Genie 3 is a “general-purpose world model capable of generating diverse, interactive environments.” Google is now letting AI Ultra subscribers in the US access it with Project Genie.
A world model simulates the dynamics of an environment, predicting how they evolve and how actions affect them. While Google DeepMind has a history of agents for specific environments like Chess or Go, building AGI requires systems that navigate the diversity of the real world.
This experimental research prototype has you describe your environment (“What does your world look like?”), including how you want to explore it — walking, riding, flying, driving, etc. — and first/third-person point of view.
After specifying your character (“Is it a person, animal, object, or something else?”), Project Genie creates a preview image, or World Sketching, using Nano Banana Pro. This lets you preview what “your world will look like and modify your image to fine tune your world prior to jumping in.”
You then “Create world” with users limited to 60-second sessions. The photorealistic worlds are in 720p resolution, while you get interaction rates of 20-24 frames per second.
When you move, Genie 3 (first previewed in August) “generates the path ahead in real time based on the actions you take.” Google is simulating the physics and interactions with “breakthrough consistency.” You can adjust the camera as you interact with the world, with the ability to download videos of your walkthrough.
Another feature lets you Remix Worlds:
Remix existing worlds into new interpretations, by building on top of their prompts. You can also explore curated worlds in the gallery or in the <randomizer icon> for inspiration, or build on top of them.
Besides the 60-second limitation, Google also warns about how:
- “Generated worlds might not look completely true-to-life or always adhere closely to prompts or images, or real-world physics”
- “Characters can sometimes be less controllable, or experience higher latency in control”
Google is working to improve Project Genie with “promptable events that change the world as you explore it.”
- Environment prompt: An endless ocean of immense, thundering waves features colossal turquoise barrels breaking under dramatic sun flares, swarming with hundreds of seagulls that fly close enough to momentarily obscure the view.
- Character prompt: The nose of a white surfboard that slices through the water, pointed directly into the hollow barrel of a massive, breaking wave.
This demo will allow Google to “better understand how people will use world models in many areas of both AI research and generative media.” Access begins “rolling out today to Google AI Ultra subscribers in the U.S. (18+), expanding to more territories in due course.”
…our goal is to make these experiences and technology accessible to more users
More broadly, world models are part of Google DeepMind’s AGI mission. Simulating real-world scenarios has practical applications for “robotics and modelling animation and fiction, to exploring locations and historical settings.”
- Environment prompt: A high-altitude open world featuring deformable snow terrain.
- Character prompt: An agile alpinist with omni-directional movement and jump mechanics.
FTC: We use income earning auto affiliate links. More.
Comments