Grildon Tundy
Member
The comment in that tweet is needlessly antagonistic, but the gif seems self-evident that it's the same model with lighting and color/"makeup" changes.
-
Hey Guest. Check out your NeoGAF Wrapped 2025 results here!
The comment in that tweet is needlessly antagonistic, but the gif seems self-evident that it's the same model with lighting and color/"makeup" changes.
Madflavor
Member
I think the bottom line is that the average person has seen one too many AI Faces, and therefore they can't unsee the AI Face in the DLSS 5 versions of these character models. It's uncanny and it's off putting to a lot of people. You're not going to gaslight them into thinking it's better.
Furthermore this is ultimately one more big step towards an industry where major publishers will be firing thousands of hard working and talented artists and developers, in favor of cost saving with Generative AI. You're fooling yourself if you think publishers aren't salivating at the idea. Some of the tech might be impressive, but it's implications for the industry paint a dark future.
Some of the greatest stories and games ever made came from the passion, discipline, and imagination of talented creators working within constraints. They didn't have limitless tools or automation to fall back on, so they had to push what was available as far as it could go to realize their vision. That pressure is what led to distinctive styles and memorable design choices. Limitation breeds creativity. When technology starts filling in the gaps with minimal human effort, that creative edge risks getting dulled.
Furthermore this is ultimately one more big step towards an industry where major publishers will be firing thousands of hard working and talented artists and developers, in favor of cost saving with Generative AI. You're fooling yourself if you think publishers aren't salivating at the idea. Some of the tech might be impressive, but it's implications for the industry paint a dark future.
Some of the greatest stories and games ever made came from the passion, discipline, and imagination of talented creators working within constraints. They didn't have limitless tools or automation to fall back on, so they had to push what was available as far as it could go to realize their vision. That pressure is what led to distinctive styles and memorable design choices. Limitation breeds creativity. When technology starts filling in the gaps with minimal human effort, that creative edge risks getting dulled.
FoxMcChief
Gold Member
I guess I just don't get the big fuss. I think it looks interesting.
Is it just that "anti-Ai" hate boner some have?
Is it just that "anti-Ai" hate boner some have?
SegaSnatcher
Member
No, many of us just honestly believe it looks like hot garbage.
Most people here love DLSS upscaling, but this is something entirely different.
DLSS 5.0 doesn't have any concept of a model though. It only has access to the rendered frame + depth buffers + motion vectors. If DLSS 5 was a system that developers fed their lighting data into, and it produced corresponding light values for the scene, I think there would be zero issues.
