I played around with the AI image generator, and what I found out was the same prompt yields totally different results in either of my two publications.

Prompt: Anthropomorphic Rat smoking a cigar after a battle, style: etching

First result: Rat Says; second result: Conspi-Rat

The former produces great pictures, the latter random weirdness.

I wonder if it’s related to the fact the former publication has paid subscribers (er, in fact a paid subscriber).

Anyone willing to take part in an experiment? ;)