Why many dislike NVIDIA DLSS 5 (but will eventually appreciate it)

      Upscaling, or real-time frame reconstruction for video games, is a highly debated practice. Purists are opposed to it, while users with less powerful gaming systems appreciate the increased smoothness it provides. NVIDIA offers this technology, as do AMD and Intel. However, chaos ensued when Nvidia revealed its latest version of super-sampling technology, primarily due to the overly AI-enhanced appearance of the visuals, particularly in human faces.

      DLSS 5 Off (left) vs DLSS 5 On (right) NVIDIA

      The past few weeks in the tech industry have been tumultuous, especially for anyone following the DLSS 5 (Deep Learning Super Sampling) saga, which has included moments of excitement, confusion, and disdain. Here’s a summary of the DLSS 5 drama, from the initial excitement to the current perception of it as a “2D filter.”

      The Background: The “GPT Moment” That Wasn’t

      It all began when Jensen Huang took the stage at NVIDIA’s GTC 2026 and announced DLSS 5. NVIDIA was not merely upscaling pixels anymore; they were generating new visual interpretations. Jensen referred to it as the “GPT moment for graphics,” claiming AI would take on the task of achieving visual realism, handling aspects like skin textures, fabric reflections, and dynamic lighting. Unfortunately, the excitement fizzled out in less than a day.

      Awful actually😑 The more you look at it the worse it gets. This is basically Nvidia DLSS 5 applying its AI photorealism to existing games, removing artistry from the experience. For instance, Resident Evil Requiem from Digital Foundry pic.twitter.com/qgoK4itztF— SpawnYaard 🎮 (@SpawnYaardReply) March 16, 2026

      Within hours, the internet was inundated with comparisons between Resident Evil Requiem and Starfield. The community's verdict? “AI Slop.” Instead of enhancing visuals, DLSS 5 was “Yassifying” characters by softening realistic skin textures, inadvertently applying makeup, and making characters resemble 2022 Instagram influencers.

      So … it turns out the studios that “created” those DLSS 5 demos were just as in the dark about it as we were. It was clearly NVIDIA’s handiwork, hence the uniform appearance. pic.twitter.com/Ubg3iECF6p— Beyond FPS (@beyond_fps) March 18, 2026

      Then came the “Betrayal.” As per Insider Gaming, significant game developers were caught off guard. Artists at Ubisoft and Capcom reportedly learned about the DLSS 5 demonstrations simultaneously with the rest of us. NVIDIA rushed to manage the fallout, vowing to provide a “Full Creative Control” SDK with adjustable intensity sliders. However, the final blow was delivered just days ago: An email interview between YouTuber Daniel Owen and NVIDIA’s Jacob Freeman revealed that DLSS 5 isn’t leveraging the game's deep 3D geometry. Instead, it functions as a sophisticated 2D post-processing filter applied over the visuals. The so-called “Neural Revolution” ultimately turned out to be merely an expensive superficial enhancement.

      Why “Better” Isn't Necessarily Better

      In theory, DLSS 5 seems magical. In some instances, it is. When observing a landscape or a still environment, the AI-generated shadows and highlights do appear “cleaner.” However, therein lies the issue: cleaner doesn’t always mean better.

      Video games are a form of art, and art is driven by intention.

      If a developer dedicates years to perfecting a dimly lit, atmospheric corridor in a horror game, they do not want AI to come in and “improve” it.

      NVIDIA

      DLSS 5 tends to brighten shadowy areas and eliminate atmospheric fog, mistakenly identifying these elements as “flaws” that need correction. The surprising reaction from developers about the demo is a significant red flag. It reflects a classic corporate structure: the executives approve a deal with NVIDIA for marketing hype, while the creative teams remain uninformed. Had NVIDIA genuinely collaborated with the artists, it could have utilized 3D data models and designs to inform the AI.

      NVIDIA

      Imagine if the AI was aware of a character’s specific scar location or how a particular fabric should reflect light. In fact, as Veedrac showcased on Reddit, games using DLSS 5 with tone-mapping can look impressive. This indicates that the technology can be effective, but only when guided by human input. By launching it as a “black box” filter, NVIDIA essentially overlooked the very individuals who contribute to creating engaging games.

      Moreover, there's a pressing concern: Data Sovereignty. As a creative designer, why would I want to share my original character designs and lighting maps with an AI model? We’ve seen this play out before. The AI utilizes that data to “learn,” ultimately creating based on your hard