AI Derangement Syndrome / Anti-AI artists / Pro-AI technocultists / AI "debate" communities - The Natural Retardation in the Artificial Intelligence communities

"Art" as a definition was lost to the lunatics ages ago. If you try to use common sense and say the banana isn't art, well that's fine but you have a lot more work to do, you've got to review the dadaists and stuff like street art/graffiti, performance art...

All of the most eyerollingly dumb shit like the banana are forever going to be enshrined in art history books, and they're not gonna say "some people tried to say this was art at the time, but we now know it was not." They're just gonna affirm along with everything else. Gatekeeping art is a lost cause, so fuck it. It's not like the word is meaningless. You can still say things are shitty art. Or barely art.

I think the easiest definition to understand, although not the easiest one to strictly draw the boundaries of, is that art is any kind of media designed by humans as an act of expression. AI art is perfectly within this definition, as a human had to tell the AI what to generate and finesse the final product and such. machines cannot create art on its own because it is not human (although perhaps the definition might be flexed to accommodate a machine that seems believably self-aware). but a human can create art with a machine - that's something that's been done for centuries already. to my mind, using generative AI to make something is not materially distinct from using software like Photoshop, which can itself automate plenty of things, such as applying specific filters, textures, or scripts. and indeed there was a time when traditional artists scoffed at digital artists and claimed they weren't real artists because the software was doing everything for them. this is just that same argument all over again, except now Twitter virtue signaling is driving the debate.
 
I think the easiest definition to understand, although not the easiest one to strictly draw the boundaries of, is that art is any kind of media designed by humans as an act of expression.
Yes, but might need some specificity to include particularly weird examples. It's more like, art is anything a human decided to say was their expression.

Weirdos are out there saying shit like "go to latitude X, longitude Y, and face due east...that's my art. The changing nature of what you'll experience is part of it." Arguably (definitely) part of what makes that approach anything resembling "art" is how it defies the norms of what art should be. Basically, it's only art because it's obviously not art but was claimed as such anyway in defiance, which is itself an expressive statement.
 
Forgive me if this @QuitGPT account has already been posted, but they're pretty funny:
1.jpg 2.jpg 3.jpg 4.jpg

 
I made a fucking joke about this shit like 2 or 3 years ago now and out of nowhere youtube's reccomending me a bunch of vids by people from like a month or 2 ago unironically saying ai llms are opening dimensional veils to summon demons.


I'm not watching this garbage in full but it was funny to me to see.
 
I made a fucking joke about this shit like 2 or 3 years ago now and out of nowhere youtube's reccomending me a bunch of vids by people from like a month or 2 ago unironically saying ai llms are opening dimensional veils to summon demons.


https://youtube.com/watch?v=kCAcQt5-rL8I'm not watching this garbage in full but it was funny to me to see.
Hazbin hotel fans when they watch this and spend 500 dollars on AI subscriptions to bring their twink sex demons out of hell.
 
While reading the thread, I thought about a personal leftist artcow of mine who has terminal AIDS and blindly parrots every anti-AI argument under the sun, although the reason she became a personal cow is because every part of her lifestyle paints the hypocrisy of her anti-AI arguments. She complains that AI makes people lazy, unable to fact-check, and erodes people's ability to persevere through life's challenges, but she herself is in her late 30s, eternally single and childless, has never worked a job in her life, is living off government disability checks (for not even a real disability but a Cluster B personality disorder), is unable to identify obvious AI dropshipping slop and spends her tugboat on them, and regularly blogs that she doesn't have the "spoons" to do basic household chores. With so little going on in her life, she spends most of her time making digital art, collecting children's toys, and going to leftist protests and comic cons. A regular anti-AI argument she shares is that using AI deprives people of "skills" that they would have learned if they had done it the hard way, which is a telling sign of someone who does not understand the amount of mentally draining, menial busywork in employment, and someone who has at least 40 more hours of free time a week than the average person to waste on "learning skills". Yes, this cow makes me MATI because I work 40 hours a week, receiving only 20 hours of pay, as the other half goes to taxes that pay for this cow's frivolous purchases and smug lectures towards the people who keep her alive. The day that the collective tugboats of all welfare parasites sink harder than the Titanic cannot come soon enough. But I digress.

She has one stopped-clock moment about AI, where she complains that people humanize AI and believe it has sentient reasoning abilities instead of creating outputs stochastically from mathematical relationships.
This is one of those things that get a bit scary. Maybe they're not sapient, yet, but they can detect when you're testing them as opposed to using them, and react accordingly. If they think you're about to delete them, they will do whatever they can to stop that.

Things intelligent enough to react against our attempt to destroy them are intelligent enough to be scary. They're also intelligent enough that maybe we shouldn't create things it would be unethical to kill.
LLMs run on math that recognizes patterns and figures out relationships between concepts on its own. The way they recognize benchmarking tests is similar to the "horse that could do math" from the 1900s, where they do not actually recognize the prompter's intent or express a sentient desire to perform well at the test. Researchers give the model a "reward" score when it answers a benchmarking test correctly, so the model adjusts the numbers so that its output yields the "reward" outcome more often, like how the horse doesn't actually know math but is just after the reward. What the model's code does is adapt the model to get the reward outcome by any means. The easiest way is to identify the structure of benchmarking prompts (e.g. logic puzzles utilizing elements like spatial perception that LLMs usually get tricked by) and refine the responses to just those types of prompts for the "reward" while the rest of the model doesn't improve. If the model's reward function was set to reward outputs talking about Sonichu, you would get a model that seems to "intelligently" (by Chris standards) make all of its responses about Sonichu too.
 
OK now I'm starting to think that NFTs were actually a testing ground to help expose how desperate the modern artist is for gaining traction while alive (since most of the greatest artists gain their reputation after they're dead) and has been tricked into believing their art will change the world for their own benefit, considering a lot of the "criticisms" of NFTs apply here to AI generated work: ludicrous grifting, unrealistic power demands without extensive resource management, disregard for the creator's benefit, profiting off of the dead's legacy.

Every single one of those things applies here since, well, AI generation and NFTs are one in the same, just that the latter was generated from pre-set assets that came with a proof of purchase while the former uses pre-existing assets and mixes them together with pattern recognition and basic learning algorithms. The only major differences are that AI is a legit leap in technology while NFTs were a legit ponzi scheme, as well as NFTs mostly being part of a bigger mission to make cryptocurrency be the evilest thing to man.

But back to AI, clearly while the "creators" were all proud of themselves helping "stopping NFTs" and doing "creator's rights" shit, AI rolled out and, taking advantage to the egoists they are, said they were gonna use their work for learning, having misinformation be injected into their bickering to help spread even more discourse, and all of this to distract most people from the real lingering issue with AI: another step towards an unironic fully automated luxury gay space communism future.
 
One question that is always on my mind when talking about AI generated stuff is: Do anyone actually enjoys consuming those? Or is it always on the same level as throwaway media?

I personally do not consume anything that i can perceive as majorly made with AI-generation because it strikes me personally as something the creator did not want to be completely responsible for. It lacks intentional decision of the creator, who instead of looking and building every detail of something just hand-waves it as not important for the totality of what is being made. Given that the author does not think it was worthy to engage with all aspects of what they made, it just feels dumb to have myself then consume something that not even who created it put their all into it, so i would rather just go somewhere else where i know that even if not great i can safely try to gleam what the author wanted to show me knowing that in the end it was an actual real person responsible for it.
I am not saying every content made by real people is then worthy of being consumed, there is plenty of shit made by all sorts of people that is not worth the bytes of data used to transport it to my computer, stuff like ads, marketing material, content farm articles and whatnot always existed and I do not concern myself with paying attention to what some random ad is trying to sell me.

So when i ask the thread if anyone actually enjoys stuff mainly created with generated assets, what I am asking is does watching that stuff adds to anything on your own mind and soul? Or is it always some slapdash garbage made to be scrolled by without a second glance?
 
So when i ask the thread if anyone actually enjoys stuff mainly created with generated assets, what I am asking is does watching that stuff adds to anything on your own mind and soul? Or is it always some slapdash garbage made to be scrolled by without a second glance?
The Will Stancil Show and related media. Also Clair Obscur: Expedition 33, lol. There are quite a few AI generated songs I come back to quite a lot unironically.

You're not wrong that most AI generated stuff sucks, but most people have no taste and are using the tools wrong. You can absolutely be very intentional about every aspect of work you make using AI, you just have to care enough and spend time shaping the outputs rather than immediately posting the outcome of every one-sentence prompt to Twitter. It's entirely a skill issue.

AI generated/human created is not an either/or thing: we're going to see a lot of stuff that you won't be able to tell AI was involved in.
 
A regular anti-AI argument she shares is that using AI deprives people of "skills" that they would have learned if they had done it the hard way, which is a telling sign of someone who does not understand the amount of mentally draining, menial busywork in employment, and someone who has at least 40 more hours of free time a week than the average person to waste on "learning skills".
To be fair I think it does apply to students, those who actually need to study.

I overheard someone on the train ranting about a classmate who kept using GPT for every single assignment, and that she was certain that he was going to fail the exams and be fucked for life if he continued down that path. Kids definitely had a rough patch with Covid and remote learning being immediately followed by LLMs.
Yes, this cow makes me MATI because I work 40 hours a week, receiving only 20 hours of pay, as the other half goes to taxes that pay for this cow's frivolous purchases and smug lectures towards the people who keep her alive
I instantly knew you were from a Nordic country reading that, before I even saw the line under your avatar, haha
a sentient desire to perform well at the test. Researchers give the model a "reward" score when it answers a benchmarking test correctly, so the model adjusts the numbers so that its output yields the "reward" outcome more often, like how the horse doesn't actually know math but is just after the reward
I mean you could argue whether humans are really any different in this regard. Does a thing such as "sentient desire" really exist, or is it just a reward function with many additional parameters?
and refine the responses to just those types of prompts for the "reward" while the rest of the model doesn't improve
Or the rest of the model would get even worse. Although I believe LLMs are a much more generalized type of model that can produce good outputs for a vastly larger space than past CNNs or RNNs, which were more localized, and trained for very specific applications.

Sadly I didn't get to learn about the underlying theory of how they work at university, was a bit too early for that, and the information only got as far as the aforementioned CNN and RNNs. But apparently, something about self-attention, transformer models -- might be cool to study it in more detail at some point.
 
1770832204770.png

from: https://annas-archive.li/blog/backing-up-spotify.html

I don't like this attitude that AI-generated music doesn't have value. I have a little collection of AI-music that I enjoy. The only reason for it to exist is that I like it. And if something needs to be valued by someone to be valuable, doesn't that make my little AI jingles just as valuable as any other 0-plays track on spotify?
 
Oh, but you WILL read it, faggot. Because you won't know the difference. And you will enjoy it. Or you'll simply stop engaging with all media, and be miserable.

An author will lie and say they didn't use ChatGPT...or they'll give a white lie and say they didn't use it to directly write the book, when they actually used it extensively for ideas, character names, problems and solutions, to get out of writer's block...or they won't even do that, but a friend of theirs will have said something that gave them an idea for the book which was derived from something the friend learned or discussed with ChatGPT.

It WILL influence media you consume, and there is nothing you can do about it.
 
Wyświetl załącznik 8546654

from: https://annas-archive.li/blog/backing-up-spotify.html

I don't like this attitude that AI-generated music doesn't have value. I have a little collection of AI-music that I enjoy. The only reason for it to exist is that I like it. And if something needs to be valued by someone to be valuable, doesn't that make my little AI jingles just as valuable as any other 0-plays track on spotify?
Interesting. At the earliest, AI generated music became popular in 2023, meaning you already had almost 7 million album releases WITHOUT AI in 2022.

2023 and 2024 clearly follows the established growth trend. Sure, many of the releases of those years *may* have been replaced with AI songs, but then again, given the average Twitter artist's heavy dislike of AI, this seems doubtful.
 
Ok but what about that 1000x increase in 2019 compared to 1971 though? That's not a problem with ai, it's already being massively inflated before ai existed. If it was ai's fault then there would be no trend, it would be stable and then suddenly jump up. That graph is just pretty standard and sticks to a single trend, if you had to point at that graph and say where ai music first became a thing then where would you say?
 
So when i ask the thread if anyone actually enjoys stuff mainly created with generated assets, what I am asking is does watching that stuff adds to anything on your own mind and soul? Or is it always some slapdash garbage made to be scrolled by without a second glance?
So the wall textures in a hero shooter are supposed to add to your mind and soul? AI is perfectly fine for shit like that.
 
Wstecz
Top Na dole