As a lone fan of clipping 3d models I’ll certainly miss some of the past jank, lol. But if AI makes it easier to develop the game, it’s understandable why you’d use them. And as far as I can see the models are still there, only enhanced by the AI. I haven’t played yet, but my only worry is how this’ll affect animations, as they can be very hit and miss with AI editing. I’ll edit this reply once I get there.
Edit: So it’s a mix between AI images and non-AI animations, I feel like that’s optimal. Though using just the models was great for keeping the sizes consistent, I guess some kind of trade-off had to be done.
Edit 2: Some thoughts on what works better and what works worse: (I elaborated the worse parts more which is why the section is longer, don’t get scared by it please lol)
Better: torn clothes, thin faces, boobs, asses, bellies at certain angles, skin dentation from tight clothes, individual scenes that aren’t a direct progression of the other, sweat (seriously, the AI is great with that one)
Hit or miss: fat faces (Sometimes they’re amazing, sometimes they don’t exist. Tough to get right)
Worse: arms (the AI often struggles with them, which causes artifacts and muscle-esque deformations), stretch marks (they were there when Akane first got them, but later I feel like they disappeared, again something AI struggles with), bellies at odd angles (same problem as arms), eating (it reminds me of the infamous Will Smith eating spaghetti videos at times), scenes that happen one after another (for example Akane waking up, sometimes her entire body seems to shift and her hair changes colour)
Not a comprehensive list by any means, but I feel like it’s important enough that it could shape the way in which future scenes look, in order to maximize the potential of your chosen method.
