money could be used for their better dinner, better movies, but they are willing to pay for our robot and after they receive it, they use them in their real life and give feedbacks. In that way, I can ensure that all this money is being spent on the right direction rather than like I have my belief, I have a big picture and I tell this to the investors. They are capitals. Of course, they can decide, spend millions of dollars into one company. But many cases happens when the story, the belief don't work in real life and in that way the resources of the society is wasted. Yes, like you want to be a real company based on real usage out there. You don't want to be just like one of these Silicon Valley startups that get valuation of a billion and they don't ship anything. But on the other hand, sometimes the consumer needs to be poked in a direction and the poking happens with something that has to come from maybe some kind of investment, I guess. But the best way is if you can invest just what you make selling your product and if that's enough to make a big leap, then that's amazing. So it's all about balancing. It's about the pace, how much you make money, how much you raise money, how much you sell. I'm kind of conservative in that process. I take every step firmly. It's also like when I was a child, I learned all those fundamental science concepts, calculations very carefully and then I can ensure in later decisions there won't be any mistakes when I'm doing more complex calculation. Running a business is also like doing math. If some of your first early principles are wrong, then it could be totally wrong. I guess you're working on the holiday. It's hard work to run your own company. It's really hard, but how do you enjoy this Shenzhen life compared to the USA stuff? Because my feeling is Shenzhen is like the city of opportunity. It's the place where everything is possible in a way that nobody, nowhere else in the world. This is a very special place. When you are saying so, it's almost like talking about USA 100 years ago. People would say USA is full of opportunities and so on. For Shenzhen, I think it really shows some switching between the center of manufacturing of technology from one side of the world to another. All this happening, I think, all depends on people. On Shenzhen, there are so many people who are sacrificing their living quality, like the out job time or weekend time. They are just so focused on their work and trying to make something moving fast. There's something called the Shenzhen speed. If I draw a PCB, I can get it in 48 hours or even 24 hours. If I want to discuss a problem with the manufacturer, they would drive maybe two hours to your office and stay here for the Saturday night. If this happens in the USA, sometimes it's considered as unacceptable. People want the balance of life and work. I really appreciate the concept that balance of work and life. My parents and my friends also appreciate it. That's why they also gave me pressure that you shouldn't work so hard. But things just happen when you are at an opportunity, like the time window for technology. How do you say? You just see the chances that you can make something happen. It's not related to how much money you are making. For example, I'm losing money now, but I just see that I have all these imaginations of how this robot can evolve. The only limitation turns out to be myself. If I could work harder, I could work faster with my team, then it can come out faster. I can validate that it's really useful. People really love it. That's the passion that pushes me and almost locked me on this seat in this room just to keep working. If you can catch the right opportunity at the right time, that's how you change the world. This can happen in the city. If you do the right thing, the right time, the right strategy, if you catch the wave, you can change the world. Changing the world is not my purpose. I believe that the earth can rotate without anyone. It's more about the feeling that you are making useful stuff for people. The most enjoyable moment is not when I'm counting the income, but when I see users who post their videos on how they are using their robot in their home, in their lab, all around the world. I haven't been to, but my robot has been to. That's the best moment.