As far as I know it is possible right now. There are some really impressive video generative AIs out there. Us regular people probably cannot do it, as nobody who runs such a model would let us use their resources to that extent.

But if someone with such a model decides to go all in, they could probably do so now. However, “now” does not mean instant. They could push the button now, and then the computers start to churn, and considering the size and complexity of a full movie it’s going to take a while.

And then you have the first pass done. The first of what will probably be many. I find it hard to believe that you’d create the entire movie in one go. Because if the AI fucks up just a few details, the entire movie will be garbage. It is better to do it scene by scene, and then stitch them together afterwards. So you run multiple passes for every scene until you have what you want, you stich them together, and then do whatever else you need to do to release this end result.

I’m sure the NVIDIA and AMD CEOs are currently having wet dreams of someone deciding to do this, because it’s going to cause GPU prices to skyrocket.