Imagine when these faces start talking, tracking objects with their eyes with a perfectly synthesized voice all, generated in real time.