The way we create and explore virtual worlds is evolving at lightning speed. From rapid hackathon prototypes to cutting-edge AI research, developers and companies are experimenting with technologies that turn imagination into immersive, explorable environments. Recently, builders and researchers alike have demonstrated just how far AI-generated worlds can go: from robotics simulations and interactive museums to near-infinite AI-driven landscapes.
The first-ever World Labs Hackathon (WL-HACK 01) brought together 70 hackers in San Francisco to experiment with Marble, a platform for building interactive worlds. In just 3.5 hours, 32 teams created a dazzling array of projects spanning AR/VR interfaces, games, art experiences, and real estate tools.
Best projects included:
- Musée du Monde – visitors can step inside famous paintings, from Van Gogh’s bedroom to worlds inspired by Vermeer and Matisse.
- Jar of Marbles – a semantic interface for organizing Marble worlds into clusters, navigable via computer vision gesture controls.
- Augmented Virtuality Room – a robotics simulation where a robot uses LiDAR to explore a Marble environment, generating semantic maps and commenting on its surroundings in real time.
The hackathon highlighted the creativity and potential of AI-driven world models, demonstrating how quickly innovative ideas can take shape when builders have the right tools.
Google is also advancing interactive AI worlds with Project Genie, an experimental prototype now available to Google AI Ultra subscribers in the U.S. Powered by Genie 3, Nano Banana Pro, and Gemini, Project Genie enables users to create, explore, and remix AI-generated worlds in real time.
Unlike static 3D snapshots, Project Genie dynamically generates the environment as users navigate, simulating physics, interactions, and environmental changes. Sessions are currently limited (often around 60 seconds in early testing), but worlds can be remixed, explored in curated galleries, or regenerated for new experiences.
Project Genie offers three key capabilities:
- World Sketching: Generate worlds from text or uploaded images, and preview them using Nano Banana Pro to fine-tune style and perspective.
- World Exploration: Move through the world in first- or third-person perspectives while the AI renders the path ahead in real time.
- World Remixing: Transform existing worlds with new prompts or create entirely new experiences, then export explorations as videos.
Though still experimental, Google emphasizes responsible development, noting occasional input lag, limited session duration, and some inconsistencies with physics or character behavior.