Before you decide to ragepost in the comments about how terrible either of these guys are, please take a moment and look at how happy they are. Wouldn’t this be a nice outcome for a change?
Before you decide to ragepost in the comments about how terrible either of these guys are, please take a moment and look at how happy they are. Wouldn’t this be a nice outcome for a change?
I really doubt this is AI, the faces maybe, but everything else screams photoshop job on real images.
I play around with AI every day, and there is no way such detailed scenes have zero fucked up cooking tools. It’s jpg as fuck, but I cannot find one AI artefact anywhere in any of those images.
No artifacts? What are you talking about? All of those scenes are utter nonsense. Upper left, nothing to note, it’s super generic. But the upper right is a fall themed picnic… in the spring because of the Sakura blossoms. And nothing in the “picnic” makes any sense.
Bottom right, does that look like a believable scene to you? Cause, there’s way too much fruit, not enough workspace, the pans on the wall in the back make no sense, and the pie is something out of a nightmare. And we come at last to the worst of the bunch. I call it… something to do with string. Cause what they’re making is indecipherable. Voodoo dolls? Creepy, tiny yarn balls with red tubes? And the background is “vaguely Christmas”, complete with ornament on a branch, connected to nothing
And yeah, I do see artifacting. It’s most obvious in the kitchen scene. The “grapes” are mostly collections of blocks. Blocks that intrude into the line dividing the top from the bottom. The “box” in the picnic scene was another example.
Also, the focus is faked. Things in the foreground are more in focus than what should be nearby objects. Clear fields of view wander from focus, to out-of-focus, to back in focus.
The compression artifacts and low resolution do a good amount to hide the problems, but there’s nothing believable about these images. They’re clearly ai generated, with all the defects that implies.
I agree, GenAI has gotten pretty good as of late, but 0 artifacts at all is still too far for current tech and I couldn’t find any either
Could be generated and then touched up by a human Editor. I suspect that’s going to be the optimal workflow pretty soon
No. This is clearly AI generated. See my other comment in this comment tree.
It could be both. Plenty of AI generated images that get touched up by humans still show artifacts, because the human touching them up doesn’t notice every single one.
No, this is very clearly entirely ai generated. Please see my other comment in this comment tree.
deleted by creator