Hey! I may be grumpy, and bitter, and lash out at children, and trick people into neglecting God's Word but... What was that other thing you said?
Hey! I may be grumpy, and bitter, and lash out at children, and trick people into neglecting God's Word but... What was that other thing you said?
You're viewing a single thread.
This is AI generated, right? Looks off.
nah i saw this last decade
Doesn't look like that to me, it just looks like someone quickly drew it up in that general style.
The thumb looks cartoonish, the fold under "...but then why are they" is inconsistent and has the bottom of letters too far under it, the words at the bottom of the page look especially borked (look at "are needed"). Pretty certain this is a generated image.
Google reverse image search has it showing up on ytmnd in 2006. I don't think it's AI generated.
Ah nice find, thanks! I think this was used as a reference to generate an image of someone holding an actual printed copy, then
I suppose it sort of looks like that, but it also looks like it could have been an image that was smoothed/upscaled by an "AI" algorithm, but not generated by it directly.
I've upscaled a lot of images before, and the slight font blurring, crinkle lines becoming smoother, translucency of the fingernail looking off, etc all looks to me more like upscaling or smoothing than it does a complete generation.
I'm not ruling it out completely, but generally, this length and quality of text isn't something an AI model can consistently do. It will deteriorate very quickly if you try and have a GenAI model output this much text.
I think it's worse than what you've described -- the closer I look, the worse it gets. The font isn't slightly blurry, it's garbled and the letters are inconsistent. The crinkle lines aren't just smoother, they don't look like paper crinkle (loose point on my part, I'll concede -- hard to describe). The fingernail isn't oddly translucent, the entire thumb looks like a cartoon thumb. Someone else found the original image on ytmnd, I think using that as an input is how the length and consistency of the text was barely maintained. The bottom right part of the background behind the goat demon isn't right either, unlike the original.
I think that given that context, the idea they used it as a base image and generated the rest on top of it does actually seem quite likely. Especially if they'd used a tool like ControlNet, that would explain how the text was still as consistent as it is, albeit not perfect.