tygertyger413

[tygertyger413]

doodls

[tygertyger413]

[tygertyger413]

hey uh new type of ao3 spam comment just dropped. (I know it's spam because the fic they left this comment on . doesn't have chapters. lmfao). Report this kinda comment as spam and don't take it personally it is literally recycled bullshit

[tygertyger413]

[tygertyger413]

“Babe are going to bed? How did you get under there???”

“Mreh!” >:(

“Oh, I’m sorry! G’night!”

[tygertyger413]

Donphan -- GOSSAN

[tygertyger413]

Source

Video of Tama

Follow Ultrafacts for more facts

[tygertyger413]

every day i kinda wanna cancel the show

[tygertyger413]

You’ve probably been asked this before, but do you have a specific view on ai-generated art. I’m doing a school project on artificial intelligence and if it’s okay, i would like to cite you

I mean, you're welcome to cite me if you like. I recently wrote a post under a reblog about AI, and I did a video about it a while back, before the full scale of AI hype had really started rolling over the Internet - I don't 100% agree with all my arguments from that video anymore, but you can cite it if you please.

In short, I think generative AI art

Is art, real art, and it's silly to argue otherwise, the question is what KIND of art it is and what that art DOES in the world. Generally, it is boring and bland art which makes the world a more stressful, unpleasant and miserable place to be.

AI generated art is structurally and inherently limited by its nature. It is by necessity averages generated from data-sets, and so it inherits EVERY bias of its training data and EVERY bias of its training data validators and creators. It naturally tends towards the lowest common denominator in all areas, and it is structurally biased towards reinforcing and reaffirming the status quo of everything it is turned to.

It tends to be all surface, no substance. As in, it carries the superficial aesthetic of very high-quality rendering, but only insofar as it reproduces whatever signifiers of "quality" are most prized in its weighted training data. It cannot understand the structures and principles of what it is creating. Ask it for a horse and it does not know what a "horse" is, all it knows is what parts of it training data are tagged as "horse" and which general data patterns are likely to lead an observer to identify its output also as "horse." People sometimes describe this limitation as "a lack of soul" but it's perhaps more useful to think of it as a lack of comprehension.

Due to this lack of comprehension, AI art cannot communicate anything - or rather, the output tends to attempt to communicate everything, at random, all at once, and it's the visual equivalent of a kind of white noise. It lacks focus.

Human operators of AI generative tools can imbue communicative meaning into the outputs, and whip the models towards some sort of focus, because humans can do that with literally anything they turn their directed attention towards. Human beings can make art with paint spatters and bits of gum stuck under tennis shoes, of course a dedicated human putting tons of time into a process of trial and error can produce something meaningful with genAI tools.

The nature of genAI as a tool of creation is uniquely limited and uniquely constrained, a genAI tool can only ever output some mixture of whatever is in its training data (and what's in its training data is biased by the data that its creators valued enough to include), and it can only ever output that mixture according to the weights and biases of its programming and data set, which is fully within the control of whoever created the tool in the first place. Consequently, genAI is a tool whose full creative capacity is always, always, always going to be owned by corporations, the only entities with the resources and capacity to produce the most powerful models. And those models, thus, will always only create according to corporate interest. An individual human can use a pencil to draw whatever the hell they want, but an individual human can never use Midjourney to create anything except that which Midjourney allows them to create. GenAI art is thus limited not only by its mathematical tendency to bias the lowest common denominator, but also by an ideological bias inherited from whoever holds the leash on its creation. The necessary decision of which data gets included in a training set vs which data gets left out will, always and forever, impose de facto censorship on what a model is capable of expressing, and the power to make that decision is never in the hands of the artist attempting to use the tool.

tl;dr genAI art has a tendency to produce ideologically limited and intrinsically censored outputs, while defaulting to lowest common denominators that reproduce and reinforce status quos.

... on top of which its promulgation is an explicit plot by oligarchic industry to drive millions of people deeper into poverty and collapse wages in order to further concentrate wealth in the hands of the 0.01%. But that's just a bonus reason to dislike it.

[tygertyger413]

People have written a lot of touchy-feely pieces on this subject but I thought I’d get right to the heart of the matter

[tygertyger413]

when the mutuals are infodumping about an interest i know nothing about

[tygertyger413]

homesick

[tygertyger413]

Magical girl Annette ✨🪓✨

[tygertyger413]

Colored this sweet DimiDue piece for @/fireboxemblem on Twitter/X! 💙🩵✨

[tygertyger413]

a familiar face from long ago

[tygertyger413]

warm-up

[tygertyger413]

a quick post-timeskip sylvain for a friend 💕