speaker-0: Yeah. This robot can be the coolest thing in the world. The nicest thing in the world? But if the unit economics doesn't make sense for someone to pay for that, nothing makes sense. We've cleaned and inspect more than five million square feet of facade in the last 12 months. The hardest thing to copy is not one component. It is the entire field and the operation system. I was in the C and the NVIDIA event and you see it all over the time. You see these videos of robots that are doing boxing match. This is why you invested so much money and so much resource to build robots. And this is how you choose to promote it? Solve the problem.
speaker-1: Hi, I'm Kim Ismerg, Super Intelligence Editor in Chief, and today we are joined by Ido Genosar. I hope I pronounced your name correctly, co-founder and CEO of Verobotics. Welcome.
speaker-0: Welcome. Thank you very much for having me today.
speaker-1: Such a pleasure having you. So let's jump right into it and right into the question. For someone seeing Verobotics for the first time, what problem are you solving in one sentence?
speaker-0: Okay. the robotics replace dangerous and inefficient manual work on building exterior, working on height with a lightweight robotic platform that clean, map, inspect and documents the facade.
speaker-1: That is very concise, very precise, Peter.
speaker-2: So Ito, you you came to this this business originally through construction, facades, building management, not from a robotics lab. So what did you see in that world from which you came that robotics first-time founder would have missed?
speaker-0: So my family has been in the facade business for the last forty years and I had an active role in the company on the innovation side. And I think that probably the biggest understanding that I had was that the problem wasn't the person that is sitting inside the basket of the BMU, which is the billed management unit. It's the huge crane that you see going outside of the building. The problem was the existing infrastructure and the access itself that was broken.
speaker-2: So can you can you clarify that a little bit when you say the the existing infrastructure and access, what does that mean?
speaker-0: Yeah, so today when you are doing exterior job for building, you either do it with a crane, which is basically the building management unit. You live in New York, Peter, so you see these cranes everywhere. And the other way to do
speaker-2: Building I can see the the I can see the Chrysler building right now.
speaker-0: Yeah, so you either see the cranes or you see the scaffolding everywhere. And the other method is with trope access. It's people that are literally rapping down the building. And I think that my perspective from the business was the understanding that every building is different, every roof is different, every facade is different, every access is different. And you can solve the problem using the existing inca infrastructure the building has. And for me this is something that I figured out quite in the beginning and that we have to change the access to the building. While in contrast, if you look at other solutions today in the market and other that operate before, they said, Okay, we have a platform, we have the crane, we wanna improve the safety and the efficiency, so instead of putting a person inside the basket of the crane, let's put a robot. But the assumption, to my opinion wasn't the right assumption because the problem wasn't the person in the basket. The problem was the basket.
speaker-2: Right. I I'm I'm curious though, if if that's the case and you're saying the problem was the basket, like I personally I'm not afraid of very many things, but I don't like heights. So when I think about like being one of the guys on on one of those rigs going up and down the side of a building hundreds or thousands of feet above the ground, that would not be something I would pick. It's very dangerous, e even though those guys are all s connected by climbing ropes. So how much of what you're doing, what you're solving is about keeping people away from harm versus improving the speed and the cost. And then maybe another aspect is like how does it affect the tenants of the building or are they aware of this at all?
speaker-0: Yeah. So safety is the starting point, right? But eventually the economics it's what really make the adoption and the scalability. we are moving people away from spending hours exposed to the side of the building, exposed to wind, other elements and of course the risk of walking on heights, right? I would also say that it's not that there is a lot of casualties every year. There is a lot of accident. But the ev effect of the accident is big, right? Because you involve the fire department and the police and the insurance and the liability and the tenants. So it's a big headache, right? And our goal is really to avoid as much as possible of having people dependent. And at the same time it's to reduce the fraction with the building, with what we're doing. on the tenant perspective, for them Again, I can also say what on my my perspective that I'm traveling a lot and you know every time that you go to a hotel in cities like Tokyo or Hong Kong or New York or San Francisco, most of the time you'll have some small note on the table, be aware that they're cleaning the window, for your privacy, close the curtain. Right? So I think that for us the fact that no one is over there and suddenly arrive from somewhere is something that it's nicer. It's even cooler in a way on the on the perspective of the customer and Going back it's really around the safety, the efficiency and the visibility for the biddingers.
speaker-2: Got it.
speaker-1: Okay. so my wife and I we just bought a house and obviously our winners are Dirty all the time, we have like a twelve month old baby, a son, so he's dodging all like all the windows all the time. And as far as I know, there isn't really much of like autonomous cleaning robot. So why has and especially like high rise facade cleaning and inspection been so difficult to automate? So I just brought up this personal like information because I know that it is it it it is a difficult task, right? So but why?
speaker-0: Think first of all it's building exterior is not a factory floor. Right? It's not like you know a car manufacturing site or other stuff like that. The fact that every building is different, every roof, every facade, the materials, the access condition are different, the wind, the light, the reflection, there is changes that is constant around the building and around the day. Right? You have a scenario that It's not just that every building is different. Every side of the building acts differently. Right? And we're looking at a system that has to operate hundreds of feet around the ground, carrying tools, carrying camera, carrying power, carrying communication, have a lot of safety logics, right? So solving this requires a lot of mechanics and controls and sensors and deployment infrastructure and software and field operation to work together. Right? And Right. Even though we've cleaned and inspect more than five million square feet of facade in the last in the last twelve months, I always say there is a reason why ninety nine point nine percent of the be of the buildings around the world are still being cleaned with people. Right? And the buildings are getting more and more complicated with the design.
speaker-1: Okay. Right. so and where is the system genuinely autonomous and where does the human still make the decision, especially in the end?
speaker-0: Yeah so I'm a big believer in saying that in in robotics there will always be the last mile and the last mine will be operate will will be part of the loop, right? In our case the robot is autonomous inside a defined and qualified operating area, right? Meaning that we have the operator that he placed the robot on the facade, the robot knows what he needs to do, how he needs to clean, what he needs to inspect, the navigation it on it's autonomous. But eventually we have a person in the loop that is there to deploy the robot and to interfere when there is any issues or just to make sure everything is working flawlessly.
speaker-1: So it's a little bit like like with the Waymo, for example, where they're announced as full self driving cars, but they're still like per person supervising if everything aligns with it, right?
speaker-0: Okay, okay. So humans still qualify the building and he plan the deployment and install the equipment and he put the define the the the boundaries of the project and supervise it, right? And I think that it is a very complex environment, right? You have pedestrians, you have the weather, you're walking on height. It will always like even like a decade from today you will still need to require a person that is there keeping an eye on the platform.
speaker-1: Quick s
speaker-2: So for example, you control the you control the robots, you have the systems, and you control and train all of the people, and that will remain the same throughout the world, or is it going to be a question of the contract relationship with the ten the the owner of the building?
speaker-0: that's a that's a good question. eventually for us as a company, we are focusing on metropolis with high density of building. Right? However, at the same time every geography acts differently, right? Tokyo has the highest density of building around the world. For the robotics to establish and to operate as the robotics directly, it will be very challenging. Right. So in Japan, for example, we'll work with partners that they will deploy and they will do the service. In contrast to Israel, we have our own service company, so we control the supply chain. In the US, every city and every state acts differently, right? So we need to adapt according to the different area that we are working.
speaker-1: Okay. So I mean AI is like in everyone's mouth, right? So AI is like one of the big terms. So let's be a little bit more precise here. What does DAI actually do inside the product or in your product especially today? Perception, localization, real planning coverage, defect detection, or something completely else?
speaker-0: Yeah, no, I think that kind of all of the above, including the safety factors as well. again for us safety is probably the biggest and most important component for the robotics, right? so we're invest a lot of resource on building the logic around it. But in going one step back, we look at it is in the processing of close range, georeference images, identifying, classify potential visible conditions. multiple perspective and identifying the technical review, right? And as you mentioned, the localization is is a huge factor, right? And we have the routing of the robot, right? It's think it's kind of connecting to the continue of the discussion. But some of the challenges that we have it's around the vision, right? Because imagine that you have a building made out of glass. Right. Right. And you have reflection and glare. And shadowed. And it could be that the robot will be at the same location and what he sees is changes constantly because of the weather and the sun and clouds and reflection. Right? So we've built a lot of mechanism in order to help us identify because the faster the robot will be, the better the unit economics will be for anyone, right? And it's all depends on the autonomy of the system.
speaker-1: You know, it it also reminds me on full self driving cars again. It it sounds so simple in when you when you think about it, right? What could go wrong? A car driving down a road, you have like the middle lane, how could this be like difficult to to figure out for an AI, right? And obviously there are so many factors that need to be counted in co calculation. And you brought it up. For example, when there is dirt, the the model needs to understand if is it dirt, is it a shadow, is it maybe something completely different? So I mean you you already mentioned s many Different things and maybe you could highlight a little bit more. So so the a DAI, so to say, is is helping detect like what it really is facing, so to say, and what is actually the most difficulty you're facing? Is it like like the wind or is it really the dirt or or the reflection? What is causing you the main problem?
speaker-0: Yeah. So earlier you mentioned Waymo, which I always try to use Waymo as an example if I'm not mistaken. And the thing for them is that they had to build everything from scratch. Right? They spent the years on building the navigation and the visual navigation and understanding the the the environment. And this is what we did also on the first four years of the of the company life. Right? it's not I mean other than the actuators that we have, nothing is off the shelf. Right. So imagine that we need to build the entire spec of the technology, of the hardware, of the software, of the navigation, of the lamo conversion, and another big factor is the operation, right? We have to build the operational playbook because no one ever deployed robots on the roof. So we spend probably we've been in more than one hundred roofs of building in order to build everything. And gathering all of this data help us And the experience of course helped us really to build the entire infrastructure of the robotics.
speaker-1: Hmm, okay. And how do you prove the robot covered the full surface rather than just producing plausible roads and a set of images? I mean you you make sure you have to make sure that it like did its job pretty well, right?
speaker-0: Yeah. So I'll I'll I'll give an answer that, you know, we have a few months ago we did a pilot with one of the biggest cleaning companies in the US and we've told them that we've built a sensor that is helping us to close the loop on the cleaning. And then I asked them how do you validate today if the building owner was satisfied and the job was good? Yeah. And they said if we got paid, we know we did a good job. Right? However in our case we are always being measured. more right and I think this is probably every every every entrepreneur that builds robotics or even like innovation in software, you will always be measured when you try to replace a person. In our case we create an auditable coverage record. So it's not just route visualization. So when we are onboarding the building before the project we divide the building based on elevation If the building owner have the cards of the building, we are kind of putting it with our system, we're building kind of a digital twin of the facade. And based on that the route optimization has and if not, we're doing it manually. And when the map collect the data back against the plan, meaning that when the robot traverses the facade, with the movement of the actuators, we know exactly the location that he was there and he has the mission that he needs to do. So the entire process is it's it's quite smooth.
speaker-1: interesting. Just just a question that's popped up in my head. Yeah how how how did you train that model to so so was there like some kind of reinforcement learning, using verifiers to to ensure that the model is like yeah, get getting getting like like yeah those verifiers for finishing the task? how how did you do it? I I'm just curious.
speaker-0: blood, sweat and tears mostly. But we've try to spend as much time we can on different buildings in order to build the visual dictionary that we build. Right? So as I mentioned we've spent probably more time on more than one hundred buildings, also on the roof but also on the facade in order to build our internal l libraries. Right. And this is also we didn't have any place that we can say, you know, whether if it's like a drones that you can have so many open sources that you can train it or autonomous car or humanoids or whatever it is, right? We didn't have like a simulation environment. We have to build everything internally.
speaker-1: Okay. Interesting. And I mean you already described the collected data as the backbone like of the digital twin, right? I think you mentioned it once. So what does the twin contain beyond images and and a location map? So a little bit more technical question.
speaker-0: Yeah. So the robot traversed the building. today we are using the same navigation cameras to collect the data. Right? So we're kind of right building a map of the facade, but it's very different than the digital t I don't like to use the the term digital twin so much for the robotic case. No, but it it happens, right? it it happens because this is what we've been used to, right? Yeah. it's not like that you are getting That you're getting like the full scaling images that you see with the drones that are doing inspection. in our case, what you are getting is kind of a visual map that is like a puzzle. Right? So we have four four cameras on the robot. And imagine that every meter that the robot traverses the facade, we're taking images and we're building kind of a huge puzzle of all of the images. And for the building owners, for building operators. for facade companies, for engineering companies, insurance, yeah, municipalities, today it was kind of a black box that no one had idea, right? What's going on. Right. I I like to call it Pandora box, right? Because we are covering what no one wants to see in a way. And so I think this
speaker-1: I like the analogy.
speaker-0: yeah. And it's not just about, you know, every conversation that we have with customers and investors, wow you can use it for defense as well to take this data and you can use it for weather. And you can use it for city metrics or whatever it is, right? We're agnostic and we've designed a platform that's we have the robot, I'm sure that you're going to show the video, but we have the robot that traverses the building and he can connect any payload that we want up to twelve pounds, right? So the first payload is the cleaning and the inspection and we have some really cool things down the sleeve that we are planning to launch next summer.
speaker-1: Super excited, super excited. Peter, why don't you take it over the business side? I'm excited and curious about this as well.
speaker-2: Yeah, you know, it's funny. I'll I'll just tell a quick story, keto. The last video that Kim sent me of his son was him playing in a like a toy car. And I was thinking about him last night because a firetruck, a big red firetruck, went by in New York and little kids love mechanical things. So you guys are in a part of the business or the field of AI that would be very exciting to kids, like things that are physical and real versus stuff that's behind a screen. So I guess The question is, are you guys actually a cleaning company or are you a data company?
speaker-1: Hmm.
speaker-0: So if you would ask Ido in twenty twenty six I will say both. If you would ask me five years ago, I would still say both. But if an investor would have asked me this five years ago I would say we're only a product company. You know, the world is shifting. Everything happening with AI in robotics, even today like to say that you build a robotic company, it it makes sense. But when we started five years ago saying that we are going to build a robotic platform, everyone looked at us like we're crazy. Right? But the world is shifting. And for us, from day one, in in my in my well in my heart, I knew that we're going to do both. Because I think that a company like the robotics that is the really building a new category within a business, if they want to win, they need to own the entire ecosystem and the supply chain. Right? We need to control our own destiny. And by the way, you see it even with other companies today. Right? Soft bank robotics, this is a public information. They bought a cleaning company in the US in order to implement the robots for cleaning. They did the same in Australia and the same in Japan. Right? And I think that in our world is going to places where soft services is is dead unfortunately because someone wants to do this job. You need to be able to promote the technology and to take ownership on the entire supply chain. So on the robotics case we do both. Cleaning is the mo is the is the wedge, is the Troyan horse in a way. And I think that the inspection it's what will build a very unique economic standpoint for the entire ecosystem of real estate.
speaker-2: Mm. Well now that we've talked about that I bet Kim wants to talk about the reality.
speaker-1: Yeah, yeah, absolutely. So your team told us Varebotics completed a project with Nvidia. What exactly was the project? What did each site contribute and what actually shipped?
speaker-0: Yeah. so in Israel we are working very closely with C BRE, which is one of the biggest facility management company around the world. And we're able to build a relationship with them when we had a clunky MVP like years back. And once we started to work and to deliver, we've started to get more and more projects and we're mainly focusing kind of a class A campus of companies. Nvidia is being managed by CBRE. we have a contract with them through cleaning And to do inspections once a year. And it was a real deployment in NVIDIA campus. We have a three year contact with them. And with the robotics we provide the cleaning services and the inspection services. The cleaning is twice a year and the inspection we do once a year or once two years, it really depends.
speaker-1: Hm, cool. Peter.
speaker-2: You know, in your stuff, in in the materials that we've read about your company, it says that a robot can clean a building f about four times faster than a human crew. and that one person can run several units for your company. So I I'm curious, like how did you guys figure out that four X number? What was the set of conditions? Like were you just taking data from stuff that you deployed or were there was it a a test that you ran under certain control conditions?
speaker-0: So I I think that it was part of the first question that you on my background. I think that coming into this venture, I knew all the bits and bytes in window cleaning and inspection, right? So I knew the cost of human labor, whether if it's the salaries, the equipment, the insurance, and I knew all of the things that associate to that. And we've designed the platform, we knew what is our north star on the performance, right? Because this robot can be the coolest thing in the world, the nicest thing in the world. But if the unit economics doesn't make sense for someone to pay for that, nothing makes sense, right? so the way that we structure the entire ecosystem of robotics is to be able that today we're not there, but the vision is that every robot will reach ten thousand square feet per day and we're deploying multiple robots. Right. So in a way, if you have three people today that are doing window cleaning with the BMU, they can reach around eight thousand square feet, right? So we want with less amount of less less than three people with four robots you can do significantly more than that because it's all around unit economics, right? people want people need to justify the fact that they're paying for something and if you're not giving them the same value, why would they pay?
speaker-2: It it's interesting, you know, I have friends who've worked in fields where they do physical deployment of technologies, whether it's medical devices or robotics. And a lot of companies will get to the point where they've done a bunch of testing but they never actually get their product into into the field. So like when you guys are doing a pilot, what's the rate of success from pilot to implementation and then you've got a new customer?
speaker-0: So I would say that in the last twelve months we are mainly focusing on Israel, because we are here and the R and D is here, even though we're an American company. here it's one hundred percent rate. All the project that we're doing today is deployable and it's commercial project. same as the project that we did in the US. I would say that the place where we saw it's not objection, it's just a process, is that we had a full full paid pilot in Tokyo at the end of last year that were successful. However, to take a solution to Japan and to deploy it, it's a very long journey that requires a lot a lot of efforts. But today for the robotics, at least in Israel, which is a market that we're very active, we don't do pilots. All the project are paid project with a two to three years contract.
speaker-2: So is just one quick question. in i in Japan, is the different implementation a bureaucracy question that's or is it a question about corporate structure?
speaker-0: Both and you can also add culture. Right? So for a company, for a for a technology company like the robotics that is redefining an existing industry, it's a it's a process because you need to completely localize the company, right? It's not just about changing the language of the dashboard to Japanese, it's transf it's moving the entire cloud to Japan, it's building a Japanese company. It's it's a very complicated process. In contrast to the US, you know, in the US we've met C level a few weeks ago of one of the biggest real estate companies in New York. And within one, two meetings you're closing like the first project, right? And in Japan it's just a big process. However, for me, I think that Japan is going to be one of our most significant markets. Not sure if you spent some time in Tokyo. Tokyo has the highest density of the world. They have a huge problem with labour and if in the US the class A billing owners clean their billing three times a year, in Japan the norm is four times a year. Right? So it's a very promising market for us, but it's something that we are prioritizing probably for Q four next year.
speaker-2: So you guys have built something that's really difficult. I think a lot of people, probably most of the people who hear our podcast or read Kim's newsletter know that, understand AI, but many of them may not understand how hard robotics really is. You guys have built a very complicated a a very good solution for a complicated problem. And you have, you know, worked with NVIDIA, who's like the eighth pounder pound gorilla in ships in this industry. But like what would you say is the one thing that
speaker-0: I would say that the harder things the hardest thing to copy is not one component. It is the entire field and the operation system. A competitor can build a camera, he can build an actuators, it can build a cleaning or you can copy any individual feature. But I think it is much harder to reproduce the knowledge of how to qualify buildings, how to deploy from different buildings, how to manage wind. and obstacles, how to supervise the multiple robots and how to connect all of the data that we've built. Right? I think that one of the unique things that we've had is the dataset of the images that we have. That the use of the images is also for the movement of the robot, but it's also on the inspection mode. And you know it's such a complex environment and and and we still see since we've started Like five years ago there was a lot of companies that unfortunately closed. And we see some companies that are coming and building the solutions today. however they are still doing things that are very similar, right? And I think that the combination of the experience that we have on the last five years, this is a huge defensibility for the robotics.
speaker-2: So you just said five years ago we were here. Five years from now, where are you gonna be? Where's the value in your business?
speaker-0: So five years from today, as I mentioned earlier, it's going to be a hybrid company on the go to market. In some cases we're going to own the entire chain, in some cases we're going to lease it and to have different models. But in five years the vision is to really be leading the category of exterior services. And I think that Looking on the data aspects, it's probably one of the strongest thing that we are going to have. Right? It's the robotics going to work with architects and general constructors and municipalities and insurance companies and really to owning this category. Okay.
speaker-1: Okay. so specialized robots get far less attention than humanoids, obviously, right? So everyone is talking about humanoids routing the laundry, yeah, unfortunately, right? So the question is, are we telling the wrong story about what commercially useful physical AI will actually look like? And what's the one thing you wish more people understood about robotics in the built environment?
speaker-0: Yeah, I am totally agree with what you said. I think that there is a huge huge buzz around humanoids and I think that probably as the the VC ecosystem will change every every year there is a new trend, right? And we often being asked why won't the humanoid do it? Let's first have a humanoid doing things that are more basic. I think that for robot working on a building exterior, it doesn't need to look like a person. It needs to perform a specific, repetitive, dangerous and expensive job. It could have two arms, three arms, five arms, six arms. Yeah. No one cares. Right? We've designed the platform the way it is today to solve an operational problem. And we've believed that the architecture that we have today is the best architecture in the market because it's allowing us to be To have zero dependency on the environment.
speaker-1: It it just crossed my mind. But I mean several companies are building humanoid robots because the world around us is built for us humans, right? So this is why they try to mimic us and our in our physical behavior and whatnot. But looking at at at for example, the the task you're trying to solve, like cleaning difficult areas, I mean it's in in in the first place A place where it's not practical to be human at at all in the first place, right? So maybe you have more room to develop above like humanoid physical stances, so to say, right? So is it like maybe even an opportunity to go beyond having humanoid robots, developing it further? Because you you just mentioned like having more arms or legs or s or something like this, right? It so that the the the difficulty could be an opportunity, right?
speaker-0: Yeah. It's it's also opportunity, but it's also to think in a way to think out of the box. Right. Yeah. you know, if you're looking like in all of this robotics for surgery for medicine. Yeah. So you know the there were there is a companies that started a few years ago doing lycroscopy and other surgeries, which is just kind of a few devices and today you've connect them to a humanoid. I don't know why, maybe something psychology a bit more on the human mind that make us feel like, okay, this is nice. But I don't know, do I want a robot in my home folding my Landerid? Why doesn't the machine do it itself, right? true. So there you you know there is a lot of also I think that the there is also this robots now doing Mills, the guy that sold his company to Amazon or it's Mike Mark Aaron if I'm not mistaken. You know, you build a machine that can produce thousands of dish per day. It doesn't look like a humanoid. Right? I mean o honest honestly, who cares how it look? As long as it's delivering and it justifies the unit economics and it justifies the demand. And it's not let's you know, I'm I was in the C and the NVIDIA event earlier this year. And you see it all over the time. You see these videos of robots that are doing boxing match.
speaker-1: Yeah, yeah. I mean
speaker-0: Okay, I'm sorry to say that, but this is why you invested so much money and so much resource to build robots that will compete on each other, who is beating each other, and this is how you choose to promote it? Solve the problem.
speaker-1: Yeah. Yeah, yeah. I I fully agree with you. I fully agree with with you. And to be to be frank, I'm a little bit I don't want to say annoyed, but probably bored by seeing another robot doing some MMA martial arts stuff or or s or another robot doing some dancing. Whereas as as you rightfully said, so why don't we have like good videos of robots doing Yeah, useful stuff that we as as human need. And I really like what you were just saying. So in a way it forces you to think out out of the box. And this makes totally sense to me. And actually I think this is where we should should be heading and think more in in a way, right? Making use of such technologies and think out of the box.
speaker-0: I think that Figure did something like a few months ago with organizing the mail, if you remember and this was cool. This is what people want to see. I want to see what the robot is actually delivering. Yeah. Right? Yeah. And and I think that most people, you know, always say that it's very similar to to autonomous cars, right? Like in two thousand seven in two thousand ten everyone said by the end of this decade all the cars will be automated. You won't have you won't own a car. Right, but we're almost in twenty thirty. Right? And only now Vegas approved to deploy the couple hundreds of Teslas if I'm not mistaken. So that's true. It's taking a time same as for for the humanoids, which is it's it's a great, but they just still need to prove themselves before it will be a commodity.
speaker-1: Yeah, f first slowly and then all at once. It it it it needs its time to adjust, to develop, but when you when you reach this certain threshold it's like it's going to explode. And hopefully in the field of robotics we're slowly approaching this third threshold where it now really takes off, right? Yeah.
speaker-0: Yeah, definitely. I I totally agree on that.
speaker-1: Peter, you you had a question?
speaker-2: I had one quick question, Ido. It it seems like the most the closest analogy to what you guys are doing is actually the robot that the most people have seen before, which is one cleaning the rug in your house. Yeah or sh mowing your lawn in your backyard. So I it's odd that people are obsessed with humanoid robots to do a task that they're already accustomed to seeing done by something that doesn't look at all like a person.
speaker-0: yeah. I I do agree although that you already see videos today of the humanoids taking the law doing them on, right? I I don't really understand the logic. But yeah, I mean you know iRobot was the pioneers of that. And they did a really huge change in the industry. And you have Dyson, which is probably mm the biggest innovator in the world. That if he's hearing this, I want to meet him because I think that he made the huge impact. And it's really you know, when we started the robotics like a few years ago building the pitch, I always say that yeah, it's like the first vacuum of Dyson. You know, Dyson took out the bag, they you had this ugly plastic bag in the vacuum back there and people said wha no, it's impossible, right? And so there's a lot of shifting also in kind of in in in people understanding, but it was always like this technology. just one less story about innovation, that a few years ago I worked in a company named Bring. And they were kind of the pioneering in less mile technology for SAS companies. So it was really around transforming logistics for company. You know, imagine companies used to work with pen and papers and Google Maps and like zero proof of delivery, zero innovation, I think. Yeah. And we had a big client in Australia and I I was there, I did the the deal there and I also went there for like for onboarding. And you know, I'm explaining about the driver app and route optimization that we save so much time and improving efficiency and Everything, right? Completely transforming the business. And there was a guy in the audience that's like, Hey do you also do 3D scanning for optimizing how I load the boxes into my car? And I'm like, Man, like yesterday you used to go to the printer to print it, to go to Google Maps, to check the address. You have everything automated today, right? And it's always part of people expectation. This is why I said it For us when we do the cleaning people will always try to look more for like a proof, right? If today on the traditional window cleaning most people will say nothing. For us because they know it's a robot, they will even want to do like a bigger investigation.
speaker-1: Mm.
speaker-2: Ido, can I ask I have one qu personal question based on something you just said. Yeah. So you were off working on it for a different company doing a completely different thing, and you're implementing stuff that you learned in your own family business. Yeah. Like you're bringing your experience back. What does it feel like to have this legacy family business that you decided to walk away from and now you've come back with your experience sort of like the prodigal son and you're changing the way it works?
speaker-0: We we need to open another podcast for that to tell them Pastor.
speaker-1: Yeah.
speaker-0: That's that's that's that could be a that could be a Netflix show if you want.
speaker-1: I'd love to do a show about it.
speaker-2: I I asked because one of my f closest friends d took over his family business and grew it by a hundred and fifty times and just using totally different techniques. So I
speaker-0: it's a huge company in the Israeli landscape, it's the biggest. And I always say that leaving the company was the most difficult thing that I ever did in my life. like one hundred percent. And today everything that we've able to accomplish it's by ourselves. And I think it's it's interesting because all of the project, the innovation, the methodology, the technology, the know everything is I mean it could be nicer if we had the support, right? Things would probably happen faster. But it is what it is.
speaker-1: This is a perfect ending. I know this was excellent. Thank you so much. Where can people find their robotics and follow what you are building?
speaker-0: so obviously the robotics.com, LinkedIn, Instagram, Twitter and soon Q4 we are going to have a few projects in the US and when you're looking at the building, it's not a bird and it's not Superman, it's probably our robots over there.
speaker-1: Thank you so much for joining us and thanks to everyone listening. You can find us at get superintel.com. I don't it was such a pleasure. Thank you so much.
speaker-0: I appreciate it guys.
speaker-2: Thank you. Thanks everybody. Thanks for coming and we'll see you next time.
speaker-1: Hi, I'm Kim Ismerg, Super Intelligence Editor in Chief, and today we are joined by Ido Genosar. I hope I pronounced your name correctly, co-founder and CEO of Verobotics. Welcome.
speaker-0: Welcome. Thank you very much for having me today.
speaker-1: Such a pleasure having you. So let's jump right into it and right into the question. For someone seeing Verobotics for the first time, what problem are you solving in one sentence?
speaker-0: Okay. the robotics replace dangerous and inefficient manual work on building exterior, working on height with a lightweight robotic platform that clean, map, inspect and documents the facade.
speaker-1: That is very concise, very precise, Peter.
speaker-2: So Ito, you you came to this this business originally through construction, facades, building management, not from a robotics lab. So what did you see in that world from which you came that robotics first-time founder would have missed?
speaker-0: So my family has been in the facade business for the last forty years and I had an active role in the company on the innovation side. And I think that probably the biggest understanding that I had was that the problem wasn't the person that is sitting inside the basket of the BMU, which is the billed management unit. It's the huge crane that you see going outside of the building. The problem was the existing infrastructure and the access itself that was broken.
speaker-2: So can you can you clarify that a little bit when you say the the existing infrastructure and access, what does that mean?
speaker-0: Yeah, so today when you are doing exterior job for building, you either do it with a crane, which is basically the building management unit. You live in New York, Peter, so you see these cranes everywhere. And the other way to do
speaker-2: Building I can see the the I can see the Chrysler building right now.
speaker-0: Yeah, so you either see the cranes or you see the scaffolding everywhere. And the other method is with trope access. It's people that are literally rapping down the building. And I think that my perspective from the business was the understanding that every building is different, every roof is different, every facade is different, every access is different. And you can solve the problem using the existing inca infrastructure the building has. And for me this is something that I figured out quite in the beginning and that we have to change the access to the building. While in contrast, if you look at other solutions today in the market and other that operate before, they said, Okay, we have a platform, we have the crane, we wanna improve the safety and the efficiency, so instead of putting a person inside the basket of the crane, let's put a robot. But the assumption, to my opinion wasn't the right assumption because the problem wasn't the person in the basket. The problem was the basket.
speaker-2: Right. I I'm I'm curious though, if if that's the case and you're saying the problem was the basket, like I personally I'm not afraid of very many things, but I don't like heights. So when I think about like being one of the guys on on one of those rigs going up and down the side of a building hundreds or thousands of feet above the ground, that would not be something I would pick. It's very dangerous, e even though those guys are all s connected by climbing ropes. So how much of what you're doing, what you're solving is about keeping people away from harm versus improving the speed and the cost. And then maybe another aspect is like how does it affect the tenants of the building or are they aware of this at all?
speaker-0: Yeah. So safety is the starting point, right? But eventually the economics it's what really make the adoption and the scalability. we are moving people away from spending hours exposed to the side of the building, exposed to wind, other elements and of course the risk of walking on heights, right? I would also say that it's not that there is a lot of casualties every year. There is a lot of accident. But the ev effect of the accident is big, right? Because you involve the fire department and the police and the insurance and the liability and the tenants. So it's a big headache, right? And our goal is really to avoid as much as possible of having people dependent. And at the same time it's to reduce the fraction with the building, with what we're doing. on the tenant perspective, for them Again, I can also say what on my my perspective that I'm traveling a lot and you know every time that you go to a hotel in cities like Tokyo or Hong Kong or New York or San Francisco, most of the time you'll have some small note on the table, be aware that they're cleaning the window, for your privacy, close the curtain. Right? So I think that for us the fact that no one is over there and suddenly arrive from somewhere is something that it's nicer. It's even cooler in a way on the on the perspective of the customer and Going back it's really around the safety, the efficiency and the visibility for the biddingers.
speaker-2: Got it.
speaker-1: Okay. so my wife and I we just bought a house and obviously our winners are Dirty all the time, we have like a twelve month old baby, a son, so he's dodging all like all the windows all the time. And as far as I know, there isn't really much of like autonomous cleaning robot. So why has and especially like high rise facade cleaning and inspection been so difficult to automate? So I just brought up this personal like information because I know that it is it it it is a difficult task, right? So but why?
speaker-0: Think first of all it's building exterior is not a factory floor. Right? It's not like you know a car manufacturing site or other stuff like that. The fact that every building is different, every roof, every facade, the materials, the access condition are different, the wind, the light, the reflection, there is changes that is constant around the building and around the day. Right? You have a scenario that It's not just that every building is different. Every side of the building acts differently. Right? And we're looking at a system that has to operate hundreds of feet around the ground, carrying tools, carrying camera, carrying power, carrying communication, have a lot of safety logics, right? So solving this requires a lot of mechanics and controls and sensors and deployment infrastructure and software and field operation to work together. Right? And Right. Even though we've cleaned and inspect more than five million square feet of facade in the last in the last twelve months, I always say there is a reason why ninety nine point nine percent of the be of the buildings around the world are still being cleaned with people. Right? And the buildings are getting more and more complicated with the design.
speaker-1: Okay. Right. so and where is the system genuinely autonomous and where does the human still make the decision, especially in the end?
speaker-0: Yeah so I'm a big believer in saying that in in robotics there will always be the last mile and the last mine will be operate will will be part of the loop, right? In our case the robot is autonomous inside a defined and qualified operating area, right? Meaning that we have the operator that he placed the robot on the facade, the robot knows what he needs to do, how he needs to clean, what he needs to inspect, the navigation it on it's autonomous. But eventually we have a person in the loop that is there to deploy the robot and to interfere when there is any issues or just to make sure everything is working flawlessly.
speaker-1: So it's a little bit like like with the Waymo, for example, where they're announced as full self driving cars, but they're still like per person supervising if everything aligns with it, right?
speaker-0: Okay, okay. So humans still qualify the building and he plan the deployment and install the equipment and he put the define the the the boundaries of the project and supervise it, right? And I think that it is a very complex environment, right? You have pedestrians, you have the weather, you're walking on height. It will always like even like a decade from today you will still need to require a person that is there keeping an eye on the platform.
speaker-1: Quick s
speaker-2: So for example, you control the you control the robots, you have the systems, and you control and train all of the people, and that will remain the same throughout the world, or is it going to be a question of the contract relationship with the ten the the owner of the building?
speaker-0: that's a that's a good question. eventually for us as a company, we are focusing on metropolis with high density of building. Right? However, at the same time every geography acts differently, right? Tokyo has the highest density of building around the world. For the robotics to establish and to operate as the robotics directly, it will be very challenging. Right. So in Japan, for example, we'll work with partners that they will deploy and they will do the service. In contrast to Israel, we have our own service company, so we control the supply chain. In the US, every city and every state acts differently, right? So we need to adapt according to the different area that we are working.
speaker-1: Okay. So I mean AI is like in everyone's mouth, right? So AI is like one of the big terms. So let's be a little bit more precise here. What does DAI actually do inside the product or in your product especially today? Perception, localization, real planning coverage, defect detection, or something completely else?
speaker-0: Yeah, no, I think that kind of all of the above, including the safety factors as well. again for us safety is probably the biggest and most important component for the robotics, right? so we're invest a lot of resource on building the logic around it. But in going one step back, we look at it is in the processing of close range, georeference images, identifying, classify potential visible conditions. multiple perspective and identifying the technical review, right? And as you mentioned, the localization is is a huge factor, right? And we have the routing of the robot, right? It's think it's kind of connecting to the continue of the discussion. But some of the challenges that we have it's around the vision, right? Because imagine that you have a building made out of glass. Right. Right. And you have reflection and glare. And shadowed. And it could be that the robot will be at the same location and what he sees is changes constantly because of the weather and the sun and clouds and reflection. Right? So we've built a lot of mechanism in order to help us identify because the faster the robot will be, the better the unit economics will be for anyone, right? And it's all depends on the autonomy of the system.
speaker-1: You know, it it also reminds me on full self driving cars again. It it sounds so simple in when you when you think about it, right? What could go wrong? A car driving down a road, you have like the middle lane, how could this be like difficult to to figure out for an AI, right? And obviously there are so many factors that need to be counted in co calculation. And you brought it up. For example, when there is dirt, the the model needs to understand if is it dirt, is it a shadow, is it maybe something completely different? So I mean you you already mentioned s many Different things and maybe you could highlight a little bit more. So so the a DAI, so to say, is is helping detect like what it really is facing, so to say, and what is actually the most difficulty you're facing? Is it like like the wind or is it really the dirt or or the reflection? What is causing you the main problem?
speaker-0: Yeah. So earlier you mentioned Waymo, which I always try to use Waymo as an example if I'm not mistaken. And the thing for them is that they had to build everything from scratch. Right? They spent the years on building the navigation and the visual navigation and understanding the the the environment. And this is what we did also on the first four years of the of the company life. Right? it's not I mean other than the actuators that we have, nothing is off the shelf. Right. So imagine that we need to build the entire spec of the technology, of the hardware, of the software, of the navigation, of the lamo conversion, and another big factor is the operation, right? We have to build the operational playbook because no one ever deployed robots on the roof. So we spend probably we've been in more than one hundred roofs of building in order to build everything. And gathering all of this data help us And the experience of course helped us really to build the entire infrastructure of the robotics.
speaker-1: Hmm, okay. And how do you prove the robot covered the full surface rather than just producing plausible roads and a set of images? I mean you you make sure you have to make sure that it like did its job pretty well, right?
speaker-0: Yeah. So I'll I'll I'll give an answer that, you know, we have a few months ago we did a pilot with one of the biggest cleaning companies in the US and we've told them that we've built a sensor that is helping us to close the loop on the cleaning. And then I asked them how do you validate today if the building owner was satisfied and the job was good? Yeah. And they said if we got paid, we know we did a good job. Right? However in our case we are always being measured. more right and I think this is probably every every every entrepreneur that builds robotics or even like innovation in software, you will always be measured when you try to replace a person. In our case we create an auditable coverage record. So it's not just route visualization. So when we are onboarding the building before the project we divide the building based on elevation If the building owner have the cards of the building, we are kind of putting it with our system, we're building kind of a digital twin of the facade. And based on that the route optimization has and if not, we're doing it manually. And when the map collect the data back against the plan, meaning that when the robot traverses the facade, with the movement of the actuators, we know exactly the location that he was there and he has the mission that he needs to do. So the entire process is it's it's quite smooth.
speaker-1: interesting. Just just a question that's popped up in my head. Yeah how how how did you train that model to so so was there like some kind of reinforcement learning, using verifiers to to ensure that the model is like yeah, get getting getting like like yeah those verifiers for finishing the task? how how did you do it? I I'm just curious.
speaker-0: blood, sweat and tears mostly. But we've try to spend as much time we can on different buildings in order to build the visual dictionary that we build. Right? So as I mentioned we've spent probably more time on more than one hundred buildings, also on the roof but also on the facade in order to build our internal l libraries. Right. And this is also we didn't have any place that we can say, you know, whether if it's like a drones that you can have so many open sources that you can train it or autonomous car or humanoids or whatever it is, right? We didn't have like a simulation environment. We have to build everything internally.
speaker-1: Okay. Interesting. And I mean you already described the collected data as the backbone like of the digital twin, right? I think you mentioned it once. So what does the twin contain beyond images and and a location map? So a little bit more technical question.
speaker-0: Yeah. So the robot traversed the building. today we are using the same navigation cameras to collect the data. Right? So we're kind of right building a map of the facade, but it's very different than the digital t I don't like to use the the term digital twin so much for the robotic case. No, but it it happens, right? it it happens because this is what we've been used to, right? Yeah. it's not like that you are getting That you're getting like the full scaling images that you see with the drones that are doing inspection. in our case, what you are getting is kind of a visual map that is like a puzzle. Right? So we have four four cameras on the robot. And imagine that every meter that the robot traverses the facade, we're taking images and we're building kind of a huge puzzle of all of the images. And for the building owners, for building operators. for facade companies, for engineering companies, insurance, yeah, municipalities, today it was kind of a black box that no one had idea, right? What's going on. Right. I I like to call it Pandora box, right? Because we are covering what no one wants to see in a way. And so I think this
speaker-1: I like the analogy.
speaker-0: yeah. And it's not just about, you know, every conversation that we have with customers and investors, wow you can use it for defense as well to take this data and you can use it for weather. And you can use it for city metrics or whatever it is, right? We're agnostic and we've designed a platform that's we have the robot, I'm sure that you're going to show the video, but we have the robot that traverses the building and he can connect any payload that we want up to twelve pounds, right? So the first payload is the cleaning and the inspection and we have some really cool things down the sleeve that we are planning to launch next summer.
speaker-1: Super excited, super excited. Peter, why don't you take it over the business side? I'm excited and curious about this as well.
speaker-2: Yeah, you know, it's funny. I'll I'll just tell a quick story, keto. The last video that Kim sent me of his son was him playing in a like a toy car. And I was thinking about him last night because a firetruck, a big red firetruck, went by in New York and little kids love mechanical things. So you guys are in a part of the business or the field of AI that would be very exciting to kids, like things that are physical and real versus stuff that's behind a screen. So I guess The question is, are you guys actually a cleaning company or are you a data company?
speaker-1: Hmm.
speaker-0: So if you would ask Ido in twenty twenty six I will say both. If you would ask me five years ago, I would still say both. But if an investor would have asked me this five years ago I would say we're only a product company. You know, the world is shifting. Everything happening with AI in robotics, even today like to say that you build a robotic company, it it makes sense. But when we started five years ago saying that we are going to build a robotic platform, everyone looked at us like we're crazy. Right? But the world is shifting. And for us, from day one, in in my in my well in my heart, I knew that we're going to do both. Because I think that a company like the robotics that is the really building a new category within a business, if they want to win, they need to own the entire ecosystem and the supply chain. Right? We need to control our own destiny. And by the way, you see it even with other companies today. Right? Soft bank robotics, this is a public information. They bought a cleaning company in the US in order to implement the robots for cleaning. They did the same in Australia and the same in Japan. Right? And I think that in our world is going to places where soft services is is dead unfortunately because someone wants to do this job. You need to be able to promote the technology and to take ownership on the entire supply chain. So on the robotics case we do both. Cleaning is the mo is the is the wedge, is the Troyan horse in a way. And I think that the inspection it's what will build a very unique economic standpoint for the entire ecosystem of real estate.
speaker-2: Mm. Well now that we've talked about that I bet Kim wants to talk about the reality.
speaker-1: Yeah, yeah, absolutely. So your team told us Varebotics completed a project with Nvidia. What exactly was the project? What did each site contribute and what actually shipped?
speaker-0: Yeah. so in Israel we are working very closely with C BRE, which is one of the biggest facility management company around the world. And we're able to build a relationship with them when we had a clunky MVP like years back. And once we started to work and to deliver, we've started to get more and more projects and we're mainly focusing kind of a class A campus of companies. Nvidia is being managed by CBRE. we have a contract with them through cleaning And to do inspections once a year. And it was a real deployment in NVIDIA campus. We have a three year contact with them. And with the robotics we provide the cleaning services and the inspection services. The cleaning is twice a year and the inspection we do once a year or once two years, it really depends.
speaker-1: Hm, cool. Peter.
speaker-2: You know, in your stuff, in in the materials that we've read about your company, it says that a robot can clean a building f about four times faster than a human crew. and that one person can run several units for your company. So I I'm curious, like how did you guys figure out that four X number? What was the set of conditions? Like were you just taking data from stuff that you deployed or were there was it a a test that you ran under certain control conditions?
speaker-0: So I I think that it was part of the first question that you on my background. I think that coming into this venture, I knew all the bits and bytes in window cleaning and inspection, right? So I knew the cost of human labor, whether if it's the salaries, the equipment, the insurance, and I knew all of the things that associate to that. And we've designed the platform, we knew what is our north star on the performance, right? Because this robot can be the coolest thing in the world, the nicest thing in the world. But if the unit economics doesn't make sense for someone to pay for that, nothing makes sense, right? so the way that we structure the entire ecosystem of robotics is to be able that today we're not there, but the vision is that every robot will reach ten thousand square feet per day and we're deploying multiple robots. Right. So in a way, if you have three people today that are doing window cleaning with the BMU, they can reach around eight thousand square feet, right? So we want with less amount of less less than three people with four robots you can do significantly more than that because it's all around unit economics, right? people want people need to justify the fact that they're paying for something and if you're not giving them the same value, why would they pay?
speaker-2: It it's interesting, you know, I have friends who've worked in fields where they do physical deployment of technologies, whether it's medical devices or robotics. And a lot of companies will get to the point where they've done a bunch of testing but they never actually get their product into into the field. So like when you guys are doing a pilot, what's the rate of success from pilot to implementation and then you've got a new customer?
speaker-0: So I would say that in the last twelve months we are mainly focusing on Israel, because we are here and the R and D is here, even though we're an American company. here it's one hundred percent rate. All the project that we're doing today is deployable and it's commercial project. same as the project that we did in the US. I would say that the place where we saw it's not objection, it's just a process, is that we had a full full paid pilot in Tokyo at the end of last year that were successful. However, to take a solution to Japan and to deploy it, it's a very long journey that requires a lot a lot of efforts. But today for the robotics, at least in Israel, which is a market that we're very active, we don't do pilots. All the project are paid project with a two to three years contract.
speaker-2: So is just one quick question. in i in Japan, is the different implementation a bureaucracy question that's or is it a question about corporate structure?
speaker-0: Both and you can also add culture. Right? So for a company, for a for a technology company like the robotics that is redefining an existing industry, it's a it's a process because you need to completely localize the company, right? It's not just about changing the language of the dashboard to Japanese, it's transf it's moving the entire cloud to Japan, it's building a Japanese company. It's it's a very complicated process. In contrast to the US, you know, in the US we've met C level a few weeks ago of one of the biggest real estate companies in New York. And within one, two meetings you're closing like the first project, right? And in Japan it's just a big process. However, for me, I think that Japan is going to be one of our most significant markets. Not sure if you spent some time in Tokyo. Tokyo has the highest density of the world. They have a huge problem with labour and if in the US the class A billing owners clean their billing three times a year, in Japan the norm is four times a year. Right? So it's a very promising market for us, but it's something that we are prioritizing probably for Q four next year.
speaker-2: So you guys have built something that's really difficult. I think a lot of people, probably most of the people who hear our podcast or read Kim's newsletter know that, understand AI, but many of them may not understand how hard robotics really is. You guys have built a very complicated a a very good solution for a complicated problem. And you have, you know, worked with NVIDIA, who's like the eighth pounder pound gorilla in ships in this industry. But like what would you say is the one thing that
speaker-0: I would say that the harder things the hardest thing to copy is not one component. It is the entire field and the operation system. A competitor can build a camera, he can build an actuators, it can build a cleaning or you can copy any individual feature. But I think it is much harder to reproduce the knowledge of how to qualify buildings, how to deploy from different buildings, how to manage wind. and obstacles, how to supervise the multiple robots and how to connect all of the data that we've built. Right? I think that one of the unique things that we've had is the dataset of the images that we have. That the use of the images is also for the movement of the robot, but it's also on the inspection mode. And you know it's such a complex environment and and and we still see since we've started Like five years ago there was a lot of companies that unfortunately closed. And we see some companies that are coming and building the solutions today. however they are still doing things that are very similar, right? And I think that the combination of the experience that we have on the last five years, this is a huge defensibility for the robotics.
speaker-2: So you just said five years ago we were here. Five years from now, where are you gonna be? Where's the value in your business?
speaker-0: So five years from today, as I mentioned earlier, it's going to be a hybrid company on the go to market. In some cases we're going to own the entire chain, in some cases we're going to lease it and to have different models. But in five years the vision is to really be leading the category of exterior services. And I think that Looking on the data aspects, it's probably one of the strongest thing that we are going to have. Right? It's the robotics going to work with architects and general constructors and municipalities and insurance companies and really to owning this category. Okay.
speaker-1: Okay. so specialized robots get far less attention than humanoids, obviously, right? So everyone is talking about humanoids routing the laundry, yeah, unfortunately, right? So the question is, are we telling the wrong story about what commercially useful physical AI will actually look like? And what's the one thing you wish more people understood about robotics in the built environment?
speaker-0: Yeah, I am totally agree with what you said. I think that there is a huge huge buzz around humanoids and I think that probably as the the VC ecosystem will change every every year there is a new trend, right? And we often being asked why won't the humanoid do it? Let's first have a humanoid doing things that are more basic. I think that for robot working on a building exterior, it doesn't need to look like a person. It needs to perform a specific, repetitive, dangerous and expensive job. It could have two arms, three arms, five arms, six arms. Yeah. No one cares. Right? We've designed the platform the way it is today to solve an operational problem. And we've believed that the architecture that we have today is the best architecture in the market because it's allowing us to be To have zero dependency on the environment.
speaker-1: It it just crossed my mind. But I mean several companies are building humanoid robots because the world around us is built for us humans, right? So this is why they try to mimic us and our in our physical behavior and whatnot. But looking at at at for example, the the task you're trying to solve, like cleaning difficult areas, I mean it's in in in the first place A place where it's not practical to be human at at all in the first place, right? So maybe you have more room to develop above like humanoid physical stances, so to say, right? So is it like maybe even an opportunity to go beyond having humanoid robots, developing it further? Because you you just mentioned like having more arms or legs or s or something like this, right? It so that the the the difficulty could be an opportunity, right?
speaker-0: Yeah. It's it's also opportunity, but it's also to think in a way to think out of the box. Right. Yeah. you know, if you're looking like in all of this robotics for surgery for medicine. Yeah. So you know the there were there is a companies that started a few years ago doing lycroscopy and other surgeries, which is just kind of a few devices and today you've connect them to a humanoid. I don't know why, maybe something psychology a bit more on the human mind that make us feel like, okay, this is nice. But I don't know, do I want a robot in my home folding my Landerid? Why doesn't the machine do it itself, right? true. So there you you know there is a lot of also I think that the there is also this robots now doing Mills, the guy that sold his company to Amazon or it's Mike Mark Aaron if I'm not mistaken. You know, you build a machine that can produce thousands of dish per day. It doesn't look like a humanoid. Right? I mean o honest honestly, who cares how it look? As long as it's delivering and it justifies the unit economics and it justifies the demand. And it's not let's you know, I'm I was in the C and the NVIDIA event earlier this year. And you see it all over the time. You see these videos of robots that are doing boxing match.
speaker-1: Yeah, yeah. I mean
speaker-0: Okay, I'm sorry to say that, but this is why you invested so much money and so much resource to build robots that will compete on each other, who is beating each other, and this is how you choose to promote it? Solve the problem.
speaker-1: Yeah. Yeah, yeah. I I fully agree with you. I fully agree with with you. And to be to be frank, I'm a little bit I don't want to say annoyed, but probably bored by seeing another robot doing some MMA martial arts stuff or or s or another robot doing some dancing. Whereas as as you rightfully said, so why don't we have like good videos of robots doing Yeah, useful stuff that we as as human need. And I really like what you were just saying. So in a way it forces you to think out out of the box. And this makes totally sense to me. And actually I think this is where we should should be heading and think more in in a way, right? Making use of such technologies and think out of the box.
speaker-0: I think that Figure did something like a few months ago with organizing the mail, if you remember and this was cool. This is what people want to see. I want to see what the robot is actually delivering. Yeah. Right? Yeah. And and I think that most people, you know, always say that it's very similar to to autonomous cars, right? Like in two thousand seven in two thousand ten everyone said by the end of this decade all the cars will be automated. You won't have you won't own a car. Right, but we're almost in twenty thirty. Right? And only now Vegas approved to deploy the couple hundreds of Teslas if I'm not mistaken. So that's true. It's taking a time same as for for the humanoids, which is it's it's a great, but they just still need to prove themselves before it will be a commodity.
speaker-1: Yeah, f first slowly and then all at once. It it it it needs its time to adjust, to develop, but when you when you reach this certain threshold it's like it's going to explode. And hopefully in the field of robotics we're slowly approaching this third threshold where it now really takes off, right? Yeah.
speaker-0: Yeah, definitely. I I totally agree on that.
speaker-1: Peter, you you had a question?
speaker-2: I had one quick question, Ido. It it seems like the most the closest analogy to what you guys are doing is actually the robot that the most people have seen before, which is one cleaning the rug in your house. Yeah or sh mowing your lawn in your backyard. So I it's odd that people are obsessed with humanoid robots to do a task that they're already accustomed to seeing done by something that doesn't look at all like a person.
speaker-0: yeah. I I do agree although that you already see videos today of the humanoids taking the law doing them on, right? I I don't really understand the logic. But yeah, I mean you know iRobot was the pioneers of that. And they did a really huge change in the industry. And you have Dyson, which is probably mm the biggest innovator in the world. That if he's hearing this, I want to meet him because I think that he made the huge impact. And it's really you know, when we started the robotics like a few years ago building the pitch, I always say that yeah, it's like the first vacuum of Dyson. You know, Dyson took out the bag, they you had this ugly plastic bag in the vacuum back there and people said wha no, it's impossible, right? And so there's a lot of shifting also in kind of in in in people understanding, but it was always like this technology. just one less story about innovation, that a few years ago I worked in a company named Bring. And they were kind of the pioneering in less mile technology for SAS companies. So it was really around transforming logistics for company. You know, imagine companies used to work with pen and papers and Google Maps and like zero proof of delivery, zero innovation, I think. Yeah. And we had a big client in Australia and I I was there, I did the the deal there and I also went there for like for onboarding. And you know, I'm explaining about the driver app and route optimization that we save so much time and improving efficiency and Everything, right? Completely transforming the business. And there was a guy in the audience that's like, Hey do you also do 3D scanning for optimizing how I load the boxes into my car? And I'm like, Man, like yesterday you used to go to the printer to print it, to go to Google Maps, to check the address. You have everything automated today, right? And it's always part of people expectation. This is why I said it For us when we do the cleaning people will always try to look more for like a proof, right? If today on the traditional window cleaning most people will say nothing. For us because they know it's a robot, they will even want to do like a bigger investigation.
speaker-1: Mm.
speaker-2: Ido, can I ask I have one qu personal question based on something you just said. Yeah. So you were off working on it for a different company doing a completely different thing, and you're implementing stuff that you learned in your own family business. Yeah. Like you're bringing your experience back. What does it feel like to have this legacy family business that you decided to walk away from and now you've come back with your experience sort of like the prodigal son and you're changing the way it works?
speaker-0: We we need to open another podcast for that to tell them Pastor.
speaker-1: Yeah.
speaker-0: That's that's that's that could be a that could be a Netflix show if you want.
speaker-1: I'd love to do a show about it.
speaker-2: I I asked because one of my f closest friends d took over his family business and grew it by a hundred and fifty times and just using totally different techniques. So I
speaker-0: it's a huge company in the Israeli landscape, it's the biggest. And I always say that leaving the company was the most difficult thing that I ever did in my life. like one hundred percent. And today everything that we've able to accomplish it's by ourselves. And I think it's it's interesting because all of the project, the innovation, the methodology, the technology, the know everything is I mean it could be nicer if we had the support, right? Things would probably happen faster. But it is what it is.
speaker-1: This is a perfect ending. I know this was excellent. Thank you so much. Where can people find their robotics and follow what you are building?
speaker-0: so obviously the robotics.com, LinkedIn, Instagram, Twitter and soon Q4 we are going to have a few projects in the US and when you're looking at the building, it's not a bird and it's not Superman, it's probably our robots over there.
speaker-1: Thank you so much for joining us and thanks to everyone listening. You can find us at get superintel.com. I don't it was such a pleasure. Thank you so much.
speaker-0: I appreciate it guys.
speaker-2: Thank you. Thanks everybody. Thanks for coming and we'll see you next time.
