'Bad ending: now every game is slop': Game developers share mixed reactions to DLSS 5

After teasing it as "the future of real-time rendering", Nvidia has finally announced DLSS 5, an AI-led model for photorealistic visuals…that mostly looks like an AI filter slapped over Resident Evil Requiem. However, it's game developers, with their honed art direction and style, who will care the most about DLSS 5, and that reaction has been mostly mixed to put it lightly.

Over on X, New Blood co-founder Dave Oshry shares a meme calling the tech "Pure Slopium", a reference to AI slop. 'Is this a 3D model?', an account dedicated to teaching X users about 3D models, similarly, labels DLSS 5 a "slop filter" and calls Nvidia "an absolute joke".

The word slop is used a lot in regard to Nvidia's announcement. Indie developer Guselect says, "bad ending: now every game is AI slop."

Over on BlueSky, Karla Ortiz, a Puerto Rican artist who worked for Ubisoft, Blizzard, Marvel and more, says:

"This is so disrespectful to the intentional art direction of devs. If devs wanted to lean in to hyper realism they would," she says. "This also drastically changes key aspects of visuals like character features, focal points, lighting and so on. What a terrible invention. Nvidia should shelve this one."

This is so disrespectful to the intentional art direction of devs. If devs wanted to lean in to hyper realism they would. This also drastically changes key aspects of visuals like character features, focal points, lighting and so on. What a terrible invention. Nvidia should shelve this one ????

— @kortizart.bsky.social (@kortizart.bsky.social.bsky.social) 2026-03-17T13:46:17.934Z

Some fears about the tech are more existential than simply perceiving something as looking bad. Jon Ingold, narrative director and co-founder of Inkle, shares a screenshot of its game Heaven's Vault, arguing that DLSS 5 would remove its main character. Heaven's Vault is a game largely about archaeology, history, and a desire to preserve culture, so this post implies DLSS 5's AI beauty filter style would actually erase the identity of characters that don't fit certain beauty norms.

This same fear is expressed by illustrator Corey Brickley, who says: "What if we blended every famous woman into one woman and then that's every woman you see now."

Chris Gardiner, narrative director at FailBetter Games, calls this AI beauty standard the "Scarlett Johansesonification of videogames." Catchy.

Kansai-based game dev Alwei critiques what the model does to lighting, plus expressions, and art direction, stating: "The artwork created by artists has a solid intention behind it, and if that can't be controlled, it has no meaning."

Sam Barlow, known for his work on Immortality, Her Story, and Telling Lies, argues that the choice of games intentionally ignores those which use face models, too. "Can you imagine the legality of and just the optics of your game starring Lea Seydoux and Elle Fanning changing their faces? Making Conan O’ Brien look like a catalogue model Chris Hensworth?"

Funny thing with the DLSS stuff - they only picked games that don’t use actor likenesses. Can you imagine the legality of and just the optics of your game starring Lea Seydoux and Elle Fanning changing their faces? Making Conan O’ Brien look like a catalogue model Chris Hensworth?

— @mrsambarlow.bsky.social (@mrsambarlow.bsky.social.bsky.social) 2026-03-17T13:46:18.119Z

Not all game developer reactions are negative towards DLSS 5. In the announcement for DLSS 5, Todd Howard shared, “When Nvidia showed us DLSS 5 and we got it running in Starfield, it was amazing how it brought it to life. We've played it. We can't wait for all of you to do so as well.”

For positive reactions outside of Nvidia's official press release, Kazuya Okada, ex-Epic Games software engineer, wishes "we could use super-high-quality images and videos—created in advance in the game engine's editor, ignoring the load—as training data, and then apply them to actual upscaling somehow…"

3D model creator tarava777 says, "It's ultimately an optional process handled by the GPU, essentially a kind of 'aftermarket mod.' You can't do it without a GPU, and it can be turned ON/OFF at will.

"This is just a tech demo, not something that's going to be released as-is. We need to take those aspects in stride with a cool head, right? Personally, I think advancements like this are in a realm where 'whether you like it or hate it, there's no stopping them anymore'."

The 'optional' element of that is certainly notable. Developers who don't want to use it, or gamers who don't want to see it, can choose not to do so. And there's some excitement in favour of using it too.

CG artist and the writer of the CG art blog "3Dnchu" argues, "It's amazing that this can be done in real time...This feels like an evolution that's gonna stir up a lot of buzz in all sorts of ways, huh."

Some like actor Rahul Kohli mainly met the reaction with dumbfoundment, "DLSS5 has to be a joke right? Right? Guys?"

Bruno Diaz, who worked as a senior writer and lead narrative systems designer for Failbetter Games, argues that prohibitive hardware requirements could stop it from making a splash anyway. "If DLSS 5 actually ships, it'll probably be so performance costly that few people will be willing to use it."

The demo for the tech used two RTX 5090s, which would cost the average gamer around $8,000 to get ahold of. It is worth noting that these games are running at 4K max settings, so cranking resolution down to 1440p or even 1080p will lower power requirements. Nvidia says it has got the tech working on a single GPU (presumably just one measly RTX 5090), but that is still a lot for your rig to be able to run it.

One can assume the tech has been revealed because Nvidia assumes gamers will actually be able to run it once it launches in the Fall, but we'll need to go hands-on to understand its hardware limitations.

Ultimately, this tech is very new and still relatively early in its development. Most have not had the chance to go hands-on with it, and how it will perform in real-time will be the most telling aspect of its release. Game developers online, or at least the most vocal of them in the West, seem rather sceptical, and Nvidia's Grace Ashcroft AI makeover perhaps isn't the best selling point for it.

That's not helped by the fact that DLSS 5's presentation is just over one minute of a two-hour GTC keynote presentation, with little explanation on how the model works, how much it can be tweaked, among many other things.

Nvidia says devs have "artistic control" with DLSS 5, but it could go so much further to specify how. Whether or not this will really be the future of real-time rendering will largely depend on mainstream adoption (both from gamers and developers) and on how easy it is to run, and that's not something easy to grasp from initial reactions. With DLSS 5 set to launch later this year, that will be the real testing ground for whether or not people actually want it, and if that's better or worse for the games, as a result.