Image corresponds to a sentence or sometimes just a single word. What ChatGPT does is closer to video generation - a much harder task. I expect video generation

Image corresponds to a sentence or sometimes just a single word. What ChatGPT does is closer to video generation - a much harder task. I expect video generation models to be much larger than LLMs.