Recently, Nvidia’s public expose of DLSS 5– and its “generative AI” improved glow-ups of video gaming scenes– drew prevalent condemnation from the video gaming neighborhood. In a podcast released Monday, however, Nvidia CEO Jensen Huang attempted to separate the innovation’s optional, artist-guided visual improvements from the “AI slop” that Huang states he’s not a fan of.

As part of an almost two-hour-long interview with the Lex Fridman Podcast, Huang was asked to discuss the “drama” around DLSS 5 and “the players online [that] were worried that it makes video games appear like AI slop.” Huang reacted that he “might see where they’re originating from, since I do not like AI slop myself … all of the AI-generated material significantly looks comparable and they’re all lovely, so … I’m understanding towards what they’re believing.”

At the very same time, Huang stated DLSS 5 is distinctly different that type of “slop,” due to the fact that it “is 3D conditioned, 3D assisted.” The artists behind a video game are still the ones developing the in-game structural geometry and textures that form the “ground fact structure” that DLSS 5 works from, Huang stated. “And so every frame, it improves however it does not alter anything,” he stated.

For the a lot of part, however, players have not been stressed over DLSS 5 developing trippy brand-new material from the ground up like some generative AI world designs. Rather, the concern is that DLSS 5’s visual “improvements” might wind up raveling lots of diverse video games towards a single, flattened, homogenized photo-realism requirement.

That’s a misconception of how DLSS 5 works, Huang stated. It’s not an innovation where a video game ships in one state and “then we’re gon na post-process it,” he stated. Rather, DLSS 5 “is incorporated with the artist, therefore it’s about providing the artist the tool of AI, the tool of generative AI.”