It probably is a screen-space effect, like is the case with all DLSS models (outside Ray-Reconstruction.) I don't think the engine is feeding in game-data on material properties and light sources and then these are corrected before rendering the frame, as an example. It's not a PINN (Physics Informed Neural Network.) But I also think there are limits to what 3D artists and tech media personalities know about deep-learning models and they think this is a big-gotcha. 

Image-segmentation/semantics feature modeling is extremely good in the year 2026, even for tiny models, and Nvidia has already mentioned that this model uses semantics to help it obtain information about objects in the scene. This is a far more efficient way to do it than to feed it heavy game-data. 

Sure there is a risk of hallucination going this route, but what they're missing is that training objectives matter. Image and video models, even "filters", aren't trained on the same objectives as any DLSS version, including DLSS 5. There is no reason for one to believe that two models pre-trained on different objectives and with different data distributions, even if they use the same neural-network architecture, will have the same limitations. 

There's no way we can have a safe, livable world in an AI future. Once AI develops a survival instinct it will kill us all.

All this is happening because billionaires wanted to be richer and eliminate human workers.

So many of those A/B tests look absolutely awful with DLSS5. Like, my god, the original screenshots I saw were bad but it is even worse than I thought. 

Pretty much all of this. I wrote a long post about this a couple of months ago, so I'll keep it short-ish, but so-called "A.I." is indeed a massive net negative. All of this destruction, all of this effort to force Plagiarism Bot 3000 on us, as if every new gee-whiz tech fad from Silicon Valley has to be "THE INEVITABLE FUTURE™ SO GET USED TO IT, LUDDITE!!!" They don't care if they have to destroy everything from the environment to the very soul of human creativity in the process. "A.I." reeks of a get-rich-quick scheme, a desperate last-ditch attempt at making the line go up in a moribund economic system. Capitalism will be the death of us all, and fossil fuel and tech companies are leading the charge. If it were up to me, "A.I" would be banned and all those new data centers being put up would be bulldozed and replaced with something actually useful.

That's not really true anymore, at least in advanced economies. On a per capita basis, electricity use did indeed climb substantially over the 20th century, which makes sense considering more and more households had electricity to the point of being near universal by mid century, then after WW2 more and more households started buying more creature comforts like household appliances and central heating & air. The latter is actually the largest driver of household energy use. But now just about every household that wants or needs those things has them, and for the past decade per capita electricity use in most advanced economies has not only stopped growing about 15-20 years, but has actually declined a bit thanks to increased energy efficiency. Turns out that household energy use does have a rough upper limit and doesn't continue to grow exponentially. In fact, U.S. electricity generation has remained relatively flat since the mid 00s, despite the increase in population, hence the downward trend in per capita use.

Anything that causes an explosion in electricity generation in developed nations is purely optional. Up until this point, we've been generating all we need for pretty much all residential and commercial purposes.

I think it was Slavoj Zizek who said "it's easier to imagine the end of the world than the end of capitalism."

Almost every ethical problem being attributed to "AI" really is a problem with the socio-economic system we have, as these problems manifest themselves in other technologies as well.

But just like people can imagine the end of the world before they can imagine the end of capitalism, they also imagine that we can somehow erase a broad range of technologies before we can end capitalism.

But really if one wants to solve these ethical problems one needs to be a serious anti-capitalist. If you're not, then you're not very serious about these issues, in my opinion.

^Why we must destroy all AI servers.

This video has some interesting insights, and it's straight from the horse's mouth at that. Many issues that people bring up in this thread seem to be addressed, and it's worth a watch (hint; the detractors may have it right this time). 

Not to mention every other issue with AI. It's killing the economy and soon it will kill us all.

I think one issue I have with this sort of video is that any discussion with any single Nvidia rep is going to be constrained on the reps end, especially when there is question-begging. 

Having said that, we do get useful information in this video, so I think it is very valuable for that alone. 

The conclusions are a bit hasty though. The framing of "only a 2D image and motion vectors" for example makes it seem like that isn't enough, when having access to the buffers is a big advantage DLSS 5 has over VLMs (which need to build heuristics in their features for object motion.) 

And yes, material properties and lighting need to be inferred, but DL models are pretty good at that, and have been for a while. 

What we've seen so far is a lot more temporally stable than early versions of DLSS 2, for example. 

The biggest issue is really the style shift issue. That will have to be fixed through artists, rendering engineers, etc working iteratively. It might also mean that they use neural shaders (which do directly access the material values) in their render pipelines when DLSS 5 is applied, to get that fined-tune control.

Even if they fix these issues it won't detract from the the fact that AI will kill us all one way or another.