One of my best friends who is familiar with creating A.I. generated video and is aware of my WAM proclivities was going to send me "pie fight video" for my birthday a few weeks ago but told me that even after a few hours she was unable to create:

1. A woman or man walking over to another woman or man while holding a pie, splatting the pie in their face and having some after shots of it dripping.

AND

2. A pie flying in from off camera and splatting someone in the face.

Coincidentally, these are the same 2 things I seem unable to create. It's truly annoying. It's more annoying than a RobbyWLP corny joke and that's saying something.

I'm currently using a combination of images and KLING.

Thoughts or ideas?

You can call me,

Al

I tried making a few videos on Krea. On some of them and others I've seen here rather than a pie striking the face, the goo suddenly appears and spreads out from the face.

I get that too. Although I have a workaround for it, kinda sorta.

I just can't believe some website won't allow us to do either of the two.

You can call me,

Al

One scenario I haven't managed to get AI close to is a tableau where the model's lost a welly in calf-deep mud, so she's standing in the mud on one leg with a sock on the other foot and the lost welly sticking out of the mud nearby.

Another one is a gunge-filled transparent hazmat suit over normal clothes.

AI seems to have a general issue with things that are in layers. It has obviously got much better at covering people in mess, but can still struggle with 'specific items of clothing covered in specific mess' and in particular your prompt where something needs to be worn over two other things but also contain something and be transparent.

You can have a woman flying into (or SLIDING into) a large pie, but you cannot have a large pie flying towards the face of a woman.

Go figure.