Nvidia’s DANGEROUS Plan!

Nvidia's DLSS has gotten even better than what we originally thought. But there's a serious issue with it that no one's really talking about. To illustrate my point, I'm going to go over some of the new stuff that dropped, proving that yes, DLSS is almost like magic, but then I'm going to tell you exactly why it's a serious problem and everyone should be worried about it, but also what to do. So, when DLSS 4.5 was first announced with its second generation Transformer model up to six times frame gen, etc., I went over the impressive tech as well as some of the videos coming out that showed just how good it looked. Since then, Nvidia has released a full stable version and some of the newest videos are wild. Case in point, we have this video here from this guy who's gone in and forced games to render at stupidly low internal resolutions and then upscaled it to 4K with DLSS 4.5. All in an attempt to see when the image becomes usable. And by low resolutions, I'm talking as low as 38x 22, not 2200 or 220, but 22, essentially at a 1% render scale of 4K. And as you'll quickly note, it looks terrible. Though when we look at 136x76, you can see that it's actually night and day better than the original render. It's almost shocking. In fact, he found out that at just 764x430, the resolution a potato could output at just 20% of 4K, it actually doesn't look that bad. Not amazing, but it looks playable. And this is where things get scary. If you remember, when Nvidia originally launched their RTX 4000 series, the company made some wild claims about performance to the point that everyone was really excited. I was skeptical at the time, but then we found out for sure that those numbers only came when using their new at the time frame generation. Fast forward to their 50 series launch, and Nvidia did it again with their 5070 issuance at just $549. I was there at CES when it was announced and I remember being really surprised that some people, even some journalists in the audience, didn't seem to know that this was only with their newlyannounced multi-frame gen. I figured everyone would have understood that, which is why I went back to my hotel room, made 100% sure that I was right, and then made a video on it. And that's the part that really has me worried. Of course, everyone today now knows that the 5070 claim deserved a giant asterisk by it. But the simple fact is that Nvidia is clearly planning to continue this trend which could be a serious issue for gamers and hardware in general. But the thing is that it actually does come from a need. Simply put, Moore's law is dead. Of course, it was never a law anyway. But that aside, the simple fact is that performance gains in PC hardware are getting harder and harder to come by. Companies have done some incredible things to get around limitations like chiplet technology, newer packaging like 3D vcache, but they still need smaller transistors and that is getting more and more expensive. I'm talking TSMC's 2nm node is looking set to be double that of their 5nmter node. Now, the increase in density is helping with that, but it's not completely offsetting it. Tooling is more expensive. It's just harder than ever to do. On top of all this is a huge push for ray tracing. Now, I will say that gamers aren't really the ones pushing for it, but it is the logical next step in graphics. Ray tracing is nothing but a simulation of natural light in a scene. Rasterization tricks can do a great job at faking light with shadow maps, fake reflections, and all that, but it's never going to get us to ultra realistic. I understand there's an argument from gamers not wanting realism, art style, trying to be an escape, not real life. I get it. I mean I still play Runescape from time to time but we are for the most part trying to approximate real life. So ray tracing is definitely required for that. The issue there as many have noticed is that ray tracing is very computationally expensive. So we have two problems that hardware makers are trying to solve. We need way faster GPUs than what's possible right now. And it's really hard and expensive to get even a little faster. Meaning it's understandable that Nvidia, AMD, and Intel would release technologies like upscaling and frame generation. It's a way to solve a lot of these problems. But what happens when the solution becomes the crutch? We know that game developers have already leaned on DLSS so they don't have to better optimize their games. Now, imagine hardware makers doing the same as a way to make their cards seem better than they are, which could easily incentivize Nvidia to make DLSS better instead of their actual hardware. Because let's not forget, making the hardware better is extremely expensive and just difficult. Why come up with creative new ways to solve hardware limitations when you can just train the AI better? Now, some people might think that's okay because in the end, you get a GPU that plays a game at a higher FPS. But besides the fact that we want innovation to continue, imagine paying $1,000 for a card that's really just as fast as a 3050. The only difference is that it can play this new DLSS that last gen can't. And that's the real issue I can see happening soon. Nvidia would obviously love that. They have a huge incentive to push this. Their profit margins could skyrocket and I have no doubt that it's crossed their mind. But don't worry, I'm not just going to complain that something's happening and tell you that nothing can be done. I actually have a few ways that we as gamers can help stop this. For one, don't let Nvidia or any company get up on stage and make absurd claims about a GPU's performance. I'm not saying that upscaling and frame generation aren't important technologies or anything. They are, and it is impressive, but we shouldn't let companies make a hardware claim about a new software, or at the very least, they need to be extremely transparent about it. Things may have felt really muddled lately as to what's with DLSS, what isn't, etc. And we have to push back against that. This confusion is almost certainly intentional. And second, we need to push for more competition. Even if you're an Nvidia fan, you want AMD to succeed and vice versa. Because if either of them win, your GPUs will objectively get worse. Especially if they can just throw some new software out that technically gives you higher FPS. For example, do you really think Intel would be releasing a 52 core CPU, potentially with a huge amount of cash, if AMD hadn't clocked them with their X3D chips? Of course not. Competition drives innovation, and it helps to force prices lower. If Nvidia can name their price, they will. And that's true for AMD, Intel, or anyone. We want competition in the market, but that's of course not necessarily one that we have a ton of control over. So, finally is to control what you buy. And I don't necessarily mean to buy the competition or anything because buying AMD just despite Nvidia doesn't help AMD's products get any better. And let's be honest, while Nvidia's DLSS is currently winning, if AMD's overtakes them, they could try the same thing. No. The next time someone makes a claim about a GPU that isn't true, tell them you aren't going to buy it and why. DLSS should be a bonus, not the foundation. Because once software becomes the selling point, hardware becomes optional.

Nvidia has a plan that's not good. And it all comes from DLSS. We can't let them get away with it! Stay tuned... ****Items featured in this video available here**** Newegg (Affiliate): https://geni.us/newegg1 Amazon US (Affiliate): https://amzn.to/3b9UjKB ►GET THE MERCH: https://meldstore.com/ ►CHAT ON DISCORD: https://discord.gg/gamermeld ►SUPPORT ON PATREON: https://www.patreon.com/gamermeld ►GAMER MELD SPONSORS: https://www.gamermeld.com/sponsors X: https://x.com/GamerMeld Facebook: https://www.facebook.com/gamermeld SOURCES: https://www.youtube.com/watch?v=DKCyk3CeUFY https://www.youtube.com/watch?v=FEdKM98n5pg&t=1686s https://www.computerhistory.org/siliconengine/moores-law-predicts-the-future-of-integrated-circuits/