Its a tiny fraction of the problem, the issue is with datacenter buildout taking all the available supply, if it was dlss directly related we would have seen ram prices increasing much earlier then we did.
But this specific iteration of DLSS is deeply wrapped up in the AI hype machine which is directly related to the datacenter buildout. The reason 5 has been a hot button topic centers around that AI upscaling showcase which only exists because NVidia is deeply invested in AI. This is what I mean that nothing in is a vacuum.
But you’re combining two things that should be understood separately to have an accurate picture of what is happening.
All I’ve seen about 5 is people complaining about how it looks, “not the original artistic vision” or just generally mocking it as bad. (Which is a personal opinion not a hard fact)
Nothing is in a vacuum sure, but let’s take the “personal responsibility” discussion around greenhouse emissions as an example, yes ordinary people and their actions contribute, but the vast vast vast majority is NOT and focusing on the least impactful level is not going to move the needle when the bigger issue is things like private jets, corporations ignoring environmental regulation, etc.
Weird because if you check search terms there’s a shitload of people asking “how do I enable dlss” and tons of hobbyist projects for backporting it into games that didn’t have it initially.
It seems to me a lot of people do actively want it.
Maybe I’m the one out of my depth, but my understanding is that the objection is not to DLSS overall, but the specific implementation of DLSS 5 and it’s AI-driven focus. I am not claiming nobody is asking for DLSS, I’m claiming nobody is asking for the AI-driven focus of the upcoming iteration.
But to your prior point, DLSS is necessarily AI driven, it just uses the specific sections of the gpu in order to store/transform the image with local data (as far as I understand some “pre trained” info from NVDA for the specific game along with recent prior frames) to help do the upscaling. And dlss5 just takes that even further but still done locally on the device.
I do see the point that it’s clearly adjusting it to a higher extent then prior dlss versions, however since none of us actually have it in our hands to see, all we can go off of is what is being said/shown. A lot of people are basing their opinions simply off the images from the showcase but NVDA has also mentioned its extremely tunable by the developers directly (and likely by more advanced end users if the devs don’t make those tweaks available in their games) kinda like how UE has a ton of hidden tunable settings that most devs don’t make actively available but are trivial to adjust for end users if they know which file to modify.
I apologize for my passive aggressive responses, I’ve mostly been unable to have a nuanced discussion about this topic with other lemmy users and I’m kinda just taking it out on you.
Nothing exists in a vacuum; those impacts are real even if indirect.
Its a tiny fraction of the problem, the issue is with datacenter buildout taking all the available supply, if it was dlss directly related we would have seen ram prices increasing much earlier then we did.
But this specific iteration of DLSS is deeply wrapped up in the AI hype machine which is directly related to the datacenter buildout. The reason 5 has been a hot button topic centers around that AI upscaling showcase which only exists because NVidia is deeply invested in AI. This is what I mean that nothing in is a vacuum.
But you’re combining two things that should be understood separately to have an accurate picture of what is happening.
All I’ve seen about 5 is people complaining about how it looks, “not the original artistic vision” or just generally mocking it as bad. (Which is a personal opinion not a hard fact)
Nothing is in a vacuum sure, but let’s take the “personal responsibility” discussion around greenhouse emissions as an example, yes ordinary people and their actions contribute, but the vast vast vast majority is NOT and focusing on the least impactful level is not going to move the needle when the bigger issue is things like private jets, corporations ignoring environmental regulation, etc.
Because, going back to the original metaphor, nobody asked for cinnamon in the first place, so any amount wasted on it is stupid.
Weird because if you check search terms there’s a shitload of people asking “how do I enable dlss” and tons of hobbyist projects for backporting it into games that didn’t have it initially.
It seems to me a lot of people do actively want it.
Maybe I’m the one out of my depth, but my understanding is that the objection is not to DLSS overall, but the specific implementation of DLSS 5 and it’s AI-driven focus. I am not claiming nobody is asking for DLSS, I’m claiming nobody is asking for the AI-driven focus of the upcoming iteration.
But to your prior point, DLSS is necessarily AI driven, it just uses the specific sections of the gpu in order to store/transform the image with local data (as far as I understand some “pre trained” info from NVDA for the specific game along with recent prior frames) to help do the upscaling. And dlss5 just takes that even further but still done locally on the device.
I do see the point that it’s clearly adjusting it to a higher extent then prior dlss versions, however since none of us actually have it in our hands to see, all we can go off of is what is being said/shown. A lot of people are basing their opinions simply off the images from the showcase but NVDA has also mentioned its extremely tunable by the developers directly (and likely by more advanced end users if the devs don’t make those tweaks available in their games) kinda like how UE has a ton of hidden tunable settings that most devs don’t make actively available but are trivial to adjust for end users if they know which file to modify.
I apologize for my passive aggressive responses, I’ve mostly been unable to have a nuanced discussion about this topic with other lemmy users and I’m kinda just taking it out on you.