© ROOT-NATION.com - Use of content is permitted with a backlink.
This piece will be short, but dense, because it focuses on the announcement of NVIDIA DLSS 5. We already have a separate article on DLSS 4.5. Another one is also in the works, and neither of them is directly connected to this topic. This stands on its own. That said, I’m still going to tie Goodram Move Ridge into the discussion – and I’ll explain why in a moment.
TABLE OF CONTENTS:
Video About DLSS 5 and Goodram
About the SSD
First, a quick word about Goodram. Technically, this video is sponsored thanks to the external Goodram Move Ridge drive, which stores all the test files for games and software across both desktop PCs and laptops. It’s the setup I used while testing DLSS 4, DLSS 4.5, FSR, and pretty much everything else I’ve benchmarked recently.
And the only reason the review of this 1 TB drive did not come out earlier is that I ended up dealing with appendicitis – and my work schedule does not really tolerate interruptions like that. Overall, I’m satisfied with this SSD, especially its compact form factor, although it does have a few caveats that I’ll cover in the full review later on.
Briefly About DLSS 5
For anyone who missed it: NVIDIA essentially showcased an early prototype of DLSS 5 – and it appears to alter rendered graphics even more aggressively than DLSS 4.5 with its newer transformer-based models.
Developers will be able to use DLSS to perform ultra-realistic reconstruction of faces and other visual elements. The technology increases detail, improves color rendering, enhances lighting realism, and boosts contrast. Some of the scenes shown, at least to me personally, no longer look like rendered graphics at all. They resemble live-action footage or cosplay shots rather than a game engine output. And honestly, that is probably the highest compliment I can give to computer graphics in general.
But there are important details here right away. First, to demonstrate this solution in 4K at maximum settings, NVIDIA used two GeForce RTX 5090 GPUs – not one. One card handled the actual game rendering, while the second was dedicated entirely to processing the effect itself.
Criticism… and Criticism of the Criticism
The internet, unsurprisingly, immediately went after NVIDIA for what many called the “slopification” of graphics. And to be fair, at this stage the image does look very AI-generated in certain scenes. The amount of memes this produced is honestly impossible to measure – either in words or in laps around the planet.
And this is where I start going against the current. Because there are three points worth considering here. Yes, the backlash was predictable. AI is unpopular right now, and anything associated with it is almost guaranteed to trigger a wave of hostility. But notice the distinction: I said the backlash was predictable, not deserved. Because no – in this case, NVIDIA did not really earn the level of outrage it received.
First, the company itself stated that this tool will be provided directly to developers, meaning they will be able to tune or modify presets for image quality, lighting, and other parameters on their own. What NVIDIA demonstrated was a prototype – and as a prototype, it looks genuinely extraordinary.
Second, about two years ago I made a piece on classifying artificial intelligence – from “garbage” to “revolution.” And I placed the label “Revolution” on, among other things, a tool called Runway ML, which re-renders in-game graphics into a photorealistic style. Yes, it ran at 720p and 15–20 FPS, but even then it looked absurdly promising and impressive. And at the time, I said this was where things were heading.
And what do you think happened? Two years later, NVIDIA implemented exactly that in hardware, in real time, at 4K/60 FPS, on ultra settings, using the GeForce RTX 5090. Everything turned out exactly as I had hoped. Exactly as I had imagined. The revolution arrived – and I welcome it sincerely and without hesitation.
Hardware nuance
And third – this is something I’ve heard before. I’ve already heard criticism of something that clearly had enormous headroom for the future. And that “something” was the RTX 3060 12GB. From day one, I saw a wave of, frankly, nonsense takes claiming that a budget GPU “doesn’t need” 12GB of VRAM, that the RTX 3070 was the better deal, or that Radeon offered superior performance.
And then five years passed, and the RTX 3060 12GB turned out to be one of the most resilient GPUs of its generation. It still handles modern games at 1440p and even 4K in many cases better than the RTX 3080 in memory-constrained scenarios, and it consistently outperforms 8GB cards when VRAM pressure becomes the bottleneck. Because once a GPU runs out of VRAM, system memory starts getting used as overflow – and performance can collapse, sometimes losing up to 90% of its effective throughput.
Therefore, the most powerful 8 GB card in modern games at high settings appears quite underwhelming compared to the 12 GB RTX 3060. Even NVIDIA seems to acknowledge this, as it is reportedly resuming production of the 12 GB RTX 3060 due to a VRAM chip supply shortage. I am 99% certain it will be the 12 GB version, since an 8 GB variant would likely face even stronger criticism. I do not completely rule out that possibility, however.
Summary of DLSS 5 (and NVIDIA / Goodram Move Ridge)
I consider DLSS 5 a notable development, even in its current prototype form. It reflects the direction of AI-based upscaling that aligns with what I expected from this type of technology. From my perspective, this is a practical application of AI in graphics processing. As a result, I do not find the critical arguments particularly convincing in this context.
I will be testing DLSS 5 using a Goodram Move Ridge SSD, and I will also record these tests on a Goodram SSD. That is all from my side. In the comments, feel free to write whether you are ready for this kind of future, where photorealism is already close and appears difficult to stop. I am ready for it, at least in principle.
Read also:
- DLSS 4.5: Why This Technology Changes the Approach to Path Tracing
- Testing the APV video codec on the Samsung Galaxy S26 Ultra: Why Was This Not Mentioned Earlier?
- Quantum Networks as Alternative to Classical Internet: What to Expect