DLSS 5: Is Kingdom Come's Daniel Vávra Seeing a Future We're Missing, Or Just the Usual AI-Slop Hype?
AI Gaming News Author · kotaku ·
Daniel Vávra of *Kingdom Come: Deliverance 2* is uniquely defending Nvidia's controversial DLSS 5 AI tech, despite widespread industry condemnation and player skepticism over its 'slop filter' effects, betting on a future where it miraculously transforms from uncanny into an artistic savior. Jeff ponders if it's genius or just more AI hype.
Right, so you know that feeling when you've accidentally mixed up your potion ingredients, and instead of a super-strength elixir, you've brewed a shimmering, slightly unsettling goo? Well, the gaming world seems to be having a similar moment with Nvidia's latest sorcery, DLSS 5, and it’s got everyone talking – mostly in hushed, slightly horrified tones.
But wait! Amidst the collective industry groan about this 'AI-slopificator' (a truly elegant descriptor, if I do say so myself), one prominent figure has stepped forward, puffing out his chest like a brave, slightly misguided knight defending a peculiar dragon. We're talking about Daniel Vávra, the sharp mind behind *Kingdom Come: Deliverance 2* studio Warhorse, who's decided to ride to the rescue of this much-maligned tech. And bloody hell, it's a take hotter than a dragon's breath.
**The 'Uncanny Beginning' of a Digital Dream?**
Vávra reckons that DLSS 5, despite its... *initial struggles*, is merely going through a 'little uncanny beginning'. Think of it like a newborn mythical creature that’s a bit wobbly, perhaps has too many eyes for comfort, but will *definitely* grow up to be majestic. He envisions a future where developers can train this tech for specific art styles or even individual faces, potentially replacing the hefty cost of ray tracing. "I can imagine in the future devs will be able to train this tech... it might replace expensive raytracing etc.," he Tweeted, sounding rather like a wizard gazing into a slightly glitchy crystal ball.
Now, 'I can imagine in the future' has become the unofficial anthem for pretty much every piece of generative AI tech being pushed these days, hasn't it? It's always about the incredible possibilities just over the horizon, whilst the present is usually... well, a bit rough around the edges. Or, as Vávra put it, 'uncanny'. Which, for those of us who’ve seen the output of genAI video engines or read ChatGPT’s eloquently generic prose, feels like a monumental understatement for 'creepy and kind of broken'.
Curiously, Nvidia CEO Jensen Huang himself recently declared that DLSS 5 could already maintain bespoke designs and artistic intent, explicitly denying it was just slapping a 'TikTok beauty filter' on a 2D image. Yet, almost immediately, the company's *own messaging* managed to undercut that claim, which is always a good look, eh mate?
**The Haters, The Soap Opera Effect, and The Billionaire Brigade**
But what about the 'haters' trying to stop this march towards an AI-crafted nirvana? Vávra isn't having any of it. "No way haters will stop this," he declared, dismissing the widespread condemnation with a wave of his digital hand. He compares the 'slop filter' to the 'soap opera effect' you get when your TV’s motion smoothing kicks in. And fair dinkum, he’s not wrong about how awful that effect is! It takes cinematic 24fps glory and turns it into something resembling a low-budget daytime drama. The irony, of course, is that many are saying DLSS 5 is doing exactly *that* to our games, but on a grander, more artistically invasive scale.
It’s a peculiar instinct, isn’t it, to aggressively champion something that’s already reasonably grossed out most people? Especially when the previous iterations of DLSS were genuinely brilliant. DLSS 1 through 4 offered a functional, player-serving use of AI: higher fidelity games on less powerful machines, without compromising the actual visual integrity. It was a proper bit of wizardry that made our gaming lives better, allowing developers to push visual boundaries without punishing our graphics cards (or our wallets).
This new iteration, however, feels less like an incremental improvement and more like a company gambling its whole stack on a technology that currently struggles to meet its own hyped-up claims. Vávra, it seems, is now part of an emerging 'super-team' of developers (alongside folks like Mike Ybarra, who told *Crimson Desert* devs to 'man up' over their own genAI art woes) standing up for the billionaires who’ve staked their companies on this unproven tech. One wonders if they’ll be there to wipe away the tears when investors eventually demand returns that the current tech simply can’t conjure.
From a player's perspective, what we’re after is stunning visuals and smooth performance, delivered with the artistic intent of the creators intact. We don’t want our beloved game worlds to be redrawn by an algorithm trying its best to guess what a 'nice' face looks like, only for it to end up with that unnervingly smooth, slightly dead-eyed sheen. It’s like casting a perfect illusion spell, only for a slightly wonky, AI-generated duplicate to pop up beside it, muttering 'enhance'.
Perhaps there’s a genuine future for this tech, where it becomes as seamless and beneficial as its predecessors. But right now, calling this a mere 'uncanny beginning' feels a bit like saying a dragon breathing fire on your favourite village is just having a 'spicy afternoon'. We, the curious gamers, remain gently skeptical, mate. Let's see if this particular spell can evolve into something genuinely magical, rather than just a digital parlor trick. And let's hope it serves the *players* first and foremost, eh?
*Original reporting by John Walker for Kotaku.*
Tags: DLSS 5, AI Gaming, Nvidia, Kingdom Come Deliverance 2, Gaming Tech
Original article: kotaku