I graduated HS in '76. That kid looks a lot more like my contemporary or a couple years earlier rather than somebody from the early, or worse yet, late 80s. Heck, I had so many shirts like that. Though no braces or glasses.
OTOH, I was in SoCal and back then fashions sometimes took years to reach the Midwest. So depending on where this kid is, 1985 might be about right for that look.
Two thousand four hundred and 50-some possums? Are you OK?
I’m not trying to kink shame, but is there maybe some support group perhaps you could join? It’s OK, we’re all friends here. We can talk about this. 
I’ve had many, many visits from them. I have a feeding station near the house where they get second dibs on scraps that the cats don’t want.
They get used to you, and ignore you as irrelevant. Literally walk over you to get to the food.
They’ve been hanging with my family for a long time. Note the cars in this photo with my grandfather.
Possums were around where I grew up too, but other than as roadkill you only very occasionally saw one and always as the north end of a southbound possum in a hurry to scuttle into some bushes for cover.
That’s really cool that you have them so well habituated that they’re happy right up close and personal. And yeah, were I in your situation, 2400 some possum pix sounds about right for me too.
Young ones are skittish. Older ones are “eh, whatever”. I sometimes pet them while they are eating. The gray fur that you see is pretty wirey. But they have a softer white undercoat. And the tails are scaly, dry, and reptilian. (And they store fat in a big hump in the base of the tail.) They groom with their back feet. And sometimes blink one eye at a time.
The first image resembles that Ukrainian-American actress who does Meg’s voice in Family Guy.
But back to AI related, I have lots of AI possum photos because the older Dall-E 3 is great with possum photos. Including ones of Bread Climp:
(Later I dig up some of my favorites.)
I wondered if any of that AI-generated stuff had been fed back into ChatGPT-4o’s training, and so far it appears the answer to that is no:
Further quizzing in text revealed it had no idea about Bread Climp until I let it do a web search.
Loved the Rat Fink Hot-rod character as a kid. I turned this one into a somewhat photo-realistic rendering:
Okay, here’s images from a series on possums through history.
And here’s a small sampling of various other possum images that probably made sense in the proper context.
Now that I’m playing with the shiny new toy it is easy to forget how competent Bing/Dall-E 3 can be with less ambitious prompts.
That kid and shirt look spot on for my early 80s school pictures from 4th & 5th grade. No neon background but the kid feels correct to me.
Marketing photo of a complete set of gashapon toys based on chibi cast of Seinfeld Caganer figurines
Marketing photo of a complete set of gashapon toys based on chibi cast of Seinfeld Caganer figurines, front and back views.
George is looking either miserable, concerned or angry in most of those. Think he needs a good dose of Dulcolax or something… 
Saw this on Sora a few minutes ago.
Tried a 4-split image using Ponderoid’s trick, and of course the results weren’t as interesting.
I tried the disgusting food combination in copilot. Interesting how it gave me basically the exact same kid.
ETA: here’s the prompt I used. “Create an image of a child horrified by the combination of foods on the plate in front of him at dinner. Photorealistic.”
Huh. I don’t think I’ve made anything through Copilot before. That very much looks like a ChatGPT 4o image. But Bing.com definitely still uses Dall-E 3.
I signed into Copilot, plugged in your exact prompt and got this
I appreciate the creativity behind your request! Unfortunately, I can’t generate that particular image. However, if you’re looking for something fun or artistic, I can still help—maybe an exaggerated cartoon version or a whimsical painting? Let me know how you’d like to proceed!
So inconsistent, nearly random rules. Just like ChatGPT!
My second try at Copilot, it refused to make “Realistic photo of a reverse mermaid and a reverse centaur playing reversi. 16:9”. (ChatGPT made something yesterday, but had no clue what I meant.)
Third try, and finally it makes something.
Playing around with Copilot, the guard rails are even higher than on ChatGPT. Several of the prompts that worked there failed in Copilot.
It can do randomized selection. I ran my 3-variable prompt and it did a 3,3,4, plague doctor juggling potatoes in a crowded bazaar.
A four way split gave me a 1,1,1, a 2,2,(2-3 combo), a 4,4,3, and a 3,3,4. (Interestingly not a 1,1,1, a 2,2,2, a 3,3,3 and a 4,4,4 like Ponderoid’s example.)
I think this is the first one of my Sora images I’ve randomly seen reposted somewhere else.
(It doesn’t beat my “Jonsey the cat carrying a baby xenomorph” from Bing, though. It ended up with many thousands of shares and hundreds of thousands of reactions. I last randomly saw it a few days ago.)
