@sy0ungg@Cary_Bleasdale People in everyday life don't have hate for AI at all, I don't know what planet you're living on. I've seen people 55+ use ChatGPT for daily things in their lives.
@Cloydbrok@JoystickOW They don't like their favourite mode being the least popular so they want to force everyone to play even if the majority does not prefer it.
Put about 10 hours into it and felt like I had to drag myself to play it every time, entire gameplay loop felt pointless. I go in, get loot, get out just to do it again for no reason? There's no win condition in a match like in a BR to strive for and the moment to moment gameplay was boring just scavenging around.
Don't know why Marathon players have a hard time accepting that the majority of people do not like the game and that doesn't mean their "taste" is better.
A lot of R&D efforts are spent to get a decent product out at a price point that consumers are willing to pay, Meta could develop the same product tomorrow but they know that cramming that much costly tech into a product would result in no sales.
Find it hard to believe they met any sort of internal milestones in regards to sales considering they laid off more than 100 on the Vision team, alongside putting future headsets on indefinite hold. Unless you have some internal rep at Apple feeding you information which they haven't reported anywhere.
@ShamirahZ@FervorGamingYT@Waydots Doesn't matter if it's good for a few months, it's a new game, new engine, all new content. Not just WoW trying to recapture all the players that have quit, again.
@stefjh@MoroseJerk Only ever there comutting to and from work so it’s just people getting their food and coffee. Melbourne Central is awful to be during weekends.
Wuchang with DLSS 5, off/on. Easily my favourite game from 2025 and brings back the long lost feeling of playing Dark Souls 1 for the first time again. Played through it 3 times back at launch. Very underrated title. #Wuchang, #DLSS5
@floptinaboots@RicharD1ckJohn Yeah growing up with two sisters I would always walk around in boxers and sometimes without a shirt in summer and they were the same in a shirt and undies, never once thought it was strange or anything off about it.
Just bait the dart, now just bait the suzu, now just the teleport, now just bait the hook. Great, now you have a 2 second window to maybe do something. A Kiri can make 85 mistakes, a Dmon can play with her brain off entirely, cat can do anything on the planet but a DPS has to play perfectly.
On reading futher into it and reading Nvidia's whitepaper on DLSS 5 (barely a whitepaper), you're right. I was under the impression that Nvidia's DLSS 5 was sourcing more data than just motion vectors. They do during training to infer what to apply where, but that only achieves one result which is pure photorealism.
They are attempting to do different things with different methods, Nvidia's DLSS 5 is meant to be used alongside pathtracing as we no way to cast thousands of rays per pixel in real time to achieve photoreal results while AMD is trying to cheapen path tracing altogether. I'd like to see AMD's implementation outside of a research demonstration though as to whether it could ever be viable, I feel we'll see their radiance caching first which will be used in conjunction with this version of neural rendering rather than the paper demonstrates, I also expect them to ship their own version of DLSS 5 in the same way.
DLSS 5 is able to achieve photo real results due to path tracing, the off-screen lighting is already in that colour buffer so it can enhance it without developers feeding it additional information.
I wouldn't consider DLSS 5 to be sloppy but I know a lot of creatives who dislike it, a lot of user implementation is (especially those abhorrent multipasses) but when applied correctly it can greatly improve the lighting. It's not 100% accurate nor is it entirely artist driven but I've worked with a lot of lighting artists and what they can do is limited by their frame budget, or $$ limitations from us as how we rendered hair, the way light interacts with hair, skin, etc. As someone who was never on the creative side, being a rendering engineer and all, I will always appreciate tech that can push what's possible now that we've hit limits on hardware advancements.
You’re right in the sense that it sits earlier in the process (and replaces a step) but the end result is still generated image slapped onto the frame. Effectively the same end result with less performance cost. Outside of a research paper on a carefully picked frame it will hallucinate as it has no real data to work with like DLSS 5 apart from a noisy one pass GI.
@FemtoSpirit@esiuol32@grok That was a personal comment by a member of staff after working with Kojima for many years, not "they" (the company) saying anything.