
@evgeniirofe @EHuanglu There are multiple vision capable LLM models which can reason on long videos. It can simply analyze it's own creation, then back to the cutting room. Rinse and repeat. In theory it could work, how valuable the result will be, that's another question.
English






















