gm fam While humanoid robots continue its journey of improvement, teaching them about the world is an entirely new beast A model can learn a lot from roads, data on the Internet, but it still needs fresh information from warehouses, piers, loading areas, shops, streets That's what makes me excited about the idea of @vangrid_io A regular phone can capture these locations, with data getting a content fingerprint alongside with time and location grouped and attested on Base via EAS This gives the footage much better value than being tucked away on a private dashboard Someone can verify that the capture belongs to the published record rather than trust the person who uploaded it And a marketplace ties that into the actual demand Bounties can be requesting certain places, contributors can collect them, and USDC can be released once the requester accepts the result The jobs demonstrate just how specific it can get, from paint booths, pier sections, market corners, signal poles, tactile panels I think it's important that it's not just about gathering more video Creating a ground level data that robotics teams can trust and rely on is a huge bonus Physical AI needs more eyes on the real world Vangrid is trying to turn everyday phones into these eyes #Vangrid #PhysicalAI #airdrop