One part of @vangrid_io I keep coming back to is the privacy + provenance layer. If smartphones are becoming sensors for Physical AI, there’s an obvious question: How do you collect useful real-world data without turning every capture into a privacy problem? @vangrid_io tackles this at the edge. Sensitive details like faces and license plates can be blurred on-device before the data leaves the phone. The capture can then become spatial data with cryptographic provenance, creating a verifiable record of where and when it was collected. That gives the stack an interesting combination: Capture at the edge Privacy by design Spatial reconstruction Cryptographic provenance Verifiable records Because for Physical AI, fresh data is valuable. But fresh + verifiable data is where things get really interesting. @vangrid_io