Lol, nice way to wake up seeing this
Nice asbestos tent, but she still got Gollum fingers doe. 😏
Even with foreshortening, those are some freakishly short legs.
The eyes look demonic
This reminds me of camping in my youth with my alcoholic uncle.
You better not be gearing up to go all Deliverance on us with that story
Deliverance didn’t have drunk uncles.
I mean, that’s just a minor detail. And Deliverance taught me how to prepare for south of Market clubs in my 20s. IYKYK.
Your uncle is hot
I think more that the fire is hot cuz he made it in a tent
The tells
Hands issues, thumb and metal claw
Face issue, studio lighting on face, right eye is not eyeing.
Body issues, left leg is a weird tiny leg
Boot laces lack credibility
Fire in a tent
AI is bad at the details.
Prespective is not done properly. Trying to spoof 3D with a 2D model.
I’ve gotten so used to looking at the details that I didn’t even notice the blatant fire-in-the-tent.
I was staring at the hands and thinking it wasn’t too bad.
They’re terrible!
I actually kind of love that AI art can’t even pretend to do hands right. But also I had an instructor who’s catch phrase was ‘‘if you let the computer do it, it’s gonna suck’’ so. Not really a surprise AI can only defeat art that’s placid and uninspired shit. and can’t even do anything on model or sequential… it’s just… it can be a placeholder and that’s about it.
It’s funny how the generators arent getting better. They’ve plateaued pretty hard in terms of believability. Glance value? Convincing.
Under any level of scrutiny though this falls apart in at least a dozen ways.
“AI” image generation actually takes a lot of human knowledge and understanding of the models to manipulate outcomes. It is a different kind of effort, but the issues with it are based in human use of the tool.
Prompt engineering isn’t a real field bro
I think the scale of image believability is logarithmic. Going from “believable at a glance” to “believable under scrutiny” requires an exponential increase in performance compared with going from “not believable at a glance” to “believable at a glance”. The same principle applies to text generation, facial recognition, sound generation, image enhancement, etc. One of the many reasons AI should not be being integrated in many of the ways world governments and corporations are trying to integrate it.
It’s unclear if the current models can reach that level though. They seem more like they’re asymtotically approaching their limit.
Maybe I’m wrong and GPT 5 will be the end all and be all - but I don’t think generalist models will ever be consistent enough. Specialist generators focused on specifics, trained to output very defined data seem more likely to be useful than this attempt to make a single catch all LLM.
It’s more the free and public versions have plateaued
Gpt 4 premium pictures still have these major issues.
Too perfect but wrong hands
The fire also looks a bit too intents to me 😉
Wolverine somehow ended up in the model data and there you have it.
That’s a cozy fire for sure.
Apparently Ai can’t do average or ugly people.
So… I’m safe from deep fake?
tricky but doable
prompt: lemmy.world average user,
this one could be my twin.
2 fat guy pictures, but with quite the averagely-tended faces and a somewhat digital touch-up. 1 outright regular, maybe overweight person with semitic beard. 1 absolutely fabulous gentleman with furry undertones.
Not even closely ugly. Just some quantifiable stuff imo.
This one kind of reminds of Larry Ellison but looks a bit more human than the real Larry
Hey, it’s me! Spot on. But should be arch.
Are those computers running windows? That’s a very inaccurate AI.
Both look very much like .NET development with C# in Visual Studio. Each of these way too much Microsoft for a Lemmy user to touch.
Person #2 is way too masculine and attractive. The people in picture #3 are way too close to touching grass and socializing. The furry (#4) is indeed the most believable.
“Person #2” would have a hell of a time buying premade shirts.
Gloves, too.
Is nobody seeing the 8 inch difference in arm length?
Maybe is just a theme 🤷♂️
What a cozy fire.
It’s funny how similar AI generated images are to accompanying drawings for “what are ten things that are wrong with this picture?” style questions in an IQ test.
They never get the hands right lol
This AI was trained on Dr Seuss’s Wacky Wednesday
