I've used AI video already, but it wasn't good. In one video proportions of object and person were wrong, in other the woman had 4 hands in another video people were showing from nowhere. So for now I can see a lot of mistakes in short videos like this so in the end are useless if I would like to use them.
Another thing - it would be good to have an option to check if my prompt is "legal" as apparently I used some forbidden words (no idea which one) in pretty neutral scene, so I lost HBD and video hasn't generated.
there are some pretty varied quality responses with the model that is provided right now. Some of the leading models out there are substantially more expensive.
There are some things this model does pretty well