OpenAI has a Studio Ghibli problem with ChatGPT images - The Verge.

Midjourney lets people avoid anime (or any other style) by choosing what kinds of images they like and don't like in an easy tuning session. They just look at a bunch of pairs of images and pick which one that they like better. After dozens of those choices, they can turn that personalization setting on or off and get their own unique style.

Midjourney has also consistently given much better looking images without that Uncanny Valley AI look that most other AI image generators tend to have. We'll see how the new ones do.

Speaking of anime, Midjourney also has a special fork (Niji) that lets people only do anime. That can also be fine tuned to avoid things like the cutesy, huge eyed anime.

I have mixed feelings about AI, but I have to admit that Midjourney is fun to mess with. Mostly by feeding it bad puns.