For the past eight years, Pokemon Go players have helped make Niantic’s game one of the top titles in mobile gaming. Apparently, players have also been helping Niantic create a 3D model of the world, which is being expanded via AI into a Large Geospatial Model.
As reported by 404 Media, Niantic made the largely unnoticed announcement earlier this month. These plans have seemingly been in the works for a long time, and the company states that it is attempting to “help computers perceive, comprehend, and navigate the physical world.” It also compares Large Geospatial Models to the Large Language Models used by OpenAI’s GPT, Grok, and others.
Because Pokemon Go players are able to visit locations that cars can’t access, they’ve given Niantic views of buildings and areas that other models haven’t been able to see until now. Niantic explained that only the fronts of buildings have been identifiable from previous models, but the Large Geospatial Models can teach computers how to see and identify the other sides of buildings on a global scale.