Nvidia has responded directly to the growing DLSS 5 backlash, with CEO Jensen Huang pushing back strongly on criticism and insisting the concerns are misplaced. Speaking during a press Q&A at GTC 2026, Huang said gamers and critics questioning the technology are “completely wrong”, while stressing that developers still control how games look, not Nvidia’s AI.

Following the announcement of DLSS 5, some gamers expressed concerns that the technology causes games to appear worse, look homogeneous, or force a specific aesthetic defined by #Nvidia.

At a Q&A Panel at #NvidiaGTC, #JensenHuang fired back. 🔥#gaming #dlss5 pic.twitter.com/Tjr60BCcmQ

— PCMag (@PCMag) March 17, 2026

The response follows widespread debate over DLSS 5’s early demos, particularly around its AI-driven visual changes and whether they compromise artistic intent.

Jensen Huang Pushes Back At DLSS 5 Criticism

During the Q&A, Huang was asked about claims that DLSS 5 creates worse or more homogenised visuals. In response, he dismissed the criticism outright.

“Well, first of all, they’re completely wrong,” Huang said when addressing concerns raised by media and players, according to reporting from Tom’s Hardware.

He explained that DLSS 5 integrates generative AI directly into core rendering elements like geometry and textures, rather than applying superficial effects. This, he argued, allows the technology to enhance visuals without overriding a game’s original design.

Huang also emphasised a key point behind Nvidia’s positioning. “We created the technology, we didn’t create the art,” he said, reinforcing that developers remain responsible for the final look of their games.

Neural Rendering Works At The Geometry Level

A major source of confusion around DLSS 5 has been comparisons to AI filters or post-processing tools.

Huang pushed back on that characterisation, explaining that DLSS 5 operates at a deeper level. “It’s not post-processing at the frame level, it’s generative control at the geometry level,” he said during the session, as reported by multiple outlets covering the event.

According to Nvidia, this approach, described as “neural rendering”, allows the system to understand 3D environments, motion, and materials, then enhance them using AI while staying grounded in the game’s original data.

Huang added that developers can fine-tune how the generative AI behaves, meaning the final result can match a specific artistic style rather than forcing a uniform look.

Backlash Focuses On Visual Changes And “Uncanny” Results

Despite Nvidia’s explanation, the early reception to DLSS 5 has been mixed to negative in some corners of the gaming community.

Criticism has largely centred on how the technology alters visuals, especially character faces. Some examples circulating online, including comparisons involving Resident Evil Requiem characters like Grace Ashcroft and Leon Kennedy, have been cited as looking overly smooth or unnatural, according to Tom’s Hardware.

Others have compared DLSS 5’s output to AI face filters seen on social media, while some critics argue it risks making games look more homogenised or reflective of Nvidia’s own visual preferences, rather than a developer’s intent.

Hands-on impressions have echoed similar concerns, with some noting that the current implementation can feel excessive, producing hyper-realistic results that fall into the “uncanny valley”.

Community Reactions Range From Skepticism To Support

Beyond media coverage, community reactions have been equally divided, with strong opinions emerging across forums and comment sections.

Some players have questioned Huang’s dismissal of criticism, arguing that visual preference is subjective and cannot simply be labelled as “wrong”. One commenter wrote that “no one is completely wrong or even slightly wrong at all… everyone is entitled to their view and opinion” .

Others pushed back on Nvidia’s messaging more directly, with some interpreting the response as dismissive or out of touch. Comments described the stance as “arrogant” or criticised the company for telling players how they should feel about what they are seeing .

There are also concerns around how DLSS 5 was presented. Some users argue that the early demos gave the impression of “generic gen AI” visuals, suggesting Nvidia failed to communicate the technology clearly or showcase its full potential .

At the same time, not all reactions have been negative. Some community members have urged patience, pointing out that the technology is still in an experimental stage and could improve significantly before release .

Others see long-term potential, comparing DLSS 5’s current reception to earlier versions of DLSS or even ray tracing, both of which faced criticism before becoming more widely accepted.

Nvidia Says The Current Version Is Still An Early Preview

Part of Nvidia’s defence rests on the current state of the technology.

The company has confirmed that DLSS 5 shown at GTC is an early preview, with the model still being optimised ahead of its planned release in late 2026, according to its own FAQ cited by PCMag.

That means the visuals drawing criticism now may not reflect the final implementation seen in shipped games.

Huang also pointed to Nvidia’s past experience with ray tracing, noting that it faced heavy scepticism when first introduced with RTX 20-series GPUs in 2018 before becoming widely adopted.

“Everybody said ray tracing was FUBAR,” Huang recalled, arguing that the technology eventually proved its value as it matured.

Developers Retain Full Creative Control

A central theme in Nvidia’s response is that DLSS 5 does not dictate how games should look.

Huang repeatedly stressed that developers have direct control over how the technology is used. He suggested studios can experiment freely, whether aiming for photorealism or more stylised visuals, with the AI acting as a tool rather than a decision-maker.

He also described DLSS 5 as “content-control generative AI”, highlighting that it is conditioned by the original game data rather than generating visuals independently.

DLSS 5 Debate Likely To Continue Until Release

With DLSS 5 still months away from release, the conversation around it is far from settled.

Nvidia is positioning neural rendering as a major step forward in graphics technology, potentially on par with the introduction of ray tracing. However, early reactions show clear concern about how far AI should influence the artistic side of game development.

As more demos and final implementations appear closer to launch, the industry will get a clearer picture of whether DLSS 5 enhances games in a meaningful way or reinforces the concerns currently driving the backlash.

The post Nvidia CEO Jensen Huang Responds ‘They Are Wrong’ To DLSS 5 AI Slop Backlash appeared first on Shane the Gamer.

NVIDIA DLSS 5 was meant to be a breakthrough moment for game graphics, but instead, it’s quickly turning into one of the most divisive tech reveals in recent memory. While NVIDIA is pitching it as a leap toward photorealism, a growing number of developers and players are saying the opposite, arguing the visuals look artificial, uncanny, and in some cases, outright worse than the original games.

NVIDIA Promises A New Era Of Photorealism

Announced at GTC 2026, DLSS 5 introduces what NVIDIA calls a real-time neural rendering model. Instead of just upscaling resolution or generating extra frames like previous versions, this new system actively alters the image by adding AI-generated lighting, materials, and detail.

According to NVIDIA, DLSS 5 analyses a frame’s colour and motion vectors, then reconstructs it with “photoreal” lighting and materials while staying anchored to the original 3D scene.

CEO Jensen Huang described it as “the GPT moment for graphics”, positioning it as a fundamental shift in how games are rendered.

The company also emphasised that developers retain control, with tools for intensity, colour grading, and masking to preserve each game’s artistic style. DLSS 5 is set to launch in late 2026 and is already planned for major titles like Resident Evil Requiem, Starfield, Assassin’s Creed Shadows, and Hogwarts Legacy.

The Problem Players Are Seeing In DLSS 5

Despite the ambitious pitch, early footage has sparked immediate backlash. Instead of enhancing realism, many viewers say DLSS 5 makes characters look like they’ve been run through a generic AI filter.

In side-by-side comparisons shown by NVIDIA, character faces appear noticeably altered.

Features such as skin tone, facial structure, and even expressions shift in ways that weren’t present in the original game assets. Some viewers compared the effect to social media beauty filters, with exaggerated smoothness, sharper features, and a slightly “off” look that sits firmly in the uncanny valley.

One widely shared comparison contrasts Death Stranding 2’s original character model with a DLSS 5-enhanced version, highlighting just how dramatically faces can change.

Death Stranding 2 Nvidia DLSS 5 pic.twitter.com/6r4m3wEMDI

— TheGameVerse (@TheGameVerse) March 17, 2026

Just look at how good Death Stranding 2 already looks. It doesn’t need this. DLSS 5 comes in and straight up butchers the character, replacing it with something that looks like generic AI output.

Others pointed out inconsistencies like unnatural eye movement and overly processed textures, which resemble the wave of low-quality generative AI images often referred to online as “AI slop”.

This looks horrifically bad, nobody wants an AI slop filter on top of their games

— Synth Potato🥔 (@SynthPotato) March 16, 2026

This has led to a core criticism that DLSS 5 isn’t just enhancing visuals, it’s effectively replacing them with an AI interpretation.

Developers Are Publicly Pushing Back

The criticism isn’t just coming from players. Several developers and industry figures have openly voiced concerns about what DLSS 5 represents.

Dave Oshry, CEO of New Blood Interactive, didn’t hold back, calling the tech “AI dogshit” and warning that future generations might grow up thinking this kind of output looks normal. He also urged the industry to “push back harder” against it.

Arman Nouri, a senior environment artist at Epic Games, echoed similar concerns, saying it was depressing to think AI-generated visuals could become the standard for younger players.

Agree with this so much, watching my daughter grow up where AI slop is the norm and her not knowing quality of quantity really bums me out. Makes me appreciate my childhood a lot more.

— Arman Nouri (@armanguy) March 17, 2026

Michael Douse, publishing director at Larian Studios, took a more sarcastic approach, joking that Resident Evil Requiem felt “unplayable” without DLSS 5’s promised photoreal enhancements, clearly poking fun at NVIDIA’s marketing claims.

Playing some Resident Evil Requiem this morning. The game is great but I just wish it infused pixels with photoreal lighting and materials, bridging the gap between rendering and reality alas unplayable for now

— Very AFK (@Cromwelp) March 17, 2026

Even lighter reactions, like memes and jokes from across the industry, highlight the same underlying issue, that something about DLSS 5’s output just doesn’t look right.

Art Direction Versus AI Interpretation

One of the biggest sticking points is how DLSS 5 interacts with a game’s original art direction.

Modern AAA games already spend years refining character models, lighting, and materials to achieve a specific visual identity. Critics argue that layering generative AI on top risks flattening that identity into something more generic.

In examples like Resident Evil Requiem, characters reportedly end up looking like entirely different people once DLSS 5 is applied. That raises questions about whether the technology respects artistic intent, or overrides it.

NVIDIA maintains that developers can fine-tune the effect, but early reactions suggest that even with control options, the underlying AI transformation may still introduce unwanted changes.

DLSS 5 Still Has Industry Support Despite The Backlash

It’s worth noting that not everyone is against DLSS 5. Major studios including Bethesda, CAPCOM, Ubisoft, and others are already onboard, with plans to integrate the technology into upcoming releases.

From a technical perspective, the idea of bridging the gap between real-time rendering and film-quality visuals is something the industry has been chasing for decades. DLSS 5 attempts to solve that problem not through raw hardware power, but through AI reconstruction.

There’s also a practical angle. Real-time rendering budgets are extremely limited compared to film production, where a single frame can take hours to render. DLSS 5 aims to simulate that level of detail instantly, which, in theory, could unlock new possibilities for game visuals.

The Bigger Question Around AI In Game Graphics

The strong reaction to DLSS 5 highlights a broader tension in gaming right now. AI tools are becoming more powerful, but not everyone agrees they should be used to reinterpret finished artwork.

For some players and developers, the appeal of games comes from handcrafted detail and intentional design choices. If AI starts modifying those elements in real time, even slightly, it raises concerns about authenticity and consistency.

At the same time, NVIDIA is unlikely to back away from this direction. DLSS has already become a standard feature across hundreds of games, and DLSS 5 represents the next step in that evolution, whether players like it or not.

The real test will come when DLSS 5 launches later this year and players can experience it firsthand. Right now, based on early footage and reactions, the technology might be aiming for realism, but for many, it’s landing somewhere much stranger.

The post NVIDIA DLSS 5 AI Slop Backlash Grows As Gamers Slam New Visuals appeared first on Shane the Gamer.

Nvidia’s Ampere-based RTX 30XX-series GPUs continue to be very scarce, so scarce that it’s taken me six weeks to even get one loaned to me for review.

Thankfully, Asus came to the table, letting me play with one of their top-of-the-ranger GPUs, the ASUS ROG STRIX GAMING GeForce RTX 3080 OC.

The first thing that strikes you with these cards is just how big they are. The 3080 was a fair size bigger that the RTX 2080 Ti that I had in the case. I can see a lot of disappointed people being faced with buy a new case to house these monsters.

Installing the card required adding an extra 8-Pin PCI cable to the power supply unit. The Asus RTX card requires three 8-Pin connectors.

The card is factory overclocked with a 1935 MHz GPU boost clock out-of-the-box. That’s 225Mhz over the Nvidia Founders Edition reference settings. As with all the RTX 3080, the ROG card has 10GB of GDDR6X memory. Yes, that’s 2GB less than the RTX 2080 Ti that the card replaces, but, in all seriousness, it’s more than enough for today’s games.

The box contains a little pack with a thank you note, trading card, quick set-up guide and what I can only assume is a metallic sticker for your case. There’s also some branded Velcro cable ties and little ruler-come-keyring. All in all, it’s a nice package that really complements this prestige GPU.

The card has two HDMI 2.1 outputs and three DisplayPort 1.4a outputs. The USB-C that was on my old RTX 2080 Ti has been ditched for the RTX 30xx-series.

The RTX 3080’s Ampere technology features the third generation of Nvidia’s Tensor cores and the second generation of their RT cores. The Tensor cores allow the card to utilise Nvidia’s latest deep-learning super-sampling AI algorithms to produce high quality visuals without the usual processing overhead. The RT cores power the card’s real-time raytracing ability, to produce life-like reflections and lighting.

The card is heavy. Very heavy. I’d be carefully lugging your case about this this beast installed, and it puts quite a bit of pressure on the PCIe mount. This weight is down to the robust heatsink and spreaders that form the card’s cooling system. You can see the finned heatsink running almost the whole length of the card. Asus really want this card to be kept cool.

The card has a tiny switch that can be used to switch the card from the OC mode to a more sedate boost speed. Why you’d pay all this money for a card like this is beyond me. But if you want to run the card with lower performance and use less power, the option is there, I suppose.

The underside of the card is covered by a baseplate to protect the electronics.There is a box-out on the rear of the actual GPU showing off a full set of the more expensive, and apparently more robust, smaller “MLCC” capacitors.In the unlikely event of something actually going wrong, the card has a 3-year warranty.

The side of the RTX 3080 has an addressable RGB strip. This can be customised using the ROG Armory Crate software to sync it up with other ROG RGB components. The ROG Strix RTX 3080 OC is a stylish-looking card that looks good in your rig.

The RTX 3080 has really been designed to enable 4K/60fps PC gaming. The Ampere technology likes to be pushed. If you are still gaming with a 1920×1080 monitor, you may want to look at the RTX 3070 instead, or pick up a used RTX 2080. I game on a 2560 x 1440 monitor and I found games like Call of Duty Modern Warfare easily breaching 100fps maxed out with DLSS. Games like Assassin’s Creed Odyssey performed at above 60fps even with 200% super-sampled resolution. Using an Nvidia Shield TV I was able to stream Assassin’s Creed Odyssey to a 4K HDR TV with the settings maxed out and still achieve the TV’s max 60fps!

The extra power that the RTX 3080 gives your PC allows for better ray-tracing results than those teased by Nvidia’s 20xx-series GPUs. The Minecraft RTX beta, for instance, allowed me to increase the RTX effects without impacting performance.These GPUs also give virtual reality applications a powerful boost, EA’s Star Wars Squadrons looked and played great in VR using the Vive Cosmos.

Nvidia’s RTX GPUs have a built-in feature that allows “one-click” overclocking. third party programs like MSi’s Afterburner can be used to analyse a GPU’s individual attributes and create a robust, custom over-clock unique to the card.

Of course, I had to give it a whirl with the ROG RTX 3080 to see what happened. Afterburner spent about half-an-hour checking the GPU before applying an overclock that saw the boost clock exceed 2000Mhz. I added a cheeky +200Mhxz to the memory clock as well.

You can see below, benchmark comparisons between my old RTX 2080 Ti, the ROG RTX 3080 factory settings and the ROG card with the Afterburner one-click OC applied. With the Afterburner overclock applied, the card was noisier, with the fans running a bit higher to counter the additional heat.

Whilst the overclock only gave the card a marginal boost, when every frame counts, this is effectively something for nothing. The substantial gains that the 3080 has over the 2080 Ti is clearly illustrated. Whilst not as impressive with DX11 games like with the Assassin’s Creed Odyssey benchmark, The RTX-enabled DX12 Shadow of the Tomb Raider had some good gains, as did the 3DMark ray-traced Port Royal benchmark.

The Asus ROG Strix Gaming Geforce RTX 3080 OC is a prime example of Nvidia’s Ampere technology. The GPU heralds the next-generation of PC gaming, finally fulfilling the promise of 4K / 60fps gaming.

The card is a well built and robust high-performance bit of kit that should be welcome in anyone’s gaming rig.

The post ASUS ROG STRIX GAMING GeForce RTX 3080 OC Review appeared first on Shane the Gamer.

The post NVIDIA to Showcase GeForce RTX & SHIELD TV at PAX Australia appeared first on Shane the Gamer.