Imagen Video and Phenaki: Google answers on text-to-video
In one sentence A week after Make-A-Video, Google Research unveils Imagen Video and, around the same time, Phenaki: two different approaches to text-to-video, with longer, more coherent clips.
Six days after Meta, Google shows its answer on text-to-video. Two answers, actually: Imagen Video, built on the same approach as its image model, and Phenaki, which can produce longer videos that change scene (for example "a person walks in the desert, then enters a city").
Both are research demos, not products. Google again chooses not to release anything to the public, worried about misuse.
The message is clear: the text-to-video race has started, and three big labs (Meta, Google, Runway) are working on it in parallel. The public will see real fruits only from 2023 onward.
Companies
Google Research
Tools
Imagen Video, Phenaki
Tags
Sources