Nvidia CEO Defends AI 'Yassification' Feature Amid Backlash

Nvidia CEO Jensen Huang has doubled down on his company's controversial DLSS 5 feature, dismissing gamer criticism as "completely wrong" despite widespread backlash over its AI-altered aesthetics. The feature, which uses generative AI to "infuse scenes with photoreal lighting and materials," has been mocked as "sloptracing" and accused of distorting character designs with "yassified" features like hollowed cheeks and pouty lips.

During a GTC 2026 interview with Tom's Hardware, Huang insisted DLSS 5 "fuses controllability of geometry and textures with generative AI" rather than altering artistic intent. He described it as "content-control generative AI" — a term that has done little to quell confusion, especially after the demo video showed Resident Evil's Grace Ashcroft receiving an AI makeover with "hollower cheeks and poutier lips" (a visual that even Nvidia's own presentation couldn't avoid, as a Starfield character developed a "giga-nostril" in the AI rendering).

The backlash stems from DLSS 5's apparent reliance on 2D frame processing rather than the promised 3D geometry integration. When PC Gamer reporter Daniel Owen questioned Nvidia engineer Jacob Freeman about this discrepancy, Freeman admitted DLSS 5 "takes a 2D frame plus motion vectors as input" — a far cry from Huang's "generative control at the geometry level" rhetoric. The result? A visual style that feels like a "glorified Snapchat filter" (and frankly, it is), with characters' faces exhibiting uncanny valley effects that shift unnaturally during motion, creating a disorienting experience that's "less like a game and more like a video generated by an AI model shared with the caption 'Hollywood is cooked'."

Huang's insistence that developers retain "direct control" over the AI's output has been met with skepticism. The feature's default aesthetic — which reinforces "bland beauty standards" and produces "AI slop" — makes it difficult to believe that fine-tuning could prevent the homogenized look. As one Reddit user noted, "It's like an asset flip but infinitely worse; at least in those the assets are paid for and licensed." The irony is palpable: a company whose $5 trillion valuation hinges on AI infrastructure now faces accusations of contributing to the very AI "slop" it claims to combat.

Huang's stance is further complicated by his own admission to the Lex Fridman podcast that he "doesn't love AI slop myself," a statement that rings hollow given his company's role in fueling the AI content boom. Despite calling DLSS 5 the "GPT moment for graphics," the feature's implementation — which alters character geometry in ways that undermine artistic intent — has left developers questioning whether "neural rendering" is a misnomer for "AI overpainting." The feature is set to launch this fall in titles like Resident Evil Requiem and Starfield, but its adoption may depend on whether developers can resist the siren song of "neural rendering" without sacrificing their game's unique visual identity.

Whether developers will actually adopt DLSS 5 without compromising their artistic vision remains the real question.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn