yeah but this one will actually not work as expected, shoot someone in the eye then gaslight you telling you it’s your fault that you needed to buy more subscription credits for it to kinda work better
The foundation of the system is its visual model, which Cheng trained on a custom mosquito dataset. To do that, he relied on a DSLR camera with a high-magnification zoom lens, capturing detailed images of mosquitoes for training data.
So, machine vision model. Cooler project, less clicky title.
vision models is what truly broke the barrier. neil degrasse tyson likes to teplace the word “ai” with the generic term “computing”. he made the point that before “ai” every advamcement in computing was just called computing. now we give this mystical reverence to this new tech and call it “ai”. whether you hype it or hate it, you give it yok much power.
when adobe could remove an onject from a photo we said “cool tech” now its “ai”. there really are cool fun tools buy its hard yo find people online to have a down to earth conversation about them. there are people offline i talk to but it would be cool to coolaborayr more
I actually think DLSS5 isn’t that terrible and part of the reason it was so hated was because of “AI”. I see it similar to graphic mods or Ray reconstruction (DLSS 4.5), just another tech. Don’t like it? Don’t use it. Me, on the other hand, want to actually see this happen in real-time, one day. I imagine we could even see proper filters in the future (like this).
As a technology DLSS5 feels a lot like Nvidia RTX remix where game can be modded to look very different from how it was originally intentended to look. If it can be achieved with local hardware it’s fine, if it needs internet connection, subscription and huge data-center then not so much.
yeah but this one will actually not work as expected, shoot someone in the eye then gaslight you telling you it’s your fault that you needed to buy more subscription credits for it to kinda work better
It’s not that kind of AI.
So, machine vision model. Cooler project, less clicky title.
vision models is what truly broke the barrier. neil degrasse tyson likes to teplace the word “ai” with the generic term “computing”. he made the point that before “ai” every advamcement in computing was just called computing. now we give this mystical reverence to this new tech and call it “ai”. whether you hype it or hate it, you give it yok much power.
when adobe could remove an onject from a photo we said “cool tech” now its “ai”. there really are cool fun tools buy its hard yo find people online to have a down to earth conversation about them. there are people offline i talk to but it would be cool to coolaborayr more
Its unfortunate how big tech has hacked the word AI to mostly mean these LLM based chatbots or agents.
Even when LLMs are like tiny subset of AI technologies out there.
I see what you did there, you hacker.
Meant to say hijacked but I guess hacked is close enough.
I actually think DLSS5 isn’t that terrible and part of the reason it was so hated was because of “AI”. I see it similar to graphic mods or Ray reconstruction (DLSS 4.5), just another tech. Don’t like it? Don’t use it. Me, on the other hand, want to actually see this happen in real-time, one day. I imagine we could even see proper filters in the future (like this).
As a technology DLSS5 feels a lot like Nvidia RTX remix where game can be modded to look very different from how it was originally intentended to look. If it can be achieved with local hardware it’s fine, if it needs internet connection, subscription and huge data-center then not so much.
It can be done locally with like $7k in graphics cards, last I checked.
DLSS has always been local. I’ve never heard about anything cloud-related when it comes to DLSS. The DLSS5 demo was also local.
Gross