I think, the problem is that Nvidia has two customer groups. Those that buy their products and those that buy their stock options. Nvidia can produce garbage that completely misses the point of real-world usage, so long as they can convince investors that other investors will join the pyramid scheme. And for that, it just has to look like impressive tech, not actually good or artistically meaningful.
Will it smooth out a wall that is supposed to look like it can be destroyed?
Yeah, at the very least, it will throw a whole bunch of details into the general area, which will make it harder to tell what’s interactable.
We’ve had photorealistic games before, by taking literal photographs and using those as point-and-click levels. You practically don’t see that anymore these days, because not being able to tell what’s interactable was a major weakness.
Doesn’t mean that DLSS 5 or the like will strictly have the same problem, but it certainly feels like these companies are trying to throw in photorealism again, with no regards for the cost.
I mean, yeah, but you’re kind of saying what the others here were saying, too, in that when something fits the anywhere close to the “old hag” category, that the probabilities will shove it entirely towards “old hag”.
That it’s somewhat fitting for this character, I would expect to be coincidence. Like, maybe they did actually give the image generator somewhat of a system prompt for this demo, that it should make her look extra wrinkly.
But yeah, shoving all depictions of women either towards young model or old hag is quite emblematic of these image generators, so personally, I don’t think, it was even necessary…
Don’t think the playable characters are supposed to be creepy, but yeah, a yassified zombie would probably fit right in. 😅
Never cared for realism in games to begin with, so don’t particularly care to comment on how it looks, but I’ve been thinking that I genuinely find it creepy.
Not just Uncanny Valley material, but also having these faces stare at you, always fully lit, it just gives me the creeps, kind of like a panopticon situation.
I don’t fucking know, if that’s my own trauma playing into that, where for the longest time, people looking at me generally meant they’re about to bully me.
But either way, I’m about to head to bed and genuinely feel like there’s a 20% chance I’ll have a mild nightmare from that shit.
This whole AI craze has been a wild ride of all kinds of nightmare fuel, from depictions with missing/additional limbs to the weirdest warping of objects+limbs in those fucking generated videos. And the worst part is that some folks seem to just not see it or not want to see it, so they keep using the nightmare generators.
Yeah, there’s a bias in these image generator AIs, because model photographs are:
I think, why people perceive that right image as older is because the wrinkles carve much deeper. Yeah, same number of wrinkles, but it looks like they got chiseled down by 20 additional years…
So, uh, if you’re already using jq, this should also format JSON:
echo "{\"key\":\"value\"}" | jq
😅
Was recently thinking this might happen to Pinterest, too. Their webpage was never great, with how it tried to prevent you from downloading images, when that was literally the only reason I would ever visit. But at least, they did have a big database of images and a decent algorithm for detecting visual similarity.
And well, they have an even bigger database of images now, but the majority of it is not worth looking at, because the images are not real. I don’t bother visiting anymore, because you can’t find anything worthwhile on there anymore.
They did announce going all-in on AI at some point, but I don’t know, if they actually decided to generate images themselves. That seems almost too stupid.
Could be that they have some financial incentives for folks posting and that alone lead to tons of AI-generated uploads. I don’t actually know how Pinterest was supposed to work…
Yeah, management wants us to use AI at $DAYJOB and one of the strategies we’ve considered for lessening its negative impact on productivity, is to always put generated code into an entirely separate commit.
Because it will guess design decisions at random while generating, and you want to know afterwards whether a design decision was made by the randomizer or by something with intelligence. Much like you want to know whether a design decision was made by the senior (then you should think twice about overriding this decision) or by the intern that knows none of the project context.
We haven’t actually started doing these separate commits, because it’s cumbersome in other ways, but yeah, deliberately obfuscating whether the randomizer was involved, that robs you of that information even more.
Yeah, that’s my biggest worry. I always have to hold colleagues to the basics of programming standards as soon as they start using AI for a task, since it is easier to generate a second implementation of something we already have in the codebase, rather than extending the existing implementation.
But that was pretty much always true. We still did not slap another implementation onto the side, because it’s horrible for maintenance, as you now need to always adjust two (or more) implementations when requirements change.
And it’s horrible for debugging problems, because parts of the codebase will then behave subtly different from other parts. This also means usability is worse, as users expect consistency.
And the worst part is that they don’t even have an answer to those concerns. They know that it’s going to bite us into the ass in the near future. They’re on a sugar high, because adding features is quick, while looking away from the codebase getting incredibly fat just as quickly.
And when it comes to actually maintaining that generated code, they’ll be the hardest to motivate, because that isn’t as fun as just slapping a feature onto the side, nor do they feel responsible for the code, because they don’t know any better how it actually works. Nevermind that they’re also less sharp in general, because they’ve outsourced thinking.
Probably depends a lot on whether it’s an anonymous survey…
Well, bullet bras were a thing for quite a while, too:
I feel like folks are antsy, because it doesn’t look revolutionarily different from the original Steam Machines, which flopped back then.
But yeah, there actually is a revolutionary difference, which is that the vast majority of games now do run on the new Steam Machines. This has also already been proven to the public with the Steam Deck.
And I do think there is a market of folks looking at a Steam Deck and thinking they don’t need it to be a handheld and would rather have more bang for their buck.
I guess, we will have to see, but I could also imagine the cornered memory market playing into the hands of the Steam Machine, as the better memory efficiency of Linux will let you do more for less.
Oh man, I don’t want to get deep into all the politics involved, but man, this reads like complete non-sense:
The outage comes following Iranian attacks on the UAE as retaliation for US and Israeli strikes on Iran.
If they did specifically target US corporations in UAE, that would make some amount of sense as direct retaliation.
I guess, you can also attack UAE and hope that they pressure the US to stop invading.
But in any case, this seems like a really good way to drag more nations into the conflict, or at least to force them to become active, which is not in the interest of Iran.
One thing that will become important pretty quick if you continue making these scripts is that it’s almost always better to wrap your variables in quotes - so it becomes
yt-dlp -x “$a”.
Oh man, this reminds me of the joke that any program that’s more complex than Hello World has bugs – and folks still don’t even agree how to spell “Hello, World!”.
Of course, Bash is a particular minefield in this regard…
To be fair, Sarah Bond was in a similar position and is a woman of color, too.
Yet more suggestions which you didn’t ask for:
GitNote might be up your alley. Its UI is not as slick as QuillPad, nor does it have as many features, but it does faithfully keep the folder structure.
Somewhat of a deal-breaker for me personally: It can’t do reminders.
Embrace the chaos. 😅 I have a little program for managing my notes on desktop and it just dumps them all in one folder, too. If I need to find something again, I’ve got a little text search, which is basically the equivalent of grep -iR. I just make sure to mention enough keywords in each note, so that I can find it again.
Personally, I much prefer this workflow, because you can start typing (and hitting Ctrl+S) and then later ensure that it has all the right metadata, rather than having to select a folder upfront where it will be saved.
I actually tried QuillPad not too long ago and couldn’t make it work for me, but being able to save as normal files seems to be a recent addition, so maybe that’s what I was missing.
I’ll probably try setting it up to work with my desktop note system then…
Well, Kate would be the obvious choice for KDE…
Dungeon Lawl Stone Soup 🙃