(Edit: see JacobiX reply, looks like it robbed the concept, possibly because I didn't give it any direction.)
I just prompted GPT-3 (davinci-002 engine on openai) to write a children's book and include illustration suggestions. I didn't even suggest a topic, just said 'children's book'. I fed the illustration suggestions into Midjourney with the qualifier of 'children's book illustration of ...' and pulled the first image of each (it always generates four).
I then asked it for a cover page, last page and a background image for the text.
I assembled the story text written by GPT-3 and images rendered by Midjourney in this google slides preso. The only real creative input i had was the font. I did iterate on the text background image prompt a couple of times to find something that would work:
(Note this link works in incognito if you prefer not to have your google login associated)
https://docs.google.com/presentation/d/e/2PACX-1vT4XWNx2SdEg...
I've been having a lot of fun with this prompt. Part in asterisks is where you set the topic. The last 'Title of the book:' is where you submit, it really helps keeping the generative part on track.
----
Write a four page children's book with the following:
Title of the book:
Description of the cover page with enough detail for an illustrator to draw:
Description of the background imagery used for the pages with text on them:
First page text in quotes:
Description of the opposite page illustration:
Second page text in quotes:
Description of the opposite page illustration:
Third page text in quotes:
Description of the third page illustration:
Fourth page text in quotes:
Description of the fourth page illustration:
Description of the final page:
The subject of the children's book is *how to prevent buffer overflows in c code*.
Title of the book: