Nvidia launches DLSS 5 AI upscaler, mixed feedback
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Nvidia announced DLSS 5 at its GTC conference, calling it a ‘GPT moment for graphics’ that blends hand‑crafted rendering with generative AI.
- DLSS 5 uses generative AI to rework lighting, shadows, and material details in real time, delivering noticeable visual boosts at up to 4K resolution.
- Resident Evil Requiem demo shows DLSS 5 altering the protagonist’s look, adding fuller lips and intense eyeshadow, which some describe as ‘AI slop’.
- Starfield demo shows DLSS 5 increasing sharpness and bright highlights that make characters appear stage‑lit despite the in‑game lighting.
- Todd Howard of Bethesda Game Studios praised DLSS 5, saying it ‘brought it to life’ and was ‘amazing’.
- Mike Bithell criticized DLSS 5, calling it ‘nonsense’ for anyone who doesn’t want any art direction in their gaming experience.
- Nvidia says developers can control DLSS 5 intensity, color grading, and mask specific objects to preserve artistic intent.
Why it matters: It gives developers new tools to boost realism, but also risks altering the visual style that creators intended, as shown by the divergent reactions from Todd Howard and Mike Bithell. The ability to apply generative AI at 4K in real time satisfies gamers seeking photorealism, while the ‘AI slop’ concerns deter studios that prioritize artistic direction.
