Possibly, but that is a MASSIVE leap from what it can do currently. From what I've seen they can deepfake faces reasonably well but can spot problems with motion and with the best results having a simple background, OR NVIDIA can create hyperrealistic scenes, but with no people (the people LOOK CGI).
It's also somewhat restrictive because not everyone has the setup with 5-10 high end video cards, and TB's of RAM (like what's needed for many high level training computers).
Possibly, but that is a MASSIVE leap from what it can do currently. From what I've seen they can deepfake faces reasonably well but can spot problems with motion and with the best results having a simple background, OR NVIDIA can create hyperrealistic scenes, but with no people (the people LOOK CGI).
It's also somewhat restrictive because not everyone has the setup with 5-10 high end video cards, and TB's of RAM (like what's needed for many high level training computers).