The internet has spent the last couple of days completely blasting DLSS 5, Nvidia’s fancy new graphics technology that inadvertently yassifies video games. The AI company wanted DLSS 5 to be seen as the next frontier in video game visuals. Instead, it’s become a meme that keeps getting compared to “AI slop.” A negative public reaction like this would cause a PR campaign overhaul at most companies. But rather than admitting DLSS 5 has missed the mark, Nvidia is doubling down on its current message.
DLSS 5 is a feature of Nvidia’s upcoming RTX-50 line of graphics cards. The technology uses machine learning and “neural rendering” to boost the lighting of games that support it. At least, that’s the idea. Footage displaying the before and after for DLSS 5 has whipped up a storm of controversy on the internet, in large part because the technology seems to alter character designs. When it’s active on inanimate objects or elements like metal and water, DLSS 5 shines. On humans, though, it’s completely uncanny. It’s like turning on a beauty filter on TikTok.
Nvidia maintains that DLSS 5 does not actually alter the games themselves. Things like textures and renders stay the same whether the graphics tech is on or off, the company claims. Nvidia has also pointed out that game developers are opting in to use its technology, and that in many cases, these studios insist that DLSS 5 is actually helping them get closer to the developer’s original vision. It seems as if Nvidia considers the reaction to DLSS as a matter of misunderstanding facts. On social media, for example, Nvidia has clarified that DLSS is not actually a filter.
A couple of days after the DLSS reveal, Nvidia CEO Jensen Huang officially responded to the criticisms. Per an interview with Tom’s Hardware, Huang came out of the gate saying, “Well, first of all, they’re completely wrong.” Huang then goes on to repeat Nvidia’s talking points about creative control and the nature of the technology itself.
“It’s not post-processing, it’s not post-processing at the frame level, it’s generative control at the geometry level […] This is very different than generative AI; it’s content-control generative AI,” Huang says. “That’s why we call it neural rendering.”
The thing that suggests Nvidia misunderstands the situation is that Huang kicks off his response with the phrase, “as I have explained very carefully.” And sure, some people might not understand exactly what the DLSS tech is doing on a mechanical level — but arguing about the definition of a filter seems like a technicality. Ultimately, regardless of how Nvidia achieves its output, gamers are taking umbrage with the results. When gamers are used to seeing iconic video game characters like Resident Evil’s Leon Kennedy in specific ways, they’re going to have opinions about how he’s depicted. If it seems like something has changed, you can’t really debate someone to make them feel less weird about it.
And on some level, it seems like the public has grown weary of the pervasive way AI is used to depict “realistic” visuals. The aesthetic is stigmatized now. For some, the type of hyperrealism that AI-fueled tech defaults to betrays a lack of imagination. The public isn’t wrong about DLSS 5 — they just don’t like it.
