Of all the musicians I’ve shared the stage with I’ve never played with a Xylophonist
Which is closely related to what I do as a competent percussionist
No doubt I’ll find one, one afternoon
And marvel at them hammering out a melodic tune
And hopefully not have both of us be drowned out by an out-of-control saxophonist
I decided to dip my toes into Chat-GPT’s brand new free-to-use Image generator (as of 2026-04-23) and gave it a very challenging task of creating not one but five panels of story within one image and – not only that – to do it in the Wimmelbuch comic style of highly concentrated background-action imagery. (the full prompt is down below)
I gave specific instructions about the main and secondary character sizes and about incorporating the relevant text from my poem, but otherwise made no reference whatsoever to the placement of characters, or about the setting or location or about gender and ethnicity, or typeface or text location or about secondary style choices.
It’s been tough for AI to manage large workflows and deep context windows in the past, so IMHO, this looks like a step change in progress to me.
Of course, as can be expected, the background characters are where you see the AI fall apart. There are misshapen limbs and floating body parts to be found all over, and the faces of the people are lacking in detail, but for what it is, it stands up to cursory inspection. And if you think back to very recent times when AI image generation was disastrously bad at anything to do with text in images, this is amazing. Obviously, the text I wanted is visible and it is consistent and coherent, but so are other hidden text details that the AI dreamed up for itself. They are relevant to each scene, and they work.
In so many metrics this is a proper leap forward.
Interesting times…
THE FULL PROMPT
Make five double spread pages from a comic in the Wimmelbuch style using the 5 lines of a poem (at the end of this prompt) as the text. Incorporate the text for each respective spread, into that spread.
For spread one the main character is a Xylophonist. For spread two the main character is a Percussionist. For spread three the main character is the Percussionist again and the Xylophonist is a secondary character. For spread four the Xylophonist is the main character and the Percussionist is a secondary character. For spread five a Saxophonist is the main character and the Percussionist and Xylophonist are secondary characters. For every spread, make sure the main character is featured large in the foreground acting out whatever the line says about them. Secondary characters must be about half the main character’s size, and the background characters must all be much smaller, in typical Wilmmelbuch style. Make sure that each main and secondary character and their equipment and instruments are anatomically and technically correct. Do not incorporate the input within this prompt that is bracketed by underscores, into the comic – the bracketed input is for your guidance only.
_poem begins_
Of all the musicians I’ve shared the stage with I’ve never played with a Xylophonist
Which is closely related to what I do as a competent Percussionist
No doubt I’ll find one, one afternoon
And marvel at them hammering out a melodic tune
And hopefully not have both of us be drowned out by an out-of-control Saxophonist
_poem ends_