Google DeepMind’s new Genie integration starts with a map coordinate and a visual style, then turns Street View imagery into a small interactive world rather than a fixed video clip.
The key asset is scale: 280 billion images across 110 countries give Google a spatial anchor that rivals cannot simply scrape or buy.
The first use cases are Waymo-style rare-scenario training, robotics simulation and professional walkthroughs, while the limits remain visible: game-like fidelity, no full physics engine and Street View timestamps that may be years old.
Sources: TechCrunch, Google Blog, Pasquale Pillitteri, TechTimes, TNW; CocoLoop