I don’t work with classic prompt engineering .. so I won’t provide a prompt.
I work with dynamics myself 
Indeed, I also prefer to let people think for themselves and just give them inspirations. But I will give you a concept, because you have already developed something that goes exactly in this direction 
Oh Nice! I do have a different way to go with stuff. I don’t stick to one aspect too long. Depends on what I do.
Time Waits For No One - Stones fan, eh?
Someone’s gonna say something so I’ll politely refer back to one of the rules in the OP post.
It’s a shared space and we try to avoid pages of one person’s post
It’s supposed to be a quality over quantity type of thing, and the carousels are the vehicle to side step that.
One thing that I’m sure is on everyone’s mind is that this one thread alone kinda tilts the cheer meter of the website and needs to be respected.
If I don’t at least make the attempt to point this out to you it could look potentially bad for me, having invited you.
Take turns, rather than entire pages of a thread.
You can always edit your last post to add to a carousal too as that doesn’t push the cheer meter up from a bunch of posts that could have been one.
Carousel is ok? I forgot to change that and add the tag
yeah so, like 1-4 image rule we had to put in
could litterally be 1-4 four carousals…
You can add quite a bit to those.
the most important thing is be gracious to our host,
move with their speed
not be just like them,
but more like…
learning to relate with them rather than ahead of them or at them…
my best advice!
I have a few purposes with so many images:
This of course since 2022, it’s not that I suddenly made those images in 2 weeks.
You wonder what? Try the given name “Dubhchobhlaigh”, 美智子, Barack, … or “Chastity”.
GPT-Image-2 specifically inserts a lot of text output. So a bias towards written-out names is seen.
A person’s name arbitrarily is no guarantee that the output is a picture of a person. I’ve picked from trials where we have a picture of a human along with their text, instead of greeting cards or infographics or maps of Africa.
“Aibinn” gave me one “my girlfriend” type, and then revisiting, could not recover from being all East Asians, and I persisted until text.
Challenges to making people:
I tried right from the start, a clause in the prompt - in this case (which is inherently not going to give what you are searching for in underlying name meanings): don’t copy the same Japanese person in every picture; give the “ko” portrait assignments uniqueness by underlying demographic usage by portrayable ethnotypes, etc.
Funny: Sakurako gets cherry blossoms, Yukiko gets snow, Ainuko: literally Ainu? etc. The model can’t help itself, like the “Bambi” or “Cherry” I showed before.
Curious, what would you get with “Larisa”?
OK, fair, I’ll give you the VIP treatment, since you are part of the thread. 
I made it with and both accounts:
Larisa
Larissa
Laryssa
I recreated the image for non-Japanese speakers adding rōmaji. So they can pronounce such pretty names too.
But @_j made one better go check it out!
