Sure but at what cost.  I'm serious.  This needs 2 5090 cards.  Its starting to feel like they are going to use Instagram filters to push $6000 cards.  The practice of generating value by forcing the AI generated look onto games and pretending this is what people seek.

God, this looks uncanny and outright awful. This must be one, if not the worse demos nVidia has ever released.

Development costs are through the roof. AI will be leveraged significantly to lower costs.

Like it or not, it is coming.

I'm a little torn on this. The effect is pretty amazing, there's no denying that. But it's also a little uncanny valley and weird. Feels like an Instagram filter that has to hover around in real time and try to keep up with what's happening.
Of course, as anything, this will get better and better as we go. I think this will work better as a ground up solution rather than after the fact.

Beyond A.I skeptisim though there is just genuine criticism over what its representing. Even 4 years ago this bottom image would look fake or photoshopped.

DLSS <4 is different because it maintains the vision of a game with only the slightest misrepresentations (occasional textures/particles fx etc). This is several 100 magnitudes more disruptive. 

Just dissecting some of what people are responding to:
 
The level of fidelity is in no way matching the movements/animations and it comes across as extra weird/uncanney valley.

The lighting itself in the bottom remains stylised but in a way that isn't coherent to the environment. This hero lighting is sometimes done intentionally in games but here it just looks like the ai model is applying it on everyone and making each frame look like a photoshopped movie poster,  another hallmark of generic generative ai.

The ai model is clearly trying to exact a certain "look" out of the characters as opposed to just increasing the quality of lighting. Looking at the guy on the left, is there a genuine light source making is right side so brightly contrasted? I'm not sure. Again even before Ai this would come across as a highly doctored image.

You're 100% right that many normies will eat this up, but I think the criticism is also legit and not just kneejerk. They'll be a huge divide in how people take to this kind of execution, similar to how you see some people cream their pants over fan-made Mario/Zelda UE5 demonstrations whilst other people point out how souless and generic they often look. I emphasise this execution because this could definitely be refined to actual look authentic, right now it doesn't.

I'm not necessarily concerned because this is a toggle but it does reflect a direction of games away from art and laboured/intentional outcomes, into benign realism. That Assassins Creed screenshot example to me is actually the best exhibit. All tone is sucked out of the image.

IMO this was a poor showing of the technology and Nvidia+the developers should of have been a bit more careful and held back until it looked right

Oblivion Remaster was a train wreck.  Final Fantasy 16 was piss poor.  And TLoU Part 1 left a lot to be desired.  

I tend to agree with your take, I think AI makes most developers lazy.  They will sit back and expect people to use AI to fix.  

What I find funny is how the water/rain effects in Resident Evil specifically look worse. The alley behind Grace is dark, wet, then it's turned on and it looks lighter and all water reflection vanishes. It basically gaussian blurred the water effects.

I think this is interesting tech, I'm skeptical about how it will work in practice.

Lets be frank, everybody making fun of it would try it out if its simply an option in the graphics settings.

It's amusing that people keep repeating this since they have no idea how much water is used by typical agricultural and industrial processes. A large datacenter consumes roughly as much water as the beef used by six burger joints, and golf courses in the US consume around 25x more of it than all datacenters combined. Not to mention the runoff from a datacenter is warmer water... the runoff from a blue jeans factory will literally kill you.

As for errors, retrieval-augmented generation has by large reduced the rate of hallucinations in the past year or so. Even for offline mode, the hallucination rate changes considerably across models. They can actually learn to distinguish between what is encoded in their weights and what is not, though no one besides maybe Anthropic is paying much mind to solving it in reinforcement learning.

The algorithmic cost for deep learning techniques drops massively year on year. It'll probably be running on most (50-series?) consumer cards by the time it launches. At least one would hope so, Nvidia probably isn't stupid enough to release it for just a single very expensive, supply-constrained card.

Exactly my thoughts. The original version actual looks more natural even if the lighting is less detailed