NVIDIA’s DLSS 5 Reveal Was a Mistake

Four days after the reveal of NVIDIA DLSS 5 at GTC 2026, we’re still gathering reactions from the game development and modding community to the highly polarizing technology.

Yesterday, we published some thoughts from RTGI author Pascal Gilcher. Today, we can share an exclusive comment from industry veteran Denis Dyack, as part of a much larger interview that will soon be published in its entirety on Wccftech. Dyack, known to gamers primarily as the creator of classics like Blood Omen: Legacy of Kain and Eternal Darkness: Sanity’s Requiem, was quite critical of DLSS 5 as a whole while also noting that it could spell doom for triple-A developers (despite the fact that several were already on board with it by the time of the announcement), potentially reducing the usual visual gap with indie games.

The recent reveal of NVIDIA DLSS 5 was a mistake and needs to go back to the drawing board. The current release seems to go beyond enhancing the look of a video game by fundamentally changing the game’s art direction. Never mind the artifacting and other AI art issues. The AAA industry is already in trouble, as it has become very difficult to justify production costs. Making things look spectacular is AAA games’ greatest advantage over smaller budget games. If DLSS 5 is widely adopted, it will accelerate the AAA process’s extinction, as it takes away the awe of what high-production art can bring to the table.

Throughout the interview with Dyack, we had already discussed the wider AI subject at length. Dyack doesn’t hate AI, but he does think it is “way overhyped” right now and should be used to do what humans cannot, rather than replace human work.

I do think AI is a pretty big bubble right now and there’s a lot of unjustified fear that AI is going to take over or remove creativity. My background is in AI; I put a running neural network in a game back in 1992 for my master’s thesis in Computer Science. AI is a tool. It’s not a replacement for people. Corporations saying “we’re laying off X amount of people because of AI” are really going to regret that decision. You cannot finish a game with AI, or if you do, it’s awful. Putting someone who comes from AI to run a games division isn’t a first; people who run game divisions have often never done games before. But if there’s going to be a large application of AI within games, I don’t think it’s going to be that fruitful.

AI is way overhyped right now. Its ability to get results on its own is very low. As soon as someone says AI is going to save you money, you can almost universally assume that’s wrong. Technology makes you more productive, but it takes more time. If you use a lot of AI, you need a lot more people and a lot more time. The idea of laying people off because of AI is antithetical to reality.

Peter Moore recently said in an interview that he thinks all studios will eventually use AI in some form. Do you agree with that?

All studios are already using AI. It’s impossible not to. Steam’s AI disclaimer is performative theater. It’s in our compilers, our editors, our engines. If you use DaVinci Resolve for cutting videos, it’s full of AI. That ship sailed about 15 years ago. The real question is: is generative AI going to replace people? I don’t agree with that. It’s probably going to take more people to run.

[…] Right now, AI is being focused on doing things that humans already do well. I think the real win in AI is going to be focusing on things that humans can’t do, like going through large amounts of data very quickly. I think we want to focus on things AI is uniquely suited for rather than trying to replicate human work.

As mentioned earlier in the article, we’ll publish the full conversation with Dyack in the coming days, covering his next game, Deadhaus Sonata, and other prominent industry issues. We’re also still looking to bring you more opinions on the NVIDIA DLSS 5 technology. Stay tuned!

Follow Wccftech on Google to get more of our news coverage in your feeds.