So NVIDIA's DLSS (Deep Learning Super Sampling) has reached the point where it's just... generating entire human faces now. Left side looks like your character model rendered at native 480p with a potato GPU, right side is what happens when you let the AI "enhance" it. The technology has gone from "upscale some pixels" to "I'll just draw you a whole new person, thanks."
Fun fact: DLSS started as a way to render games at lower resolution and upscale them intelligently. Now it's basically doing the rendering FOR you. At this rate, DLSS 10 will just be an AI playing the game while you watch. Why even have a game engine when the graphics card can hallucinate the entire experience?
The irony is gamers will still argue about whether this counts as "real" graphics while enjoying their 240fps on a 4060.
AI
AWS
Agile
Algorithms
Android
Apple
Bash
C++