Niantic, the company behind Pokémon Go, is building a "Large Geospatial Model" (LGM) that combines millions of scans taken from players' smartphones. This AI model aims to enable computers and robots to understand and interact with the world in new ways.

The LGM's "spatial intelligence" is built on Niantic's Visual Positioning System, which uses a single image from a phone to determine its position and orientation using a 3D map built from player scans. This data is unique because it's taken from a pedestrian perspective and includes places inaccessible to cars.

Niantic's Chief Scientist, Victor Prisacariu, explained that the company has built high-fidelity 3D maps of the world using player data, including 3D geometry and semantic understanding. This development has significant implications for the tech and startup community, as it could lead to new applications and innovations in AI and robotics.