We demonstrate two multimodal examples here. From a given video context, we predict the future. On the left, we sample a future where the vehicle in front backs away and lets us drive. On the right, we sample we a future where the vehicle in front drives forward.
(video) GAIA-1 Multimodal agent reactions
