Last week, Nvidia’s public reveal of DLSS 5—and its “generative AI” enhanced glow-ups of gaming scenes—drew widespread condemnation from the gaming community. In a podcast published Monday, though, Nvidia CEO Jensen Huang tried to differentiate the technology’s optional, artist-guided graphical enhancements from the “AI slop” that Huang says he’s not a fan of.
As part of a nearly two-hour-long interview with the Lex Fridman Podcast, Huang was asked to explain the “drama” around DLSS 5 and “the gamers online [that] were concerned that it makes games look like AI slop.” Huang responded that he “could see where they’re coming from, because I don’t love AI slop myself… all of the AI-generated content increasingly looks similar and they’re all beautiful, so… I’m empathetic towards what they’re thinking.”
At the same time, Huang said DLSS 5 is decidedly separate that kind of “slop,” because it “is 3D conditioned, 3D guided.” The artists behind a game are still the ones creating the in-game structural geometry and textures that form the “ground truth structure” that DLSS 5 works from, Huang said. “And so every single frame, it enhances but it doesn’t change anything,” he said.
For the most part, though, gamers haven’t been worried about DLSS 5 creating trippy new content from the ground up like some generative AI world models. Instead, the worry is that DLSS 5’s visual “enhancements” could end up smoothing out many disparate games toward a single, flattened, homogenized photo-realism standard.
That’s a misunderstanding of how DLSS 5 works, Huang said. It’s not a technology where a game ships in one state and “then we’re gonna post-process it,” he said. Instead, DLSS 5 “is integrated with the artist, and so it’s about giving the artist the tool of AI, the tool of generative AI.”