dgrdsv
If you are going to continue to deny or obfuscate facts after evidence has already been presented to you (and even go as far as to suggest that a technology company is mistaken about how their own technology works) I don't think there can be any meaningful progress made in this conversation.
I won't bother wasting anymore of my time on it.
EDIT:
And to be clear, based on our conversation, you don't seem to understand how back propagation works. While it is true that DLSS 2.0 added temporal accumulation to improve results, the reference targets are still 64xSS images (they're just not game-specific anymore). The point of using the reference targets is not just about AA, it's about providing an ideal result for AI to work towards resolving. It has a long way to go still in getting there when the input isn't native 4k, but it allows the researchers to adjust weights based on how well it did. This is called back propagation and applies regardless of whether or not temporally accumulated data is being used or not.
And when we're talking about high quality super sampling, the AA comes from rendering at a higher resolution and then downsampling. The goal of DLSS 2X is to mimic that quality. You can't call it only an AA solution when it has to use a higher resolution to get to that AA solution.
And finally, when one technology adds pixel detail and another upscales pixel detail, they're not solving the same problem with different solutions, they're solving different problems.
With all of that out of the way, I'm done with the conversation.