I don't have issue with any of the above, also because the new incarnations are in themselves artful. DLSS 5 is for me worse both because of broader societal landscape it relates to with AI & algorithms & automation, but also in the fact that the end result is creating cohesion issues with the game world, animations and actually feeling unnatural. No one would look at Demon Souls Remake with fresh eyes and say a bad word about it but something does actually feel a bit off about many of the DLSS 5 demos and you immediately get the AI look. Not everyone is sensitive to it, but to me its very immediate.

MSG3 is probably the only recent remake where I rolled eyes at. Not even a big fan of series so I didn't care but I remember how distinct 3 looked and how generic the remake is by comparison. It feels a missed opportunity. You can just tell they didn't really respect the source material with the cinematography. I know they have a separate mode which appears closer but even the idea that they would do away with it is like someone remaking the matrix and removing the green and adding it as a hidden bonus 



Aside from the grading I think its also interesting to explore rendering and styling that evokes the original piece. Final Fantasy XVI is a prime example of how realism doesn't have to do away with distinct artstyle and depictions of people

I am meh about it, but isn't this one of the things that looks the worst now and will only improve?

Give me photorealistic graphics on PS6 gpu hardware. I think we are ready for it.

I don't think asking about whose intent is the authoritative intent or if there even is one authoritative intent in collective enterprise is just semantics. It's a meaningful question that identifies a meaningful complexity. The point of considering artistic intent is that you are considering all that goes into a final product and the full inherited context thereof from the artist(s). Identifying who they are is important. 

If the issue is instead one of artistic cohesion or consistency, then that is an entirely different argument, and I think one less supported than even the intent one. This game looks drastically different depending on which render settings you play in the current version: no-RT vs. RT vs. PT. I've experienced this first hand playing on a 5090 vs. SW2, right after each-other. 

And you can even see how the real in-game model differs from concept art, in each of those rendering contexts. 

She is drawn with quite more angular/contoured features here than in the final game model, where her features are more rounded and "younger", as an example of one difference. 

The following image is not fundamentally any more distinctive than what we've seen above. 

Whose artistic intent matters here? The concept artist's? The 3D modelers'? The producers'? In reality, this is a product of collective force, the mixed labor of many people, and you're always going to get some inconsistencies unless the team is running a very tight, and rigid ship where all decisions are centralized. 

I also don't agree that DLSS 5 is doing the same thing in every image, and think that is something people are saying without any real evidence. It wouldn't make sense from a technology perspective where deep neural networks are very sensitive to inputs. 

Oh great Netflix lighting is coming to games! /s

When it comes to games I generally turn al things like this off since it often lower the frame rates. Frame rate is way more important for my experience and impression of the picture quality than the lighting effects. In many cases, even if I somewhat know I'm wrong, I think ray tracing makes things look worse even in still pictures.

I want something like this but used instead of black frame insertions to get smooth movements but sharp pictures for watching movies without studder and without nasty artifacts from the processing of TVs that try to invent frames that does not exist.

This is exactly why I bring into discussion what is delivered in the product people have born a relationship with, what is a technical limitation versus what is not and what is established of audience expectations. Unless we worked on development, none of us are in position to actual authoritatively state where ownership lays but we can still make a judgement call on something appearing too deviating.

The point I'm making about semantics is to lean into proper understanding of someone's friction here without getting caught on 2 words when there's a whole wider sentiment to be understood behind those 2 words, which might better be expressed in nuanced ways. Unless of course you are just intent on getting into philosophical discussion about "artistic intention".

As mentioned there are many examples outside from DLSS 5 where similar discussions can be had where it seems obvious that contention over artist intention makes sense and is valid even if the feeling is not shared (Windwaker etc). That does not dictate how you personally move forward with your preference but you can understand someone else's resistance. The topic is obviously subjective, so I don't care to delve much further. I just wanted to respond in case you genuinely thought that the topic was put to rest because Capcom worked on it. I can assure you its not and by your own comments about the nuance of artistic ownership I presume you also agree. The same would apply to Nintendo working WWHD or SE working on FFX HD etc

According to Digital Foundry, NVidia used two 5090s for the demo, one for the graphics, one for the AI. 

That instantly means if you have only one 5090 gpu, a massive downgrade in resolution is needed. If you have less than a 5090 gpu (which likely 99% of the people have), then the whole thing becomes more or less unusable. Also the thing is tied to frame generation..

Updated the OP with Digital Foundary response/address video:

Will probably watch it tomorrow

This is obviously not going to be the case in the product release. With work-in-progress models like this, a reduction in model sizes in both memory consumption and parameter counts happen very rapidly as the model goes through a process of knowledge distillation iteratively. 

We also have to remember that the tensor cores are the main processing units for DLSS/AI workloads, and they are a very small share of all of an RTX 5090's cores. It's quite possible that they just needed extra tensor throughput in the current iteration and they chose RTX 5090's for the demo because they don't want to have any issues.