Name: The Robot Will See You Now
Uploaded: 2025-07-23
Duration: 11 min 39 s

Nils Pihl, founder and CEO of Auki Labs, is developing technology that lets robots navigate the physical world on their own—without human input or cloud-based maps. Together with Terri, the humanoid robot, he’ll demonstrate the posemesh protocol, a system that gives machines a shared, real-time understanding of their surroundings. It’s a step toward a future where robots can move through space with the awareness and coordination we expect from people.

The Robot Will See You Now

Conferences

Hello, hello, good afternoon. Welcome back. Thank you, thank you. Welcome back, everyone. I hope you had a great lunch and even better conversations. I certainly enjoyed mine. To kick off our afternoon, we're going to start by exploring the next frontier of robotics. Oki Labs is developing tech to help robots navigate the physical world on their own without human input or maps. Here to demonstrate, please welcome Aki Labs founder and CEO Neils Peel and his co-presenter, Terry, the robot. All right, hello everyone. My name is Nils. This is Terry. Terry is a unitary robot made in Hangzhou, and today we are going to talk a little bit about how to make the physical world AI accessible, and that's not just about the ramp over there. It has a lot to do with machine perception. I believe that over the next 10 years or so, we are going to see civilization grow to over 100 billion productive participants, most of them AI. And as a result of this growth, I think we will see the world economy grow by something like 10X in the next 20 years. In fact, that may be conservative. Uh, I was born in 1986, let's say 20 years ago, and since then, the world economy has grown 10X. Uh, the world economy has grown 3x in the last 20 years or so, so you only have to believe that we will move about 3 times as fast as we have been to do a 10X in the next 10 years. And Behind the growth that we saw over the last 20 years was that software ate the white collar world. If you are a white collar worker today, you are definitely working on a computer, you're definitely working on a phone, you are definitely using software. Software is a big industry. But 70% of the world economy is still tied to physical space and physical locations, and I actually believe this is going to increase. So What happens next? I believe that faced with the realization that most of human productivity, but also the things that bother us in the world are still tied to physical locations, we can't say that the AI revolution has truly begun. Until Terry can get a better job than being an intern. So let's talk a little bit about what needs to happen for Terry to become productive. We must make the physical world accessible to artificial intelligence. Today, out of the box, Terry is blind. He has a camera, he has a liar, but he has no modality for understanding what he's looking at. He has no, you can't ask him to guide you to the bathroom or anything like this. And we believe that the future humanoid robot stack is going to be made up of 6 layers, and since I only have 7 minutes to go, we'll go through this very quickly. Number one, of course, which you've seen already is we need locomotion. Robots need to be able to walk around without falling over, and I can tell you that I'm going to need to help Terry down the stage today because locomotion is still early. But of course we also need manipulation today. Terry cannot open a can of Coke for you or iron your shirt or anything like this. There's a lot of work left to be done with manipulation, but even if Terry knew. How to manipulate things. You still couldn't ask him to, for example, empty all the trash cans here at the conference. To do that, he's going to need spatial semantic perception as well, not just the camera, but the ability to understand where things are and what things are, and he needs to be able to put that into the context of a map so that the second time you ask him to empty the trash cans, he remembers what the trash cans are and doesn't bounce around like a first generation Roomba. And even when he has a map, he's going to need positioning to understand where he is in relation to the map. A lot of people don't know this, but the GPS is a line of sight technology, so your GPS does not know you're sitting at that table, doesn't know you're in this room, it doesn't know you're on this floor. And then finally we're going to need some kind of applications layer to tie all of these things together. In 2021 at Aki, we realized that you could get a head start on the robotic stack by skipping. The tricky part of locomotion and manipulation. If you wear a pair of glasses. And you have an AI look through your glasses. Then you could do what we call hybrid robotics, where we have human locomotion, you can have a worker that walks as well as me. You can have human manipulation. But you can still have AI perception, you can still have AI mapping, you can still have AI positioning, and you can still run AI applications. In fact, you can create AI co-pilots for physical work. And why wouldn't that be one of the stops on the way to the full AI revolution? Because today, did AI take all the engineering jobs or did AI become a fantastic co-pilot for engineers? And just 2 months ago, we closed our first multi-million dollar deal where we used this stack with just phones and glasses and soon also robots to help retailers understand what is happening in their physical environment. So for the first time ever they can capture an accurate digital twin using spatial computing and AI telling them where their products actually are so they can get better sales reporting, but it also helps them do AR navigation for the staff, reducing the walking distance by a whopping 40% for the people that pick up your online groceries, and they can also do AI and AR task management uh using the system. This is being rolled out into thousands of locations today, and later this year, together with our partner Mentra, who makes smart glasses, we will be releasing an AI co-pilot for physical labor, capable of detecting empty shelves, things that need to get done. So as the retail worker walks through, it turns out we don't replace. The retail worker, we replace the store manager. You get an AI co-pilot that helps you figure out what to do. And that same system is what will allow someone like Terry to understand where are the products and know what to build and where. So what comes after the GPS? Asia is a good place to talk about that. Uh, you need spatial semantic perception, meaning you need to understand where, where things are and what things are, you need to understand the difference between a bicycle and a stoplight. And this hit home for me when I was living in Beijing. Beijing is a big city. I lived there for 7 years. In fact, it's so big that there are more cars on the road in Beijing than there are people in Los Angeles. And if you do a quick back of the napkin calculation based on the average commute, you find that roughly the time it took to build the pyramids. Lost in Beijing traffic. Every single week. And now we might imagine that self-driving cars is going to solve this because cars are going to be coordinating with each other and we won't need red lights anymore, sure, if and only if. These robots, these cars, learn to describe space to each other and their own precise location and the precise location of things that they see. The GPS is never going to work for that because the GPS is a line of sight. This is New York, the biggest city in the West. New York has over 900 buildings taller than 100 m. Astonishing. Until you realize that Hong Kong, where I live, has over 4000 buildings taller than 100 m, and Hong Kong. It's just this little area down here. The rest of this is the Greater Bay Area. It's Shenzhen, Guangzhou, Dongguan. It's an area 2/3 the size of Greater LA but home to 87 million people. And more high-rise buildings than all of North America and all of Europe combined. So it's no wonder why Buddha, that so much machine perception work and GPS replacement work is happening here in Asia by the standards of this century. There are no big cities in the West. So what happens next? We believe that the most important work that civilization can undertake right now is to make the physical world accessible to AI so that we can as a first step, get AI co-pilots for all the physical labor that we wish we could do better and ultimately replace that drudgery with the help of companions like Terry that soon we'll be able to do these tasks autonomously once we've made the physical world accessible to AI. Thank you very much. Nils, that is just so impressive. Hi, Terry. Good to see you. I have just one question for you. You know, you were talking about the development of the robotic technologies and, you know, where we're at and where we're going. By when will Terry be able to bend it like Beckham? Uh, there's already, uh, robot soccer leagues happening sponsored by Booster Robotics. Um, it's not super exciting to watch yet. Terry, I tear you down the ramp in a second. It's like my grandma, you know, uh, needs a bit of help down the stairs, um, but yeah, it's already happening. Uh, robots fighting rings have already started happening. Uh, I think, uh, robot entertainment is, it's now. Yeah, I think I saw a Black Mirror episode about that. Thank you very much. Thank you.