[nassimsoftware]

I don't think we have solved image generation with DALLE-2 yet. There are still many challenges before we move on to video:

- generating drawing that upon close inspection don't have obvious defects.

- generate hand drawn text.

- being able to replicate the same character along multiple angles consistently with the same style.

I feel like this is going to become increasingly hard because there are many things in drawing that can't be capture quite well with words.

It's analogous to code being more expressive than no-code drag and drop.