So, when I got hated by redditors for calling DLSS 5 a cheap AI filter, I was actually right? The guy from NVIDIA literally confirmed that the tech only works with a 2D "screenshot" from the game disregarding anything that is beyond current frame. It doesn't take into consideration PBR properties or actual "depth" of the scene – that's why btw the DLSS 5 version of the image looks so flat.
So, to all DLSS 5 supporters, what would you say on that?