AI can write, code, research, and converse. But for AI to become truly useful in the physical world, it needs to understand something humans take for granted: space.
Where am I? Where are you? Where is the product on the shelf? Which way should a robot move? What object is this, and how does it relate to everything around it?
These questions are at the heart of spatial intelligence—and they are central to the work being done by Auki Labs.
In this episode of The Edge of Show, host Josh Kriger sits down with Nils Pihl, CEO of Auki Labs, to explore how robots, augmented reality devices, phones, and AI systems can develop a shared understanding of the real world.
The conversation moves from humanoid robots in Hong Kong and AI-powered retail operations to decentralized physical infrastructure networks, shared machine perception, and a future where physical AI could transform cities, workplaces, and daily life.
At the core of Auki’s vision is Posemesh: a decentralized protocol designed to help devices communicate about physical space.
Why AI Needs Spatial Intelligence
Today’s AI systems are incredibly capable in digital environments. They can process language, analyze images, generate content, and assist with workflows. But physical environments are far more complicated.
A robot cannot simply rely on text prompts to navigate a store, deliver a package, identify an empty shelf, or understand where a person is standing. It needs a reliable, up-to-date understanding of its surroundings.
Humans create these spatial maps naturally. We can look around a room, understand where objects are, recognize obstacles, and communicate directions with other people. Machines, however, do not automatically share this ability.
As Nils explains, a robot is ultimately a computer—and a computer does not inherently know where it is.
GPS can help with broad positioning outdoors, but it has major limitations. It is not precise enough to place a robot within a specific aisle of a supermarket or help it identify the exact location of a product on a shelf. It also struggles indoors and in environments without a clear line of sight to satellites.
That is where spatial computing and machine perception become essential.
The challenge is not only teaching one robot to understand a space. It is enabling multiple devices to share the same spatial understanding.
A robot, a pair of smart glasses, a phone, and another robot may all see a room differently. Posemesh is designed to help them establish a shared frame of reference so they can communicate meaningfully about where things are in the real world.
What Is Posemesh?
Nils describes Posemesh as a type of “real-world web.”
The traditional web made digital information searchable, browsable, and navigable for people. Posemesh is designed to do something similar for physical spaces and physical AI.
Through the network, venues can create and maintain maps of their environments. Devices that enter those environments can use that information to understand where they are, navigate more effectively, and contribute useful data back to the shared map.
Instead of one company trying to create a single centralized map of the world, Auki’s approach is decentralized. A venue, robot, or environment can maintain its own spatial data while connecting to the broader network.
This is especially important for the growing physical AI economy.
If autonomous machines are going to operate in retail stores, warehouses, airports, hospitals, factories, hotels, and cities, they will need more than a map. They will need a way to communicate about the physical world with other devices and with people.
Posemesh is built to support that shared perception layer.
From Smart Glasses to Robots: Shared Perception in Action
One of the most compelling examples from the episode involves smart glasses and a robot.
Nils describes an Auki Labs demonstration in which a person wearing smart glasses could direct a robot to come to them. The robot could not directly see the person, but the glasses were able to share their spatial position through the protocol.
As a result, the robot could understand where the wearer was located and navigate toward them.
This type of interaction points to a much larger future for augmented reality.
Smart glasses can become an interface between humans and physical AI. Instead of opening a separate app, studying a dashboard, or giving a robot complicated commands, users could interact more naturally through the environments around them.
Digital information can be anchored in physical space. A worker can see the location of a task. A robot can identify where assistance is needed. A new employee can follow navigation cues to find an item in a large store.
The goal is not simply to make machines smarter in isolation. It is to make human-machine collaboration more intuitive.
Retail Copilots and Robot Store Managers
Auki Labs is already applying this technology to retail.
In its lab, the company has built a mock grocery environment to test how phones, smart glasses, and robots can work together inside real stores. One of its key products is an AI copilot for retail workers.
The system helps create and maintain a detailed, real-time map of the store—not merely where headquarters planned products to be placed, but where items are actually located.
That distinction has major operational value.
In a large supermarket with tens of thousands of SKUs, it can be difficult for even experienced employees to find a specific item quickly. For a first-day staff member, the challenge can be much greater. With spatial guidance on a phone or pair of glasses, workers can follow directions directly to the relevant product.
The data can also help retailers better understand store performance. By linking product placement and sales information, companies may be able to see which parts of a shelf are generating revenue and which areas need attention.
Auki Labs is also expanding into semi-humanoid robot store managers.
These robots are not designed to replace workers by performing every task themselves. Instead, they can move through a store, identify operational issues, and alert human employees through augmented reality interfaces.
For example, a robot may identify:
- An empty shelf that needs restocking
- A display that needs tidying
- A cleaning issue
- A location that requires human attention
- An operational task that may otherwise be missed
The concept is simple but powerful: robots can help ensure important tasks are noticed, while people remain responsible for the work that requires human judgment and action.
Can Robots Reduce Workplace Friction?
One unexpected insight from the conversation is the potential social benefit of AI-assisted task management.
Nils shares feedback from a major sports apparel retailer that was interested in using robots and AI not only for efficiency, but also to reduce workplace tension.
When a manager repeatedly tells employees what to do, those instructions can sometimes create friction. Employees may feel micromanaged, and colleagues may become resentful when they must supervise one another.
The retailer’s theory was that a neutral AI system or robot could make task reminders feel less personal.
People are rarely angry at a calendar reminder for telling them about a meeting. Similarly, a robot identifying an empty shelf may feel less emotionally charged than a manager pointing out that someone missed a task.
The idea is still being tested in real-world environments, but it raises an important question about the future of workplace automation: can AI be designed not just to improve efficiency, but to improve collaboration?
For Auki Labs, that is a central part of the mission.
Building Intercognitive Capacity
Nils describes Auki Labs as more than a robotics or augmented reality company. He calls it a “memetic engineering company” focused on improving what he calls intercognitive capacity.
In simple terms, intercognitive capacity is the ability of people to think, solve problems, and work together more effectively.
Technology has often focused on empowering the individual. Personal computers, smartphones, and AI assistants can help individuals create, communicate, and work more efficiently.
Auki Labs is interested in what comes next: technology that improves the capacity of groups.
How can teams share context more effectively? How can physical environments become easier to understand together? How can people, robots, and AI systems coordinate without adding more confusion?
Spatial computing may be an important answer.
When everyone involved in a task—human or machine—can understand the same environment, collaboration becomes easier. A worker can see what needs attention. A robot can communicate what it observes. A manager can make decisions with clearer information.
This is the human-centered side of physical AI.
The DePIN Opportunity: Infrastructure Owned by Networks
The episode also explores the role of decentralized physical infrastructure networks, often known as DePIN.
Nils is candid that he does not identify primarily as a “crypto person.” However, he is strongly optimistic about DePIN as a long-term technological and financial innovation.
The basic idea is that infrastructure can be built through distributed participation instead of relying entirely on a centralized organization.
Rather than one company deciding where every piece of infrastructure should be deployed and taking all the financial risk, a decentralized network can allow participants to contribute infrastructure where they believe it will be valuable—and share in the upside when the network grows.
Nils points to projects such as Geodnet and Helium as examples of infrastructure networks that demonstrate this model.
For Auki, decentralized infrastructure can help support the physical AI ecosystem. Community members can contribute spare compute capacity through nodes, enabling devices and robots in the field to access additional computing resources when needed.
The company’s dashboard, according to Nils, shows where community-operated nodes are helping support the network.
This model could become increasingly important as robots and AI devices require more real-time processing power, mapping data, and coordination.
A Tokenized Model for the Real-World Web
If devices and robots are sharing information on a real-world web, how does that network create value?
Nils explains that robots do not watch advertisements. So a spatial information network may need a different economic model than the consumer internet.
In Auki’s model, when devices exchange useful spatial data, they can interact through the Auki utility token.
For example, a third-party robot entering a venue may want directions or information that saves time, battery power, and computing resources. The venue can provide that data in exchange for a very small transaction.
This creates a potential machine-to-machine economy, where devices can pay for useful information as they operate.
The practical goal is to align incentives for people and venues to maintain high-quality spatial data while enabling devices to access the information they need.
Physical AI Could Be Bigger Than Digital AI
Nils makes a bold prediction during the interview: the biggest AI company of the future may not be a large language model company. It may be a physical AI company.
His reasoning is that a substantial share of the global economy is connected to physical locations and physical labor. Retail, logistics, construction, manufacturing, transportation, hospitality, healthcare, and infrastructure all happen in the real world.
AI copilots have already begun transforming digital and white-collar work. The next wave may bring similar tools to physical and blue-collar environments.
Imagine workers equipped with smart glasses that offer real-time, location-aware assistance. Imagine robots that can identify issues but leave people in control of the final decision. Imagine city systems where autonomous vehicles can coordinate traffic more efficiently because they share a clear understanding of space.
Nils describes Beijing traffic as one example. If autonomous vehicles could communicate precisely with one another, traffic signals and road systems could eventually become more coordinated and efficient. That could return enormous amounts of human time and productivity.
Whether that future arrives quickly or gradually, the underlying requirement remains the same: machines need to learn how to understand and communicate about space.
The Future of Auki Labs
Auki Labs is focused on three major themes:
- Smart glasses
- Store manager robots
- The Exocortex
The Exocortex is described as Auki’s open-source internal AI tooling for improving collaboration and shared context within complex organizations.
Together, these efforts reflect the company’s broader objective: build tools that allow people, machines, and organizations to understand their environments—and each other—more effectively.
The physical AI future will bring serious questions about work, privacy, human agency, and culture. Nils acknowledges that technology is moving faster than society’s ability to fully understand its implications.
His guiding principle is to “spread good memes”: to be mindful of the behaviors, narratives, and systems we amplify as technology reshapes the world.
That is an important reminder. The future of robotics will not be determined only by technical capability. It will also be shaped by the values we build into the systems around it.
Key Topics Covered
- How robots, phones, smart glasses, and AI can share spatial understanding
- Posemesh and the concept of a decentralized “real-world web”
- Why GPS alone is not enough for indoor and high-precision robotics
- Retail AI copilots, augmented reality guidance, and live store mapping
- Semi-humanoid robot store managers
- The potential for AI to reduce workplace friction
- Intercognitive capacity and technology that helps groups work better together
- DePIN, decentralized compute, and physical AI infrastructure
- Machine-to-machine micropayments through the Auki utility token
- Fractional ownership and financing models for robot fleets
- The Exocortex, smart glasses, and the future roadmap for Auki Labs
Episode Highlights
- Machines need shared spatial context. A robot, phone, and pair of smart glasses must be able to understand the same environment to collaborate effectively.
- Retail is an early real-world use case. Auki’s technology helps workers locate products, map stores, and respond to operational issues.
- Robots can support—not replace—human workers. Store manager robots are designed to identify tasks and guide human attention.
- DePIN can help build physical AI infrastructure. Distributed networks may provide compute, mapping, and other services for the machine economy.
- Physical AI may be the next major frontier. As AI moves from screens into stores, cities, and workplaces, spatial intelligence becomes foundational.
People and Resources Mentioned
- Nils Pihl — CEO, Auki Labs
- Josh Kriger — Host, The Edge of Show
- Auki Labs — auki.com
- Posemesh
- Auki Network dashboard — dashboard.auki.network
- Auki Discord — discord.gg/auki
- Geodnet
- Helium
- PEAQ
- Ex Machina
- GAIB
- Richard Dawkins and The Selfish Gene
- Anthropic Claude
- Perplexity
- Neuralink
About Nils Pihl
Nils Pihl is the CEO of Auki Labs, where he works at the intersection of spatial computing, augmented reality, robotics, decentralized infrastructure, and human collaboration. His work on Posemesh focuses on helping devices, robots, and AI systems develop a shared understanding of physical space.
Listen to the Full Episode
Watch or listen to the full episode of The Edge of Show for a deeper look at Posemesh, physical AI, robot store managers, augmented reality, decentralized infrastructure, and the future of machines that can understand the world around them.
To learn more about Auki Labs, visit auki.com and join the community’s regular live demos and AMAs on Discord.
.jpg)
.jpg)
.webp)
.webp)
.svg)
