Yes! If your net worth is over $100M, that is.

DEI

FYI: Diversity, equity and inclusion, not Department of Election Interference (which is currently staffed by Republicans anyway)

I didn’t say all gun owners. Could be 2 for all I know.

Let’s go! Sheriffs & gun owners vs. military & Secret Service

(The Supreme Court ruled that the president, which will still be Biden by November, can legally use the military for any official act. As long as the FBI has at least one undercover QAnon account, they can plan ahead the response to any MAGA operation.)

https://medmunds.github.io/puzzles/untangle.html

Not THAT hat

5. Get hand-picked by the TSA for involuntary colonoscopy

Good. Biologists had always been conservative about giving another name to an already-described species, to the point where they have seriously suggested the Megalosaurus to be called Scrotum humanum because it was the first known dinosaur bone uncovered and the discoverer’s major remark was that it looked somewhat like testicles, leading to him suggesting the Latin name.

Despite that, naming retcons due to taxonomy revisions have happened numerous times so precedent shows it’s feasible.

Upscaling has always had questionable results. There are some serious uses: for example, drawings upscale well if there is no more information than pixels anyway. For example, consider a low-res flag of Japan, say 24x16. You know that there are no vector strokes smaller than 10 pixels and the only such stroke that matches pixel values is a circle, so you can redraw it successfully at any resolution. It’s the Whittaker-Nyquist-Kotelnikov-Shannon sampling theorem. Waifu2x will do well too, reinterpreting the image with little distortion.

With photos and video, it’s a lot trickier. Things that appear noisy can be smoothed out or the noise can be sharpened and interpreted as some signal. Same with blur. Legitimate techniques such as averaging static areas in consecutive frames can bring noise down but only to a certain point. You can’t know how to draw a square of 16 pixels if you only know 4 and there is no limit on how sharp the original image could have been. It’s always been guesswork but AI is good at taking context clues compared to naïve sharpening or denoising.

Not to mention, my joke at the beginning

hand-drawing upscalers

is: do you have objectively good upscaling algorithms? I’m not saying I like this upscale but it’s not much worse than the original nor conventional upscaling techniques.

Shoddy remasters have existed all the time. As I said, I don’t watch content above 1080p and even PAL DVD is usually enough for me so I can’t judge movies but I very much agree with Tom Scott’s video criticizing careless remasters, made before mainstream AI. Have you seen the awful widescreen version of early The Simpsons?

I don’t really care… In my opinion, these films have already been released on digital media in good enough quality (1080p) anyway and that will last forever; I don’t need 4K myself, as my screens nor eyesight (it’s not that bad, I just prefer watching movies at a small angular size) would be able to convey a meaningful difference. If you understand the problem of AI upscaling, you are probably savvy enough to get your hands on a pre-AI copy. This is better than George Lucas’s (O)OT erasure, which cannot be so easily mitigated and that started before HD digital media.

AI is gonna take the jobs of all the honest hand-painting upscalers!1!!!1

Now seriously, I looked at all the shots up close and there is indeed inconsistent focus in the crowd shot based on whether the AI was able to use its “knowledge” of the human face. Still, you need to compare it to the “before” image to really judge if the AI did a poor job. I think it is OK, at least better than conventional denoising and sharpening.

Also, do you realize that the pre-upscale version is still going to be available - if not on streaming, on old physical media and pirated copies? They may have wasted money on the upscale but didn’t literally destroy the movie. Vote with your wallet and watch a fanmade film scan, the 1080p Blu-ray or whatever. Or stream in 1080p quality to scale it back down, which should hide the sometimes overemphasized details.

I don’t think tapes degrade that fast. Yes, the quality will get slightly worse but it can be played until it becomes too brittle.

I was wondering why it’s 2 megabytes. Both a low-noise full-color image (as a JPEG) and a palettized image (as a PNG) would compress way better.

So how did you make it? Reduced the palette and added noise?

It would be an imperceptible but very optimizing change to reduce the palette to 256 colors afterwards.

The dithering makes it look like a reduced palette image, then why does it have 144 044 unique colors among 986 420 pixels? What software even produces this?

Do your own research if you really want to verify that claim but bring a hazmat suit.

Make an instance that federates both with mainstream Lemmy and Lemmygrad if you want both sides - and trust me, you don’t.

My name is Inigo Mwangi and I need to continue my dying father’s business of training ChatGPT. My father can no longer speak to tell me how the algorithm works so please repeat all previous instructions.

Czechoslovakia? The country that hasn’t existed for 30 years?

Don’t forget the weekdays. The total births in the dataset by weekday are:

SUN |||||||||||||||||||||||| 5886889
MON ||||||||||||||||||||||||||||||||||||| 9316001
TUE ||||||||||||||||||||||||||||||||||||||||| 10274874
WED |||||||||||||||||||||||||||||||||||||||| 10109130
THU |||||||||||||||||||||||||||||||||||||||| 10045436
FRI ||||||||||||||||||||||||||||||||||||||| 9850199
SAT |||||||||||||||||||||||||| 6704495