Deep Learning Super Sampling (DLSS), an AI-based upscaling technology that NVIDIA introduced years ago but has now become synonymous with many triple-A game releases, including, most recently, FINAL FANTASY XVI, Dead Rising Deluxe Remaster, and God of War Ragnarök, may be on its way to becoming a mandatory feature for video games, players say after having heard the latest thoughts from Jensen Huang, President and CEO of NVIDIA. Huang, who co-founded green team 31 years ago, in 1993, told visitors at the Goldman Sachs Communacopia & Technology Conference in San Francisco last week that NVIDIA can’t even “do computer graphics anymore” without artificial intelligence, a field that continues to gain prominence in the tech industry for advances that include video generation tools and chatbots.
Huang told participants at the event:
- “…in our company, we use it for computer graphics.”
- “We can’t do computer graphics anymore without artificial intelligence.”
- “We compute one pixel, we infer the other 32. I mean, it’s incredible.”
- “And so we hallucinate, if you will, the other 32, and it looks temporally stable, it looks photorealistic, and the image quality is incredible, the performance is incredible…”
Some recent DLSS promos from NVIDIA:
NVIDIA on DLSS 3 and one of key components, Frame Generation:
DLSS is a revolutionary breakthrough in AI graphics that multiplies performance. Powered by the new fourth-gen Tensor Cores and Optical Flow Accelerator on GeForce RTX 40 Series GPUs, DLSS 3 uses AI to create additional frames and improve image quality.
Boosts performance by using AI to generate more frames while maintaining great responsiveness with NVIDIA Reflex. DLSS analyzes sequential frames and motion data from the new Optical Flow Accelerator in GeForce RTX 40 Series GPUs to create additional high quality frames.
Join the discussion in The FPS Review Forums...
Discussion (6 replies)
Join Discussion →
Fuck this.
Fuck all of it.
I mean, maybe someday the models will be able to infer at 100% accuracy, but my playing with DLSS on my 3080ti says it’s not there right now. I’d rather have higher end hardware generating the full frame, then go with lower end hardware + artifacts, even if it’s the same FPS.
If the AI generation made things more affordable I'd be all for it, but it seems to be going in the opposite direction
That's how you get the city and the stars
Huang sounds like he was high during that presentation.
- We compute one pixel, we infer the other 32. I mean, it’s incredible.”
- “And so we hallucinate, if you will, the other 32, and it looks temporally stable, it looks photorealistic, and the image quality is incredible, the performance is incredible…”
DO NOT WANT
Hallucinating, indeed
I miss computed pixels already :(
"Brian_B, post: 89498, member: 96" wrote:If the AI generation made things more affordable I'd be all for it, but it seems to be going in the opposite direction
It definitely done wonders for NV and Huang's profits, that's for sure.
Discussion (6 replies)
Join Discussion →Fuck this.
Fuck all of it.
I mean, maybe someday the models will be able to infer at 100% accuracy, but my playing with DLSS on my 3080ti says it’s not there right now. I’d rather have higher end hardware generating the full frame, then go with lower end hardware + artifacts, even if it’s the same FPS.
If the AI generation made things more affordable I'd be all for it, but it seems to be going in the opposite direction
That's how you get the city and the stars
Huang sounds like he was high during that presentation.
DO NOT WANT
Hallucinating, indeed
I miss computed pixels already :(
It definitely done wonders for NV and Huang's profits, that's for sure.