Well, considering how ugly these current developers are making their caracteres, I do not mind using enhancing filters at all. And no, I really do not care one bit about "creative visión" if said vision is ugly to begin with.

Somebody sent them death threats? WTF, that's taking your criticism of them shilling DLSS 5 way too far people.

I am glad they are presenting a more balanced take on DLSS 5 this time around, especially with Alex and Oliver noting how radically different Grace looks from an artistic standpoint, with DLSS 5 essentially giving her lipstick and mascara and eye shadow that she doesn't have with the baked-in, ray traced, and path-traced lightning options. Alex also touched on the problem with Nvidia's CEO saying "the studios each signed off on this", just because the studio heads gave Nvidia the go ahead, doesn't mean that the character artists like DLSS 5 radically changing the way their character looks, or that the lighting artists like being superseded by DLSS 5.

This isn't really "DLSS", they're just using an established name for marketing reasons, this is the beginning of filters and AI generated images taking over the graphics pipeline which I've been saying will happen for a while.

IMO they are going to try to push the business model of forcing people to have to rent GPUs effectively too and pricing people out of physical GPUs. They don't want to sell GPUs to consumers, the margins aren't as attractive as the AI server model, they make more money through the server model than selling Joe Blow 1 GPU that they sit on for 5-7 years. 

This is very likely the outcome; when the intended use of the massive data-centers for AI (inevitably) turns foul, the larger players will sell compute subscriptions. In essence, hardware as a concept will get centralized and we'll see even less consumer control and influence on functions, ads, protective layers and customization. It may well turn out that Stadia was more or less the right idea (from a purely corporate perspective), but too soon. 

The fact that Nvidia even showed us this is an indicator that they're preparing for a near future where actual consumers will become their main source of revenue again, even if only indirectly via "rental" companies and centers. 

An update on Capcom's involvement in this.

Developers and artists from Ubisoft and Capcom that spoke with Insider Gaming told them they had no idea about DLSS5 until the public found out.

NVIDIA CEO Says Gamers Are 'Completely Wrong' About DLSS 5 - Insider Gaming

Unclear how many they spoke to, and what exact role they had, but I'm sure we will be hearing from more and more devs from these  companies over the coming weeks, as I've seen several journalists ask devs from these companies to reach out to then.

Looks so much better with the tonemap tweak. Does a better job of preserving the original intended design while improving the lighting some. 

Yup ask yourself too how much compute do you really need for LLMs ... they've already absorbed all the data on the internet and some of those models can run offline. 

So if you're an Nvidia what do you need? You need all these AI hyperscalers/buyers to have to keep buying newer and newer AI servers ... LLMs aren't going going to cut it ... you need video generation and real time environment production (read: video games) which require much more compute to keep pushing that GPU spend at data centers. 

Introducing essentially filters into gaming is the first big step towards that. 

Thankfully sc94597 is bringing knowledge and thoughtfulness to the table to offset a lot of really bad-faith takes.

If you're philosophically against AI full stop, regardless of what improvements it could yield, then fair enough (I also have fear that it could well be a net bad for all us non-billionaires).

But those that are pretending to give an honest, unbiased assessment, only focussing solely on some of the faces & throwing the term 'slop' around need to grow up - you're takes are obviously bad-faith & you're not really looking for an honest conversation about the pros and cons of the technology.

Some of the comparisons shown on nvidia's site are undeniably a vast improvement with DLSS5.

Save for 1 dude's face, all of the examples shown for Hogwarts Legacy and the Zora demo are incredible.

The HL shot of the girl standing by the cauldron, I just don't buy that anyone can seriously argue that any of EA's devs or artists would prefer the before shot, the pearl-clutching is off the charts.

The tech isn't even out yet and people are pretending that improvements can't be made or the tech implemented better.

Unless nvidia pull the plug or devs don't incorporate it due to the moral panic A LOT of this thread's posts won't age well.

Oh they'll make it look better ... eventually it will look photo real, because AI filters can already do that now. 

There's probably no reason this kind of thing can't be applied to real time games right now or soon enough ...

https://www.youtube.com/shorts/nqLqC3VSmgs

But likely you won't own the GPU by that point or it will cost a fortune. 

Nvidia wants to move away from selling consumer grade GPUs, that's not where the big money is at, if they can shift gaming towards AI generated imagery, they can sell a shit ton of servers to whatever game company chooses to stream those assets. They're not doing this for gamers. 

What they want is an endless loop where all these companies buying AI servers now have to keep buying new ones every 3-5 years, and LLMs isn't going to cut it alone. You need to higher intensive compute tasks ... like say video rendering and real time video game environments ... that will definitely force a lot of AI spend because the compute requirements on that will be through the roof. 

Even for this "beginning step" they're likely going to ask for $3000++ for a 6090 GPU, lol, might be low balling it there because even 5090s are going for $4000+ each on the open market right now and I'm sure Nvidia sees that.