Join the movement and participate on Uniswap.
Participate
We're back from another intense week of travel and speaking engagements, from podcasts in Shenzhen to giving a lecture on AI strategy to visiting Swedish CEOs in Hong Kong. We also released a new article titled "Move Fast and Break Robots," diving into the strategic differences between Western and Chinese approaches to robotics hardware testing and iteration.
One of the biggest advantages China currently holds in the robotics race is the speed of iteration. A major factor is their willingness to simply break hardware to learn faster. In the West, decision-making processes are heavily bogged down by consensus building and long timelines. As our CEO pointed out, "Western venture capital is right now irrationally afraid of letting you spend money on hardware that might not pan out right."
Western VCs might approve a $1,000,000 budget for compute to test a new world model architecture without blinking, but applying that same iterative budget to physical robots that might break causes immense hesitation. To compete in physical AI—which has a vastly larger total addressable market than software-only AI—the West must learn to act with more urgency. We must embrace the fact that "breaking some hardware is just part of the game," much like SpaceX accepts exploding rockets as part of their iterative process to reach Mars.
Long-term strategically, collaborative maintenance of spatial maps is critical. It’s not enough to stitch a map together once; we need the ability to partially update a map as the physical world changes.
This week, we demoed subtractive scanning. In a baseline scan, a shopping cart was mapped into the environment's point cloud. When a subsequent scan of that same area was uploaded, the cart had moved. The new scan correctly omitted the cart from the point cloud.
Because of our decentralized infrastructure, these updates are highly scalable. In our demo, the new scans were processed by two separate public reconstruction nodes, meaning node operators earned $AUKI tokens for keeping the domain up to date. You never have to reconstruct an entire large venue from scratch—only the parts that have changed.
To improve our computer vision and QR code scanning capabilities, we've built a new simulated testing environment. The simulation overlays QR codes in a 3D game world, allowing us to debug how the scanner behaves under various conditions—such as dealing with reflections on glossy floors.
This setup provides an environment somewhere between live, unpredictable testing and fixed datasets. It allows us to pinpoint exactly why a scanner might drop frames or lose detection when the camera isn't moving, pushing our computer vision robustness to the next level.
Auki is making the physical world accessible to AI by building the real world web: a way for robots and digital devices like smart glasses and phones to browse, navigate, and search physical locations.
70% of the world economy is still tied to physical locations and labor, so making the physical world accessible to AI represents a 3X increase in the TAM of AI in general. Auki's goal is to become the decentralized nervous system of AI in the physical world, providing collaborative spatial reasoning for the next 100bn devices on Earth and beyond.
X | Discord | LinkedIn | YouTube | Whitepaper | auki.com
