If I have a good video generation model like OpenAI's new Sora, will it be capable of doing just as well at next frame video prediction?

Full article content could not be extracted automatically. Read the original below.