With the release of Sora it seems reasonably likely that we will have realistic AI generated clips this year. Will we get realistic short films (5-10 minute range) from text prompts this year as well? "Realistic" means both the actual images are accurate depictions of what they're supposed to be depicting, and it can generate films of a "reasonable" complexity level (think multiple characters, maybe 1-3 scene changes, background details, etc). A 5 minute video of a person walking down an empty street doesn't count. No bounds on the complexity of the prompt: an AI that takes an entire script as input is perfectly acceptable.

If I have access to the tool for <= $20 I will test it myself, if not then I will look at public samples.