On Sunday, 11 January 2026 at 14:56:07 UTC, Mike Parker wrote:
### AI interlude
As we were about to close, Walter told us he’d tried very hard
to get an AI to draw a D-man. He'd failed miserably no matter
how he crafted the prompt. He kept getting something that
looked like Superman with a D on his chest.
Carsten offered to ask someone at his company who was "super
good at that" to see if they could come up with a prompt that
worked. I mentioned that some of the services let you upload a
reference image. Walter said he’d tried that, too. He'd tried
everything. All these people were saying AI was going to take
over everything, and he hadn't been successful with getting it
to do what he wanted.
[…]
Walter said
[…] He was just annoyed that he couldn't get it to generate a
D-man.
He said we had a shortage of D-man cartoons. He had a directory
with a collection of all the ones he'd seen. For him to draw
one was a long, complicated process because he had no talent
for it. He would have to spend a lot of time with the image
editor to make it look presentable. Being able to automate it
would be wonderful. Something silly like D-man jumping rope or
something. If anyone could figure out how to do it, he asked
that they send him the prompt.
I suppose one has to approach generative AI from a different
perspective to make sense of it:
It's not a tool to be used by an artist but instead it's a
(soulless) drawing robot commissioned by patron (= the user).
So, here me out. I have two ideas going forward here:
Why not open thread for collecting comic ideas that artists (or
patrons with better prompting voodoo) can turn into art?
If composing comics from templates is doable, we could also
create a collection of D-Man template resources and such to be
used in comics and similar.
(While I wouldn't advocate for it, generative AI should prolly do
a good job rendering background scenery or single sprites for
such comics.)
