Nvidia revealed its new AI-powered DLSS 5 rendering model at GTC 2026, and the internet’s kicking its ass in the comments. Promising the most photorealistic gaming experience to date, the tech certainly talks a good game. While it offers developers “detailed controls” over its various levers, when it comes to character models, all I, and many others, saw from its demos was the same uncanny look you’d find in those AI mobile game ads – you know the ones I’m talking about. If this is why we’re having to sacrifice a kidney for a stick of RAM, then we at least need to find some laughs somewhere. Fortunately, while my chances of an affordable PC upgrade are falling like the Dow Jones, the meme economy is on the rise.
So how does DLSS 5 work? According to Nvidia, “DLSS 5 takes a game’s color and motion vectors for each frame as input, and uses an AI model to infuse the scene with photoreal lighting and materials that are anchored to source 3D content and consistent from frame to frame.” It all sounds very impressive and jargon-y, and theoretically should make some of the best PC games’ myriad elements more harmonious. This is all well and good, but something about the way it impacts characters fundamentally looks… off.
Take Resident Evil Requiem, for example. Both Grace and Leon look like completely different people, and this is observable across other showcased titles, including Hogwarts Legacy and Starfield. Nvidia’s emphasized in its primer that artists have full control over everything from intensity to color grading, and can even “determine where and how enhancements are applied to maintain the game’s unique aesthetic.” Presumably, then, the likes of Capcom and Bethesda have had some level of input on these ‘yassified’ iterations.
Honestly, I’m in two minds. At first glance, I feel like I’m looking at an AI slop filter pasted over the original character. Grace in particular looks like a generic gen-AI creation. Conversely, if this is truly closer to what Capcom’s original artistic intent was for her, then it’s harder to complain. Achieving characters that look this realistic in real-time has long been a developmental challenge, and there’s historically been a strong contrast between them and the environments they exist within, due to generational leaps in that area. The prevailing question, then, must be ‘is this how I want my games to look?’ Right now, I’m not feeling it.
Anyways, onto the fun stuff. The online world has been tearing Nvidia a new one, and there are some phenomenal DLSS on/off memes flying around. Between a beefed-up Todd Howard, the return of the cursed CGI Sonic, and handsome Squidward, players are having a field day cooking up comparisons. Even the deceptively cute Cult of the Lamb got in on the action, putting the ‘slay’ in ‘slay your followers.’
Unfortunately, the gen-AI era looks like it won’t be ending anytime soon. I’m on my knees at this point, praying for the bubble to burst. Though some of what DLSS 5 has to offer is impressive, the cost feels way too high. I’m finding it difficult to believe artistic integrity is being preserved here, but I guess we’ll have to wait and see if iterative improvements to the tech and real-world use yield something a little more soulful.
