All it needs to start is a single image prompt either generated by AI or from a real-world photo. There are limitations: DeepMind says the model can generate “consistent” worlds for up to 60 ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback