Video models are zero-shot learners and reasoners

Video models like Veo 3 are on a path to become vision foundation models.

Read in full here: