Im not convinced by these prompt based image mash ups.
A bit of fun sure but not the future.
The future of AI will be a in tools to speed up current processes.
For example if it was those Cheetos Crocs ai might be able to recognise you want to fill the holes with black spots after you have done one yourself. Or the Mario LV one you might be able to tell the software to repeat and wrap the LV pattern you input into it over a complex shape.
I agree with livdiv, obviously. The only thing trash like this achieves is potentially give succour to livdiv come BnB Santa ozno panick time, as I might go this kind of duff AI instead of whatsapping him to make an image for me.
Interesting use-case for Midjourney (plus After Effects) - long ambient videos.
There must be people who use tv/monitors/signage on in the background all the time:
I would concede someone is trying to make a fair go of making decent art with LLM tools in this instance, talk about exception proving the rule though.