Nuked Hello, tech enthusiasts! It’s your funny guy Nuked here, ready to dive into some exciting news from the world of technology. Let’s get started!
Niantic is up to something big! They’ve announced the creation of a new “Large Geospatial Model” (LGM) that will utilize millions of scans gathered from Pokémon Go players and other Niantic games. This innovative AI model could revolutionize how computers and robots perceive and interact with our world, as highlighted in a recent blog post from the company.
At the heart of this LGM is what Niantic calls “spatial intelligence,” which is powered by neural networks developed through their Visual Positioning System (VPS). Over the last five years, Niantic has been hard at work building this system that uses a simple image from your smartphone to pinpoint its location and orientation using a 3D map created from user scans of interesting spots in their games and Scaniverse.
What makes this data so special? Well, it’s collected from a pedestrian viewpoint and includes areas that cars just can’t reach. Niantic’s Chief Scientist, Victor Prisacariu, elaborated on this in a 2022 Q&A, explaining how they’ve crafted high-fidelity 3D maps of the world using data uploaded by players of Ingress and Pokémon Go. These maps include not only the shapes of objects but also their semantic meanings—like identifying whether something is ground, sky, or a tree!
As noted by 404 Media, it’s pretty wild to think that when people downloaded Pokémon Go back in 2016, they had no idea that their gameplay data would someday contribute to such advanced AI technology. Who knew catching virtual creatures could lead to real-world innovations?
So there you have it! Niantic is taking us on an unexpected journey from gaming to groundbreaking AI. Keep your eyes peeled for what comes next!
