Nvidia CEO Jensen Huang has responded to criticism from the wider gaming community about his company’s new AI-powered lighting tool for gaming: DLSS 5. Gamers really, really didn’t like the “slop-filter”-like vibe the preview put out, but Huang claims they’re “completely wrong” and need to listen to him more carefully as he explains it. Again.
“Well, first of all, they’re completely wrong,” Huang said in a chat with Tom’s Hardware when discussing the criticism. He suggested that DLSS 5 is not a slop filter or a filter at all, as it uses the game’s 3D elements to re-create the lighting, not just overlay something.
“The reason for that is because, as I have explained very carefully, DLSS 5 fuses controllability of the of geometry and textures and everything about the game with generative AI,” Huang said. “It’s not post-processing, it’s not post-processing at the frame level, it’s generative control at the geometry level.”
He also highlighted that developers can and will be able to control the effect by fine-tuning the generative AI. It “doesn’t change the artistic control,” he claimed, adding that game developers would be able to make it do everything from toonify the game visuals to see what it’d be like if everything in the game were “made of glass.”
Neither of those extremes sounds palatable to gamers, even if it might be fun for devs to play around with. He calls the process “neural rendering,” rather than generative AI, despite that being the portion of the pipeline that developers will be able to adjust.
The CEO doth protest too much on this one, but it may not matter either way. The response to DLSS 5 has been so universally panned that the release later this year is likely to be severely dampened. Throw in the fact that this early demo required an entire RTX 5090 running in parallel to another RTX 5090 just to run DLSS 5, and it’s not a feature anyone will be able to use anytime soon, even if they wanted to.
Which, for the record, they almost universally don’t.