ricecake

joined 2 years ago
[–] ricecake@sh.itjust.works 2 points 3 hours ago (1 children)

Just for more clarity: they workshoped for ideas on how to improve clarity and accessibility from some editors at an event. They did some small experiments, and they then developed a plan to trial some of them and presented the plan to a wider audience for feedback. After they got feedback they decided not to.

It's not quite the editors pushing back on Wikipedia. Or rather, it's not the "rebellion" people want to make it out to be.

https://www.mediawiki.org/wiki/Readers/2024_Reader_and_Donor_Experiences/Content_Discovery/Wikimania_2024,_%22Written_by_AI%22_How_do_editors_and_machines_collaborate_to_create_content

https://www.mediawiki.org/wiki/Reading/Web/Content_Discovery_Experiments/Simple_Article_Summaries

It rubs me the wrong way when the process going how it should go gets cast as controversial and dramatic. Asking the community if you should do something and listening to them is how it's supposed to go. It's not resistance, it's all of them being on the same team and talking.

[–] ricecake@sh.itjust.works 1 points 3 hours ago

Eh, that's not quite original research. There are plenty of other examples of images and sound files created for Wikipedia. A representative example isn't research, it's just indicating what something is.

The Wikipedia article on AI slop and generative AI has a few instances of content that's representative to illustrate a sourced statement, as opposed to being evidence or something.

It's similar to the various charts and animations.

[–] ricecake@sh.itjust.works 1 points 4 hours ago

Yes, but...

https://en.wikipedia.org/wiki/Wikipedia%3ADatabase_download

That's because viewing the page uses server resources, as done API access. If you want the data you can download the database directly.

[–] ricecake@sh.itjust.works 1 points 1 day ago

I don't love an abstract legal identity. I'm capable of being happy with institutions, the culture composed if the people living there, and adoring the natural splendor.
Right now I'm actively angry at the institutions, a huge number of people have taken a sharp turn towards fascism, and I've got no problems with the forest still.
Me and the forest are cool, and that's part of why I'm mad at the institutions.

I have no desire to live in the forest because, if nothing else, that's not good for the forest. Then the people who opted to live there became insane, and decided to largely gut all of the institutions, and make it easier to destroy the forest.

"I live in a state of natural splendor, and I'm willing to fight to let you cut it down, splash me with mercury , and blot out the sun with smoke because I don't have healthcare and fuck you for asking. It's the refugees who are the problem".

[–] ricecake@sh.itjust.works 42 points 1 day ago (17 children)

Oh my God, no. "Great place to live if you want you all your neighbors to be frighteningly conservative, the closest store to be a 30 minute drive, and the nearest hospital to be an hour away and shutting down because their public funding got cut".

[–] ricecake@sh.itjust.works 2 points 1 week ago

And you didn't even read past the first sentence I see.

Saying they're the same because they both use a neural network is roughly equivalent to saying things are they same because they're both manipulating kinetic energy.

[–] ricecake@sh.itjust.works 2 points 1 week ago (2 children)

... How if flying a spaceship different from driving a car? They're both controlled applications of kinetic energy to move people or objects.

At the end of the day, it's all a pile of transistors and the only thing that is of import is the intent behind usage.

In one case it's saying you can use a neural net to take something rendered at resolution A/4 and make it visually indistinguishable from the same render at resolution A.
The other is rendering something and radically changing the artistic or visual style.

Upsampling can be replicated within some margin by lowering framerate and letting the GPU work longer on each frame. It strives to restore detail left out from working quicker by guessing.
You cannot turn this feature off and get similar results by lowering the frame rate. It aims to add detail that was never present by guessing.

Upsampling methods have been produced that don't use neural networks. The differences in behavior are in the realm of efficiency, and in many cases you would be hard pressed to tell which is which. The neural network is an implementation detail.
In the other case, the changes are more broad than can be captured by non AI techniques easily. The generative capabilities are central to the feature.

Process matters, but zooming out too far makes everything identical, and the intent matters too. "I want to see your art better" as opposed to "I want to make your art better".

[–] ricecake@sh.itjust.works 2 points 2 weeks ago

There's hardware required to shunt the display out the USB port and since it's not a super in demand feature they usually don't implement it. As such the software for looking nice while doing it isn't as developed.

But yes, it's been in developer settings for years, and was usable if your hardware supported it.

[–] ricecake@sh.itjust.works 5 points 2 weeks ago

Yes. And now it's native in all android! Samsung helped make it!

It's good when things get better.

[–] ricecake@sh.itjust.works 2 points 3 weeks ago

Yeah, the conventional ones still draw a good chunk of power, and they're not clean but they're not dirty. Same as how a grocery store isn't good for the environment but you're not looking at them first for places to clean.

They tend to be boring, and are usually not a public thing but just something owned by a company to house their computers. The only reason I know about the ones near me is I used to work at one and people would move jobs to or from other ones. (As an aside, a datacenter is a great place to nap if you like white noise).

For a sense of scale:

This is the site of an open AI data center. The yellow square is about 1 square mile and mostly encompasses the area they plan to/have filled.

That angle shows more build out.

This photo has two normal data centers in it. The yellow square is also about 1 square mile. I've highlighted the data centers in red. One is to the left of the square near the middle, and the other is down from the right side near the big piles of what looks like rocks. (Spoilers: it's rocks. They make asphalt). The sprawling complex in the upper right is a refrigerated grocery store distribution complex. The middle on the other side of the block from the asphalt is a coal power plant.

Of the things in this picture, I'm most upset about the giant freeway interchange. Coal is shit, but it's a modern plant so it's not belching soot, just co2, and the utility is phasing it out anyway. The grocery traffic is mostly dead except between the hours of midnight and 7am when they do restocks.
I can hear the freeway if I go outside.

[–] ricecake@sh.itjust.works 2 points 3 weeks ago (2 children)

I think the part you're missing is that 1) it's my community too 2) they're not talking about AI data centers, or new data centers or anything like that, they're petitioning to ban all data centers, and 3) we have multiple data centers in the city already that no one complained about until AI data centers became a thing people felt concerned about.

There's a major difference between the 2 square mile hyper scale AI data center that requires a nuclear reactor and a full water treatment plant to cool and the 2 acre data center that's air cooled and has no more ground pollution than any other parking lot and essentially a warehouse.
The state government has two in the city, at least, for processing electronic tax records, applications and hosting service sites. We have a few national insurance companies that need to process all the things they process. A research university, and a web hosting company round out the list of ones I know about.

This is my entire point about why sometimes it's really necessary to point out that what someone is referring to is only a small part of what the words they're using describe. The language being imprecise doesn't matter until someone proposes a law outlawing chemicals, shuttering all data centers, or banning AI.

LLMs are problematic. My fancy rice maker isn't.

[–] ricecake@sh.itjust.works 1 points 3 weeks ago (4 children)

I take your point. :)

It's worth mentioning in my opinion though, because if someone were to say "we should ban chemicals" it'd be worthwhile to point out what that actually means.

I don't actually think the broadness of the category is intentionally abused, it's just that it's an incredibly common thing to remove anything from the AI category that's explicable.

I feel slightly more hanlons razor about it since there's people in my city talking about and petitioning on the popular notion of banning all data centers from the state, and how it would be awful if s data center came here. I know what they mean, but it's not what they're trying to get the law to do, and our city already has six data centers I know of off the top of my head. The language drift is fine, but when it starts to conflate with policy it's another issue.

75
submitted 2 years ago* (last edited 2 years ago) by ricecake@sh.itjust.works to c/imageai@sh.itjust.works
 

Been having fun trying to generate images that look like "good" CGI, but broken somehow in a more realistic looking way.

 

Made with the Krita AI generation plugin.

 

digital illustration of a male character in bright and saturated colors with playful and fun expression, created in 2D style, perfect for social media sharing. Rendered in high-resolution 10-megapixel 2K resolution with a cel-shaded comic book style , paisley Steps: 50, Sampler: Heun, CFG scale: 13, Seed: 1649780875, Size: 768x768, Model hash: 99fd5c4b6f, Model: seekArtMEGA_mega20, ControlNet Enabled: True, ControlNet Preprocessor: lineart_coarse, ControlNet Model: control_v11p_sd15_lineart [43d4be0d], ControlNet Weight: 1, ControlNet Starting Step: 0, ControlNet Ending Step: 1, ControlNet Resize Mode: Crop and Resize, ControlNet Pixel Perfect: True, ControlNet Control Mode: Balanced, ControlNet Preprocessor Parameters: "(512, 64, 64)"

If you take a picture of yourself in from the shoulders up, like in the picture, while standing in front of a blank but lightly textured wall it seems to work best.

view more: next ›