tech
January 29, 2026
I built marshmallow castles in Google’s new AI world generator
Starting Thursday, Google AI Ultra subscribers in the U.S. can play around with Project Genie, which is powered by a combination of Google’s latest world model Genie 3, its image generation model Nano Banana Pro, and Gemini.
TL;DR
- Google DeepMind is releasing Project Genie, an AI tool for creating interactive game worlds.
- Access is currently available to Google AI Ultra subscribers in the U.S.
- Project Genie uses world models, image generation models, and Gemini to generate environments from text or image prompts.
- The release is part of DeepMind's effort to collect user feedback for developing more capable world models.
- World models are considered a crucial step towards achieving artificial general intelligence (AGI).
- The tool has limitations, including inconsistent results, safety guardrails preventing certain content, and a 60-second limit for generation and navigation.
- Interactivity and realism are areas DeepMind plans to improve.
- The race for world model development is heating up with other companies also releasing similar technologies.
Continue reading
the original article