theCUBE + NYSE Wired: Robotics & AI Media Week | David Pinn, Brain Corp

Clips
More from theCUBE + NYSE Wired: Robotics & AI Media Week

David Pinn

CEO

Brain Corp

play_circle_outline Business growth and mainstream adoption of robots

play_circle_outline Advancements in computer vision and machine learning

play_circle_outline Maximizing Business Insights Through Data Volume, Analytics, Feedback, and Intelligent Merchandising Decisions

Info
Transcript

David Pinn, Brain Corp

David Pinn

CEO Brain Corp

David Pinn, CEO of Brain Corp, shares insights on advancements in robotics and artificial intelligence during their appearance at theCUBE + NYSE Wired Robotics & AI Media Week. With extensive experience in developing innovative software platforms for robotic technology, Pinn provides a deep dive into the future of autonomous systems in commercial settings.

In this discussion, Pinn delves into Brain Corp’s journey and its role as a leader in the robotics industry. With over 40,000 robots in the field, Brain Corp focuses on commercial environments such a... Read more

explore Keep Exploring

What is the history of robots in the retail industry and how have they become more mainstream over time? add

What advancements have been revolutionary to a business in terms of using autonomous mobile robots for floor care and shelf scanning applications? add

What are the major use cases for the data collected by the fleet of robots in terms of business logic and navigation capabilities? add

bolt Powered by CUBE AI

David Pinn, Brain Corp

search

>> Welcome back everyone to theCUBE's live coverage here. The New York Stock Exchange has been a very busy week. It's upside down. Things are changing, the world is changing in robotics and AI leaders is the topic of this week's focus here as part of theCUBE and the NYSE wired communities focus on the leaders sharing their knowledge. It's a mixture of experts and we're all bringing it together. Of course, this is our East Coast studio here in the New York Stock Exchange. Of course, we've got our Palo Alto connecting Tech and Money, Wall Street and Silicon Valley all together. Really, it's the confluence of where AI is going. It's really about tech culture and next-level kinds of things. Brain Corp is here, David Pinn, the CEO, they are the leader at the forefront of robotics, innovation, driving enhancements all around robotics. Again, a lot has to work and they're going to bring it all here on theCUBE right now. David, thanks for joining us here remotely from San Diego.

David Pinn

>> Thank you very much for having me, John.

>> It's been raining in California from what people are telling me, but it's okay weather here in New York, but great to have you on and love your company. You guys have been around for a while. I want to get into some of the things you just got in the field, what you're working on, the breakthroughs, software, hardware, and a variety of other things around the business. But first, set the context, lay out what you guys do, how long you've been around funding, share some of the numbers.

David Pinn

>> Sure, absolutely. And thank you for having me on the show. So, Brain Corp has been around for about 15 years. We've got around 40,000 robots in the field and counting. Those robots do a variety of things, largely in commercial environments. Think retail, schools, hotels, hospitals, warehouses, etc. We clean the floors, we check inventory. We basically help make sure that those environments have a clean environment, the right product at the right place, at the right price, at the right time. So that's the core of our business. What's interesting about our business uniquely in the robotic space is that we're a software only company. We partner with hardware OEMs, partners that bring the physical device to market while Brain Corp provides the operating system, the underlying autonomy layer and intelligence layer to those hardware devices.

>> Yeah, you're the brain OS behind everything. You're the brains behind the operations, hence Brain Corp.

David Pinn

>> You got it.

>> Okay. You guys have quite a journey. Take us through the origination because a lot's changed. You guys have been in the arena, so to speak. From day one, really, when it really starts, the commercialization kicked in, first generation of compute starts to hit the scene, IOT edge, industrial IOT edge, mostly factories and again, workplaces have been putting robots in place for tasks. Some of them end-to-end built, but now, we're getting into a wave of more autonomy, unmanned retails a topic. We're seeing all kinds of cool things. Take us through the progression from the origination now. What's the core story of Brain Corp?

David Pinn

>> Yeah. So, as you're aware, robotics have been around for literally decades. You think about welding robots and automotive factories going back even to the 1960s. What's changed, of course since then, is we're now seeing robots not only in industrial environments but now, in commercial everyday environments. And that's really where Brain Corp has a lead. Operating robots in and around the public, as I mentioned, for example, in a retail environment where safety is paramount. You think about a robot operating in an industrial space. It's typically around highly skilled labor. You've got that black and yellow tape that separates where the machines go from where the people go. That's very different. And in contrast to what you see in a retail space or in a school or in a hotel, there where you're operating in and among the public, safety as you mentioned, is 100% paramount. And so, that's a big transition that we see in the space that we've really led in the space is helping robots move from that industrial type of space to the commercial space. Leading with safety first of course, is extremely critical in these kinds of environments.

>> How has your business changed in the past couple of years? Obviously, it's gone mainstream, Jensen Huang and GTC, his event, obviously, robots are central to his presentation with AI and big AI factories. But in everyday life, what's been the business momentum? Has the customer base shifted, TAM expanded? What has been some of the business dynamics?

David Pinn

>> Yeah, absolutely. Robots are becoming way more mainstream outside of that industrial core. When we got started 15 years ago, most of what we were focused on was R&D and research. We started commercializing our solution about eight years ago, and we were very fortunate to have early adopters like Walmart take the lead on adopting our technology inside of their environment. And as you mentioned, since then, we've gone from the early adopters really to a very mainstream product where most retailers are very comfortable now with having robots operating in their environment. Whether it's to clean the floors, whether it's to check inventory, whether it's to make sure that their merchandising is correctly placed on the shelf.

>> You got to love being the CEO of a company who's got the software, the brains and the market on the hardware side's democratized. One, you've got super computing being democratized and you've got components. There's more sensors, there's more things to connect physically. The AI physical is a hot trend, as a result that opens up the aperture for the autonomous mobile robot. That's a category you guys play in. What's been the go-to-market for you guys? Has there been any growth issues behind it? What's the impact to you?

David Pinn

>> Yeah. So, of course, as this technology matures, the pricing comes down, that helps drive adoption. So we're very excited to see, as you mentioned, all of the advancements on the compute side where you can get more and more compute on device for lower and lower prices, making the economics more and more attractive for our commercial customers. And it's the same story on the sensor side. You see the advancements being made on LiDAR's, for example. When we first got started, a 3D LiDAR costs tens of thousands of dollars. Now, you can get a very decent 3D LiDAR for maybe a couple hundred dollars. And so those kinds of advancements on the compute and on the sensor side really help to drive the cost down and drive the adoption up.

>> What's awesome about this market is that, again, you guys are software. So, first of all, robotics is a great community, very open. People talk and collaborate. So awesome to see that grow and thrive. But now, you're getting into the commercial side where you got an OS you mentioned, and you got software. So you now have that platform-developer relationship evolving quickly. Hardware and software is a key part of the successful robotics ventures that we've seen, certainly from a utility standpoint, from not only exponential scale but getting more capability in there. The role of computer vision and machine learning and now, generative AI, now also piles in. So talk about that piece of the tech because this becomes something that now we think is going to be an accelerator of net new things. So massive step function change on capability existing today, but net new because you're going to start to see applications get fused in. So talk about that platform-developer relationship and then some of the new stuff like vision and machine learning and Gen AI. How's that all intersect?

David Pinn

>> Yeah, absolutely. Yeah, so, all of the latest advancements with transformer models, Gen AI, foundational models has been revolutionary to our business. You think both on the floor care side of our business as well as on our shelf scanning side, the ability for an autonomous mobile robot to really understand its environment, understand what it's seeing, is critical for these applications. On the shelf scanning side, it's all about recognizing which products are on which shelves, what are the prices, et cetera. And in order to do that with a high accuracy that's required for these business use cases, you need really advanced computer vision, really advanced inference engines that can understand the difference between an eight ounce can of soda and a 12 ounce can of soda, for example. Nuances that are very hard for a human to detect with their own eye, these new advancements in Gen AI are able to discern the differences. And so I'm very proud of the fact that Brain Corp's on the cutting edge when it relates to computer vision, both for detecting product on shelf as well as for navigating in the physical space. Just maybe a more mundane example, something fun is that when you're cleaning the floor with a commercial-grade robotic floor cleaner and these things, they're about the size of a riding lawnmower, they weigh about a thousand pounds. You've got to navigate very carefully in these crowded commercial spaces and being able to detect, is this an obstacle on the floor that I should go around? Is this actually a spill on the floor that I should go through and clean up, that requires advancements in computer vision that we've been deploying. And so, we're very excited about this new technology that's come out that's really enabled more use cases for us. Again, both on the floor care side as well as on the shelf scanning side for us.

>> I love the computer vision angle because two things. One, vision is a lot of data. That's a discussion we're going to have, but also, the technology in vision's enhancing. But then, okay, that's hard. Not a lot of people go to school and get a degree in coding vision, but developers who were using software now can access that. So, I'm sure you're seeing a ton of development around that. You're enabling that expertise. Talk about the role of vision, the size of the data, what that does at the software layer, how you guys think about the data volume, how you're handling that from an inference reinforced learning standpoint? And then also, on the technology side around computer vision, depth, stereo is our terms. We're seeing people seeing that in real time. This is real engineering. This is not like write some code, but the coders take advantage of it because now, they can tap into it. So data size and how you're handling the data and then the killer advances in the computer vision, which is the input.

David Pinn

>> Absolutely. And you hit the nail on the head, John, regarding scale, right? You think about what we do for example, in Sam's Clubs, there's over 600 Sam's Clubs here in the U.S. We capture 23 million images a day across those Sam's Clubs on our platform. And so, when you add up those 23 million, images times the size of each image and putting that in a robust data pipeline. Just ignoring for a moment, the AI side of the equation, what you do with that data, just thinking about the amount of scale that's required to handle that data at the edge, in the cloud, being able to guarantee obviously data privacy, security, face blurring, all the things that are required to have an enterprise-grade commercial solution that can handle that level of data is a really complicated problem that we're very proud to be solving. And so, I like where you went with that question because people often forget or they get very hung up on the Gen AI side of it, which obviously, is extremely exciting, but we shouldn't forget as well that there's just some core engineering around getting the right data to the right place at the right time with enterprise grade level data privacy, safety, security, et cetera. That is a big part of this challenge. And so, when you blend that kind of traditional enterprise robustness with all of the latest advancements that are happening in Gen AI, it just becomes a really exciting opportunity for all of us.

>> On the application side, okay, you're seeing a lot of entrepreneurship emerge. Can you share some stories? Stories drive movements and robotics doesn't need a lot of help on movement because everyone loves robotics. There's some people who are scared of it, but that's a different crowd. But most people love robotics. They loves space. They love some of these areas where they see real physical value and they understand it. But on the business side, we're seeing a surge of entrepreneurship because someone might say, "Hey, I know that market. I can take productivity gains and turn them into value. I can take cost efficiencies and turn them into an opportunity to disrupt an incumbent," or frankly, someone who's not even implementing AI that's got a different slower process and more expensive and more risk. Talk about that enablement. Can you share any stories where you've seen this kind of dynamic because that's the application layer? That's the ISV's as they say. That's the next brand that's going to emerge and build the next robot for unattended retail or something we don't even know.

David Pinn

>> No, you're absolutely right, John, because you've got all of this amazing fundamental technology, but it really only comes to life when you apply it to a specific business problem. So, an example that I could draw for you is, just think about the consumer packaged goods companies and how much money they spend on marketing in the retail channel, right? Ensuring that their product is placed in a favorable location. You have a CBG that might pay for an NCAP, for example, and one of the challenges that they face is most retail stores will implement that program. So maybe 80% of a certain retailer's locations will actually put that product on the NCAP where it belongs, but the other 20% don't it. So you've got basically 20% leakage in your investment on product placement and you don't know which stores are doing it and which stores aren't. And today, that's all through shoe leather, right? You've got third-party merchandising companies, paying people to show up to stores, manually take photos, et cetera. It's not a scalable solution. The only thing people can do today really, is check on a sampling basis. With this technology, when you take an AMR plus all of the computer vision capability that you and I have been discussing, you can now automate this process and you can get 100% coverage to make sure that that merchandising program that you put forward as a Cola brand is actually being implemented in the store. So that's just one example, John, of bringing this core technology of AMRs, economic, mobile robots and Gen AI to life in a very specific business application.

>> Yeah. And I've also heard robots don't sleep either, so it's good for evening hours too. So again, back to the economics, you mentioned shoe leather. I love that comment because labor is the shoe, the leather on the shoe, feet on the streets as they say. But also, I want to get into some of the analytics because one value proposition that's popping out is data, the role of data of the actual robot working. So back in the day, we used to call it sneaker net. Not ethernet. Move the file, run it over to the next server and upload it. But now, you have the opportunity with robotics to collectively look at data, and that's become a big theme in the robotics conversations around where the value is. And also, the moat for opportunities for companies building stuff where not only do they have connected ecosystems and all these advantages you're pointing out, but the analytics are compelling. So just think retail, okay, I know what's selling, the weather's coming in, so all these contextually relevant situations can be identified. Stocking. What's selling? What shelf should it be on? All these things come into play. So, how are you guys looking at the analytics in your software? Most of the deployments have some sort of coordinated analytical data lake or some sort of system to one, feed that either back into the robot or as business logic, into the CFO.

David Pinn

>> Yeah, it's exactly right. And you've identified two major use cases for the data. So one is back into the robot. So having a fleet of 40,000 robots means that we see 40,000 different environments every single day. And that data, that telemetry on how the robot is perceiving its environment, behaving with its environment feeds back into our algorithms and our technology and our models around how robots should function. So the more robots you've got out there, the more data you're collecting, the more corner cases you're able to observe and correct for, the better the autonomous navigation works and the more robots you sell, right? So, there's this internal feedback loop that you're describing that's very powerful within the world of autonomous navigation, autonomous mobile robots, et cetera. So that's on the internal feedback loop side. And then you also mentioned the external loop, which is more around the business logic side, and here's something very, very powerful, which is, because of the ubiquity of our robots and the number of shelf images that they are able to capture autonomously, you've got this incredible scale of data. As I mentioned, 23 million images a day just in Sam's Club alone. And when you can combine that shelf intelligence with computer vision, understand for example, this skew or this brand has this percent share of shelf in this environment, and you can compare that against sell-through data, now, you've got a very powerful engine where you're able to correlate the state of the shelf with the sell-through of the product and really make very intelligent merchandising decisions that weren't available before because the data didn't exist. To really figure out what is the exact correlation between being on shelf three versus shelf four, being adjacent to this product versus that product against sell-through at a scale that this data just has never existed before.

>> It's really the convergence between predictive analytics systems and Gen AI real-time data. It's really incredible. Sell-through, just off the charts, my mind's going crazy, just thinking about with a value. The humans would have to be involved in the lag of reporting, rolling it up, squinting through the dashboards. I got to ask you on the data, what have you guys learned? Is there anything that's jumping out at you now that you've got all those autonomous mobile robots in there? And what's the data say? And then second question on that is that another big trend we're hearing in this week and in the industry is that the ability to have multi-purpose robots. In other words, precision craft is coming up, more accuracy, but the ability to not create a monolithic robot. In other words, if a task one is over, it's like the matrix. I want to learn how to fly a helicopter. That scene in The Matrix, everyone knows that if they know The Matrix. But you're starting to see with now LLMs, I can prompt potentially an LLM and get more expertise pumped into the robot. Does that create more opportunities? You guys look at that? Are you thinking about that? So, multi-purpose with an ecosystem that might be available down the road or if you learned from the data?

David Pinn

>> Yeah. So, I would say multi-purpose robot is extremely important, right? Think about a retail environment. Most retailers don't want their sales floor to look like a scene from Star Wars where you've got 100 different robots zipping around. It's a much better consumer experience if you can have one robot that does many functions. And so for us, actually, that's one of the major reasons why we started honestly with robotic floor care, is that every environment has a dirty floor that needs to be clean. And every such environment has challenges of finding labor to clean that floor. And so, you've got this universal base application that requires a robot going up and down the aisles to clean the floor. And then the natural question as you ask is, can that robot do more than one thing? And that's exactly how we got our start in shelf scanning. Is we're going up and down the aisles anyway, we may as well be taking photos of those shelves and analyzing those photos and understanding what's out of stock, what's priced correctly, what's priced incorrectly, is the merchandising laid out the way that it's supposed to be? And so, that's really how we got our start, was thinking about how to begin with a basic application and then stack more use cases on top of that same piece of hardware. And it's extremely powerful from an ROI perspective. You think about one robot, that means one set of motors, one set of batteries, one maintenance contract, and now, you can stack a lot of use cases and applications and software on top of that one piece of hardware.

>> A clean floor is a good robot, as they say. See those marks on the floor, you've got to move around. So great stuff. I got to ask you to wrap up because I love this. We could probably go another hour. We love talking about some of the tech, especially the data, I think is a huge discussion and continue will be. But the entrepreneurship side is key. I noticed you went to Cal, we were talking about that before we went on camera. There's a lot of people in business school right now, or even engineering or just in academic or job skilling, shifting careers. A lot of people are looking at this as an opportunity to either build a product, start a company. What are some of the hot areas that you'd see white spaces or big opportunities to either take down an incumbent using the software coming out like yours and others, that with the advancements of robotics, AI is showing some disruptive enablement in workflows and process that's been defined. That's in the physical world now. So, this is not just build a SaaS app and throw it in the Apple Store or build something on AWS, it's beyond that. It's now digital physical. So, I'm imagining this business plan, competitions, people are writing code right now. What are some opportunities that you guys see that you could share that would get them either more motivated or share data around what to go after?

David Pinn

>> Well, I'll just amplify your premise, which is that a lot of these new Gen AI technologies, they're magical, right? It's just a tremendous advancement that, at least I'll speak for myself, that I wasn't expecting. When we started seeing ChatGPT 2 and ChatGPT 3 starting to come out and the amazing things that became possible with multimodal LLMs, et cetera. It's just incredible that base technology. But to the premise of your question, the base technology, there's a gap... Let me say, there's a gap between the base technology and specific business problems. And so, there is a huge opportunity for my fellow Cal Bears and everybody else in engineering school and business school to really think about specific industries, specific applications, specific problem statements where they have a unique understanding of the intricacies of that problem, and then apply this general technology in a very specific way. In my view, that's really where the value comes, is combining the magic of the general tech with a very specific business problem. And so, I'm completely with you. There's 1,001 ideas out there. I don't want to share them because I want them for myself, but I am absolutely on the right track, that this is really where people are going to make a dent is bridging the gap between the amazingness of this general purpose technology with very specific business problems.

>> David, thank you very much. We'll do our part to share all the opportunities. Again, you guys have got a well competitive lead on everyone, so I wouldn't worry. You guys looking good. Congratulations on the success and thanks for taking the time to share the commentary on the area. And again, as more of the software innovation kicks in, the flywheel there, hardware integration, a lot of builders out there doing building, agents are coming. Again, a lot more software to feed intelligence into the robot. Of course, the data problem and the scale being worked on with more hardware and faster components. So, all good. Thanks for coming on.

David Pinn

>> Thank you very much, John. It was a pleasure.

>> All right. We're talking robotics and AI leaders here on theCUBE and the NYSC Wired community. It's an open community of leaders who are sharing their knowledge, participating and really sharing what's going on so that people can make a better society and also, a better tech. And then as technology continues to evolve, our lives will be changed by robotics in a good way. I'm John Furrier with theCUBE. Thanks for watching.