Sixth Sense Labs

An early look of our LiDAR stack in the field

@broodsugar Finally someone who gets it!

This guy gets it.

@PunkRawk777 Cooking!

@SensusSextus ☝️👀

Quiet on the outside. Loud in the lab. We’ve been heads-down building something that changes how robots learn. The first Repos start to drop end of May. With More coming. 👀 #DePIN #DePAI #Robotics #Blockchain.

The bottleneck in robotics isn’t compute. It’s data — and right now, every robot learns alone. We’re fixing that at Sixth Sense Labs. A #DePIN-powered refinement layer that turns real people’s real-world experience into the shared foundation for embodied AI. Natural #autonomy was built together. #Blockchain #DePAI #AI #Robotics.

Embodied AI lacks refinement and a true cognitive capability, at sixth sense labs we aim to solve just that! Soon anyone with a iPhone Pro with “li-dar” will be able to help us collect that that data while earning rewards. But to truly refine embodied AI it needs to see what we see. It needs to comprehend what we do! That’s where smart glass integration comes into play! $PEAQ $AUKI #DePIN #DePAI #AI #Robotics

We may have just created the WORLD'S FIRST REMOTE ROBOT SOCCER GAME with @Auki 😱 We did it using CTRL+R software, Auki's robots & arena, google meet, and a few dozen community members calling from all over the world - Miami, Dubai, Manchester, D.C. and more. All in real time. AND this was our first time testing it out. Only up from here🚀

@neXs_ai @Auki We love Auki!!!

▪️ $AUKI ▪️ ALPHA LEAK 🚨 Spatial intelligence hits the main stage One of the world’s largest retailers has requested a quote for 1,000+ robots: → $50M yearly potential
→ Scaling target: 10K–20K robots @Auki is about to showcase who’s truly pioneering the physical AI era

Let’s try out the @peaq sdk and bring smart glasses on chain!

The momentum behind SixthSenseLabs is real. We’re building the global refinement layer — the missing infrastructure where models meet the messy realities of the physical and digital world. Just as humans depend on collaborative science to learn and advance what we know, intelligent systems need a refinement layer to operate with precision at scale.

Why Robotics Needs a Global Refinement Layer. We’ve spent the last decade teaching robots to see, grasp, and walk. The hardware is finally catching up to the dream. Boston Dynamics flips, Figure folds laundry, surgical bots stitch tissue thinner than a hair. But here’s the uncomfortable truth: every robot is still basically working alone. Each company trains its own models on its own data in its own warehouse. A robot in Tokyo learns to pour coffee. A robot in Berlin learns the same thing, from scratch, six months later. We’re rebuilding the wheel a thousand times in a thousand basements. Humans don’t work like that. When one chef discovers a better way to dice an onion, the technique spreads. We have language, apprenticeship, YouTube. We have a refinement layer — a shared space where individual learning becomes collective intelligence. Robotics has no equivalent. And without one, we’re capping the entire field. A global refinement layer would let robots contribute edge cases back to a common pool — the weird lighting, the slippery handle, the toddler who wandered into frame. Models would improve from millions of real-world hours instead of one company’s curated dataset. Safety failures in Phoenix would patch behavior in Seoul before the next sunrise. The bottleneck in robotics isn’t motors or compute anymore. It’s the fact that every machine is still an island. The companies that figure out how to share learning without sharing IP — that’s where the next decade gets won. Foundation models for language got their internet moment. Robotics is still waiting for its. The hardware is ready. The question is whether we’ll build the connective tissue to match. #DePIN #DePAI #Robotics #AI #Blockchain

This is why robotics needs a global refinement layer!

Soon this will be a standard

Robots need a refinement layer!

Why This Moment Matters The convergence happening right now is unprecedented: foundation models capable of genuine multimodal reasoning, hardware platforms reaching cost-viability, massive capital flowing into the sector, and labor shortages creating economic pull. Companies like Figure, 1X, Agility Robotics, Tesla, Apptronik, and Sanctuary AI are racing toward deployment, while incumbents in adjacent spaces are scrambling to participate. Cognitive perception is the bridge between today’s impressive demos and tomorrow’s deployed workforce. The companies that crack it first—and build the data flywheels to keep improving—won’t just win the humanoid market. They’ll define what physical AI looks like for the next several decades. The hardware is nearly ready. The capital is here. The market is enormous and waiting. What remains is teaching machines to truly understand the world they’re about to enter—and that final unlock is closer than most people realize.​​​​​​​​​​​​​​​​ #Robotics #AR #XR #DePIN #AI #DePAI

The Market Reality, Real-World Applications. The total addressable market for humanoid robotics is staggering when you consider that humanoids can theoretically address any task currently performed by human labor in physical environments. Goldman Sachs has projected the humanoid robot market could reach $38 billion by 2035, while more aggressive estimates from firms like Citi and Macquarie suggest figures north of $200 billion in the same timeframe. Tesla’s Elon Musk has famously claimed Optimus alone could generate trillions in long-term value, and while that’s promotional, the underlying logic about labor market scale is sound. Consider the immediate addressable categories logistics and warehousing alone represents hundreds of billions annually in global labor costs, with chronic worker shortages and high turnover making robotic substitution economically attractive even at premium prices. Manufacturing, particularly in tasks requiring dexterity that traditional industrial robots can’t handle, opens markets in electronics assembly, automotive finishing work, and food processing. Eldercare and healthcare support represents perhaps the most socially significant opportunity, with aging populations in developed economies creating labor gaps that can’t be filled through traditional means. Hospitality, retail, and food service offer high-volume deployment opportunities where cognitive perception is critical for navigating dynamic, human-filled environments. Hazardous environments—nuclear facilities, deep mining, disaster response, offshore platforms—offer high-value early deployment scenarios where the cost of human risk justifies premium robotic solutions. The realistic 2030–2040 horizon could see humanoid robotics establish itself as a multi-trillion dollar industry when factoring in the robots themselves, the supporting software ecosystem, maintenance and services, and the data infrastructure underneath it all. The Hidden Goldmine, On-Site Data Aggregation Here’s the part most market analyses underestimate: the data layer. Every humanoid deployed in the real world is a sensor platform generating petabytes of multimodal data—visual, auditory, tactile, kinesthetic, and contextual. This data is enormously valuable for two reasons. First, it’s the raw material that improves the next generation of cognitive perception models. Just as autonomous vehicle companies have spent billions to acquire driving data, humanoid companies will need diverse, high-quality embodied data to train models that generalize across environments. The first companies to deploy at scale will have insurmountable data moats. Second, the data has external commercial value. Anonymized environmental data from thousands of warehouses, factories, retail stores, and hospitals contains insights worth substantial money to operations consultants, insurers, real estate analysts, and supply chain optimizers. A humanoid that works in a warehouse during the day is simultaneously generating an unprecedented operational intelligence dataset. The companies building the data refinement pipeline—the labeling infrastructure, the simulation environments for synthetic data augmentation, the federated learning systems that improve models without compromising client privacy—represent a parallel market that could rival the robots themselves in value. We’re talking about a potential $50–100 billion annual market just for the data infrastructure layer by the mid-2030s, sitting underneath the much larger robotics market.

Cognitive Perception, The Final Key to Unlocking Humanoid Robotics For years, humanoid robotics has been a story of remarkable hardware progress paired with frustrating real-world limitations. We’ve watched robots do backflips, run through forests, and dance in coordinated routines. We’ve seen dexterous hands manipulate eggs without cracking them. The mechanical engineering challenge, while not solved, is no longer the primary bottleneck. The actuators work. The batteries last longer. The sensors are precise. So why aren’t humanoid robots already restocking shelves at your local grocery store, caring for the elderly, or working alongside humans on construction sites at scale? The answer is cognitive perception—the ability not just to see the world, but to understand it the way humans do. The Hardware-Cognition Gap A humanoid robot today can be thought of as an exquisite athlete with limited situational awareness. It can execute a pre-programmed task in a controlled environment with stunning precision, but place it in an unfamiliar warehouse, a cluttered home, or a dynamic construction site, and it falters. Why? Because raw vision and sensor data are not the same as comprehension. Cognitive perception is the layer that transforms pixels and point clouds into meaning. It’s what allows a robot to look at a kitchen and know that the coffee mug on the counter is something to be carefully picked up, while the spilled coffee next to it is something to be cleaned. It’s the difference between detecting an object and understanding its purpose, fragility, social context, and how it relates to a human’s intent. This involves several interconnected capabilities: Semantic understanding allows robots to comprehend that a “chair” isn’t just a collection of geometric primitives—it’s something that affords sitting, can be moved, might be occupied, and exists within a social and functional context. Intentional reasoning lets a robot infer what a human is trying to accomplish from incomplete cues—a gesture, a half-finished sentence, a glance toward an object. Causal modeling enables prediction: if I move this box, will the stack above it fall? If I open this door quickly, will it hit the person standing behind it? Embodied common sense integrates physics, social norms, and task knowledge into split-second decisions that humans make effortlessly but that have eluded AI systems for decades. The recent explosion in vision-language-action (VLA) models, multimodal foundation models, and world models suggests we’re approaching a genuine inflection point. Once a humanoid can perceive cognitively, the same hardware platform suddenly becomes capable of thousands of different tasks without bespoke programming for each one. Why This Is the Final Unlock Every previous bottleneck in humanoid robotics has been a hardware or narrow-AI problem. Better motors. Faster compute. More efficient batteries. Improved object detection. These were tractable engineering challenges with clear solutions. Cognitive perception is fundamentally different. It’s the capability that transforms a humanoid from a sophisticated tool into a general-purpose worker. Once a robot can genuinely understand its environment, the deployment unlock is exponential rather than linear. You’re no longer selling a robot that does one task—you’re selling a workforce that can be redeployed across countless contexts. This is also the capability that compounds. A robot with cognitive perception generates training data every minute it operates. It learns from edge cases. It improves not just itself, but the entire fleet through shared learning. The flywheel that has powered LLMs—massive data collection, scaled training, deployment, more data—finally becomes possible in physical AI.

SixthSenseLabs is officially live on X 👁️In partnership with @Auki, we're building the perception layer that teaches humanoids to understand the world through real human experience.Autonomy belongs to mankind.#DePIN #DePAI #Robotics #AI

Sixth Sense Labs SixthSenseLabs.xyz Joins $AUKI | @Auki to Accelerate Real-World AI & Robotics Sixth Sense Labs is building on @Auki to rapidly train and deploy humanoid robots using real-world environments and data. $AUKI ➡️ Building the Posemesh: A decentralised universal Spatial Computing + Machine perception network for the next 100billion people, devices and Al 🌐 on earth 🌍and beyond. SixthSenseLabs.xyz ➡️ A revolutionary system combining smart glasses and LiDAR-integrated mobile devices to capture rich, multi-perspective task data for advanced robotics training. Value of the Partnership This partnership strengthens $AUKI position as a foundation for real-world AI by bringing in a team focused on practical humanoid robotics. By enabling fast environment onboarding and leveraging real worker data, it accelerates the development of more perceptive, adaptive robots. In turn, this drives stronger network effects more builders creating real use cases, more data flowing through the ecosystem, and ultimately more demand and utility for the network.

sixthsenselabs.xyz just landed our first retailer meeting to discuss a paid pilot for item processing and sorting. Meeting is set for 2–3 weeks out. A huge milestone — this unlocks our ability to start collecting real-world data from workers doing real-world work. This kind of data is essential to unlocking true cognitive abilities in embodied AI.​​​​​​​​​​​​​​​​ This will give us the ability to further enhance @Auki and its ability to cognitively perceive real world work!