*mic drop*
 Nothing Ever Happens Shirt $21.68 |
 DMT Has Friends For Me Shirt $21.68 |
Nothing Ever Happens Shirt $21.68 |
Fuck AI 'Art'
*mic drop*
Nothing Ever Happens Shirt $21.68 |
DMT Has Friends For Me Shirt $21.68 |
Nothing Ever Happens Shirt $21.68 |
Frick capitalism
This.
So much fricking this.
We all know there is only one REAL WAY FORWARD
>fricking HAIRLESS MONKEYS
>ONE MILLION YEARS spent on the face of the planet
>STILL NOT EVOLVED INTO CRUSTACEANS
>NOT EVEN DECAPODS DESPITE ALL THAT TRYHARD FINGER SHIT HAHAHAHAHA
Meaningless existence, evolutionary dead end.
CRABITALISM is the only way, the natural way, the way of the UNIVERSE
CRAB OR DIE
yes
this too
anyone who says "frick artists" and uses generative neuronal networks to create slop (trained on art made by said artists) is being an ungrateful little b***h, not only are you taking and not giving but you are being an butthole to those who made it possible even against their will, so frick you instead, making slop is free (if it isn't for you then lmao) but not being an butthole is too
post
this. why hate the tool when you should hate the creators
>inb4 we live in the reality of capitalism
why don't we work together to change that instead of breathing in copium? even if you like capitalism, at least help those whose livelihoods are at stake
Powerful. Biden will implement communism any day now, comrades.
>"FRICK COMPUTERS, TAKING AWAY MY JOB"
Imagine if everyone was as insufferable and entitled as a 21st century artist
Computers gave them that job in the first place because all of these frickers work digitally.
I wish I could frick my AI art...
I like the lines-"work". Is it a lora?
NAI V3
Total Artist Death
this. ai for life.
TSMT
F P B P
P P - -
B - B -
P - - P
ha! i get it
If you're against something just because it's a threat to your faulty busniess model you are a homosexual.
Me generating AI art
Frick Digital Artists
*prompt drop*
'ashi 666
so? give us the prompt anon, we're waiting
>email digital """"artist"""" asking for price on similar pic
>sure bro, it's $400
digital artists should cope with the fact they are plumber-tier at this point, instead of asking for hundreds of dollars for some bullshit
Prompters are gonna start sweating when AI replaces them too.
So happy animu sameface moeslop is first on the chopping block
reminder that this will never be allowed
reminder that they are already prosecuting people for 'pedo' ai images
Consoomerism is the spiritual death of this species. A population of golems born and raised to consoom will inevitably come to view art as a product, and the process of creation as a mere process of production. Under this framework, only the final product matters -- the product is viewed as a self-contained object in a vacuum, with no origin, no context and no history. This is necessary partly because the origin is usually sordid enough to devalue the product if it were part of its subjective valuation, partly because it is the consumer's "private property" and so it must come with no strings attached, or at least with no psychological strings that connect to the inner world of another individual. In a culture shaped by such a system, the method of production is incidental and irrelevant, thus the creator, who is just an artifact of the method, is also irrelevant, and so we naturally reach a point where golems will readily accept fake machine "art" because they can consoom it in endless amounts and it looks good enough in the 5 seconds of attention they're willing to dedicate to it.
But there is an upside to all this. Once AI "art" finally becomes indistinguishable from human art (from the point of view of art as a product), and the abundance of these products becomes so large that they lose their value and appeal as consumables, it's not human art that will die, but the idea of art as a product. In the long run, the original idea of art will prevail: the one that refuses to separate art from the artist, or the art from its creation process, and views both as integral to the piece itself, being parts of its identity and value.
art commissioned by kings and aristocrats were also products. The artists were providing a service. What difference does it make when the end customer is some gooner moron on xitter?
AI art is art chuds
>AI art is art
Art is a form of expression. A statistical model has nothing to express.
But the person prompting does
>But the person prompting does
Visually? Even if he did, he couldn't express it using a drawbot.
Why does it matter what tool he uses? Why does it matter if his expression isn't visual, for that matter?
>Why does it matter if his expression isn't visual, for that matter?
If you think your prompt expresses anything worthwhile, just post the prompt and call it art.
You just said that art is art because it expresses something, and now you're saying the thing it expresses isn't art
You are babbling nonsensically, but if you think your prompt is an artistic expression, just post the prompt next time.
Are you moronic?
You could use your own exact argument as a counter. Paintings people consider deep and artistic could be reduced to "clocks melting" and "man screaming with swirly background" art is and has always been bullshit its not the art itself that actually has any worth it's the story behind it, who created it and why they did so. If you use AI as a tool to acomplish your vision there is literally no issue artists do it. programmers do it, writers do it. It's a tool not magic. AI art can look good but it will because of above reasons never carry the same value as traditional art or even weird digital fetish art so i really don't understand why artists are so fricking worried. if anything "handmade art" will be more valuable in a market oversaturated by lazy AI art
>Paintings people consider deep and artistic could be reduced to "clocks melting" and "man screaming with swirly background" art
Could they be, though? You just took my point, reframed it and used it to BTFO yourself. Bravo.
ofcourse they could be, i just did didn't i? using your logic here
This. Same with digital ""artists"" who just show some jpg or png as if that has any meaning when their true artistic expression is in the psd file.
Plenty of artists do publish their psd and timeline files you know
Both of you seem to be actual, literal GPT bots, since you don't seem to have any grasp of context or understanding of what you're replying to
Is drip panting or where they have a bucket of paint over a canvas and swing it around also not art? They probably have less control than someone using SD continually prooompting and using something like controlnet
>drip panting
Your sister drips and she pants and she's taken being a prostitute and turned it into an art but this is offtopic. But on a more serious note, I don't know and I don't care about "drip painting". You can argue either way and it doesn't change the truth of what I'm pointing out.
Why do you care about gatekeeping a term in the first place? They're not stopping you from drawing or whatever. Quit being a gay about it
>Why do you care about gatekeeping a term in the first place?
It's my small contribution against the dehumanizing corpo culture you're spreading.
>Art is a form of expression
Yeah, and it's expressing me making artBlack folk seethe lmao
The content being used by the statistical model is the result of human expression
>The content being used by the statistical model is the result of human expression
Next thing you're gonna tell me the printer is expressing itself every time you print something.
Sure, why not?
Ok, golem. The way you embrace the nth degree of dehumanization only proves that the "people" pushing this meme aren't really people.
Why resort to name calling? You can't use facts and logic to argue?
What are you talking about? I'm not arguing with you. I accept your subhumn opinion and I think it speaks for itself.
Prompting is also a form of expression. Now what, homosexual?
>Prompting is also a form of expression.
Visual art expresses a visual vision. A prompt can't do this with any fidelity even if the prompter had a vision to express.
>Art is a form of expression
because drawing art as a job for some client who just wants to use it as a vehicle for marketing is the purest form of "expression". ai art might be generated and might not take skill but at the very least it allows the user to communicate complex ideas with faster iteration and no middle-man. people have had this same shitfit over cameras, photoshop, filters ad nauseam and it always ends with luddites btfo
Jeff Koons does not produce to express anything other than the dollar signs in his eyes.
>[insert literally who] doesn't express anything
Ok, and? Why does your hivemind keeps making this error?
Literally who? This is one of the foremost examples of modern artists, if his art is completely without meaning your point is entirely without merit. I don't care if AI doesn't produce anything as thoughtful as a Vermeer, because art™ today doesn't either.
>b-b-b-b-but THIS HECKIN' GUY
Ok, and? Why does your hivemind keep making the same mistake and doubling down on it? All you've proven is that some """modern artist""" isn't producing anything worthy of being called art, which we can both agree on. This doesn't refute anything I said.
Your argumentation is shit anon, your "art is expression" line of reasoning is entirely subjective.
>your "art is expression" line of reasoning is entirely subjective.
What's "subjective" about it?
First of all because there is no definition of art, some people think it is beauty, some people think it is representing an ideal, others think it is to point out social issues. I think that if you need a piece of text to explain art than you could just as well get rid of the art itself and keep the text. But just saying "art is expression" is far too facile to be of any meaning, maybe Dall-e Asuka #90433 speaks more to someone than universally acknowledged "really good artist". Most #art to me is nothing more than noise.
>First of all because there is no definition of art
There are many definitions of art, and all of them, barring a handful of the most inhuman ones, implicitly or explicitly establish the lower bound I put forward.
Then why isn't Jeff Koons art according to you? Is he not one of the most highly sought after modern artists (totally not for tax purposes btw)? Is "art is expression" beholden to solely your point of view of what is expressive? Do give some examples.
>why isn't Jeff Koons art according to you?
I didn't argue anything about Jeff Koons. Try again?
>All you've proven is that some """modern artist""" isn't producing anything worthy of being called art, which we can both agree on.
Do try again anon, give some examples of art.
That statement was contingent upon YOUR premise being true. Did you lie to me? Why do you argue like a israelite?
You can clearly not give any examples of what you consider art and the irony is that it is you who is arguing like a israelite.
Not an argument and my point remains unchallenged.
You still have not argued which art is expressive. Just one example of art would actually suffice anon.
>You still have not argued which art is expressive
What is this nonsensical babble? Why do I need to "argue which art is expressive"?
Don't bother dude, you're probably talking to an AI ironically enough.
Did you just reply to your own post? You did, didn't you? LOL
If you felt like you could make an argument you would explain or post examples of what you consider expressive/good art but you won't.
so true
Are there any statistics on how many scribble piggies have actually lost their jobs to ai? Would be interesting to see, although I fear it's not as many as people would think
Art is art, you intolerant bigot.
Wow homie, look how I don't give a frick.
adapt or die luddite
Why does AI art look better than most of the artists', despite being a fraction of the price? Sure, it's struggling with things like fingers, but a year ago it was struggling with eyes. And again, it's still better than most artists. Methinks too many people like to pretend they are artists, while demanding absurd fees for their subpar art.
Behold my art:
>clocks melting
I am now Dali.
that's a man with a vegana
So we have one AI fan claiming that when a printer prints a document, it's self-expression and another one claiming that Dali's painting is reducible into the phrase "clocks melting". Whenever you engage these "people", keep in mind that they all, without exception, believe things like this, and ask yourself what kind of impoverished inner experience, if at all, leads to one having the opinions of ChatGPT 2. Not all "people" are people.
imagine believing that art galleries are anything more than a candy store for rich BOT people wanting to put money on a non-inflatable asset
hell even most modern artists popular peices are about how much bullshit fine arts are. And still the BOTnizz men buy it like the clueless #deep asholes they are
>purely generic anti-human tripe
Aaaand there we go. Every single time. Without exception. You keep prodding these bots long enough and they start regurgitating these standard lines trying to slander humanity.
yep, it's either the talking point about money laundering and modern art, or the talking point about furries, usually followed by something about "most human artists are worse than X" or "all human artists are like diffusion models" etc.
shizo lmao
AI is not a problem for actual artists, it is however a problem for graphic designers since companies will just use some shitty prompt to generate a gross generic logo they don't wanna pay for. Im not saying AI is being used well for art rn but i don't think its the doomsday of all art like some people seem to think. It is being abused but they hype dies down eventually when people realise "AI" is just another keyword to sell useless chink products
i pay both artists and graphic designers within my line of work and i will gladly keep doing so. not becuse a computer couldnt do what they do but because theyre more creative than me and can sketch up ideas better than what i could possibly think of myself. Artists are not defined by what tools they use. Someone using crayons could be more skilled than someone with guache colors and a 3000dollar brush.
Didn't read. There is no such thing as "AI art" for the reasons previously explained.
Illiterate plebbit user want to share their opinion but not listen to someones argument how incredibly suprising
you're completely right. your opinion is the only one that matters mommy's special little snowflake UwU
You are seething but my point stands unchallenged.
>drawing art as a job for some client who just wants to use it as a vehicle for marketing
What does that have to do with anything? You can argue either way but it has no bearing on my point one way or another.
>art commissioned by kings and aristocrats were also products. The artists were providing a service.
What does that have to do with anything? You can argue either way but it has no bearing on my point one way or another.
>doesn't read counter-arguments
>my point stands unchallanged
Bro ran out of ammo
>b-b-b-but those illustrators are illustrating for reasons other than self-expression
And what follows from this, moron hivemind?
i think youre getting me and another poster mixed up my homie
Broken record. Youre starting to sound like youre the one using chatgpt to generate replies kek
now THIS is some ai generated content
You are seething but the point stands completely undisputed.
>that pic
>usually followed by something about "most human artists are worse than X" or "all human artists are like diffusion models" etc.
Like clockwork.
you DO know that restating a point isn't a rebuttal, right?
I wasn't restating any point, just lol'ing at how insanely predictable your programming is and giving credit to the poster who called it out 10 posts in advance.
>what is 1+1? no don't say it's 2
>it's 2
>lol he actually thinks it's two lmao at your npc robot programming i already predicted your answer so i don't even have to explain why it's wrong, i win my point stands undisputed
>human artists are just heckin' diffusion model it's like 1+1
Keep doubling down and making my point for me.
the point isn't about the inner workings, but the fact they both ingest the same things and create things that are undiscernable (or else why would you call it theft, and didn't two years ago when it was hot garbage)
the aristocrat point from the other anon for example is about how if ai art isn't art because it exists just to satisfy a consoomer, then commissioned art cannot be art too because the money stiffles the expression, ultimately you're seething because a man in the privacy of his home got some drawing in the style of a guy that wasn't paid, as if copycats weren't a thing before. also, some could argue that prompting is a form of art, albeit as involved as mixing colors, it could be seen as a technical feat as it's like there's millions of "primary colors", it's at least as much of a "theft" and involved as writing shitty fanfics about copyrighted characters
and the 1+1 thing is to point at you being unwilling to offer any rebuttal other than "no ure wrong"
>they both ingest the same things and create things that are undiscernable
Proof?
which part of the statement needs proof?
are you trying to say there are artists in this world that are blind from birth yet create expressive works?
or that the results are easily discernable, in which case, why are you concerned at all, and how does it get categorized as theft?
>which part of the statement needs proof?
Both of them need proof.
>are you trying to say there are artists in this world that are blind from birth yet create expressive works?
No. Are you hallucinating again?
>or that the results are easily discernable, in which case, why are you concerned at all, and how does it get categorized as theft?
What is this question even referring to? Which two specific things are being compared?
are you even a real human
artists ingest images, things they see in the real world and get influenced by other people's work, they repoduce it and distort it from reality in the process
>Which two specific things are being compared
something drawn by an artist and a piece generated by ai, moron
>artists ingest images
5 billion of them?
>something drawn by an artist and a piece generated by ai
Ok. I can distinguish pic related from something drawn by an artist, thus you got proven objectively wrong. Move along.
>thus you got proven objectively wrong. Move along.
no, you proved me right, how can it be theft if the difference is so obvious ?
Sorry, please tell me again how artists ingest 5 billion images and then create pictures indistinguishable from the pic here
.
> how can it be theft if the difference is so obvious ?
What is even the connection? You are so mentally ill it's not even funny.
what the frick is even your position
if it doesn't go anywhere beyond "ai art isn't art" then there's nothing to argue
>what the frick is even your position
Did you make the following statement?
>the point isn't about the inner workings, but the fact they both ingest the same things and create things that are undiscernable
Yes, you did. Now please tell me again how artists ingest 5 billion images and then create pictures indistinguishable from the pic here
.
>if it doesn't go anywhere beyond "ai art isn't art" then there's nothing to argue
That's a concession. It's not "art" but it is a particularly egregious form of plagiarism.
>how artists ingest 5 billion images
by living
Are you arguing there's some sort of upperbound at which it becomes, like what, i don't even see where you're going with this?
> pictures indistinguishable from the pic here
>artists ingest images
5 billion of them?
>something drawn by an artist and a piece generated by ai
Ok. I can distinguish pic related from something drawn by an artist, thus you got proven objectively wrong. Move along.
despite your cherrypicking, i see people mistaking ai art for real art all the time because they don't even know of the concept, I only need one image that's fooled at least one person to make my point stand, while the opposite requires than no ai generated image ever fooled a human, gee i wonder which is the most likely
>it is a particularly egregious form of plagiarism
proof? can you show me one image that ai generated and a blatant plagiarism of any artist?
>by living
See
. Not reading any more of your shart. Every member of your hivemind has a severe psychotic condition. Society needs to put you down like the rabid animal that you are.
that was always their position. "new thing isn't real, unlike old thing". totally new and groundbreaking idea, not at all repeated throughout the ages for every new 'degenerate' movement or whatever.
>how artists ingest 5 billion images
by having eyes and a brain that learns continuously, the stuff that pops in your head isn't divine inspiration.
Even if an aspiring artist studied one image per minute, every minute, for the rest of his life, he couldn't even come close to absorbing 5 billion examples of other people's work. You are mentally ill.
he doesn't have to because his brain is absorbing imagery 24/7, you're mentally deficient.
>his brain is absorbing imagery 24/7
He'd have to consume other people's work at a rate of 2 per millisecond, for every millisecond of his life, to make it to 5 billion. You have a very severe mental condition.
what does it even mean, would you recognize ai as real art if it was 500 images of training?
>what does it even mean
It means you have a very severe psychotic condition where you keep hallucinating things that are obviously impossible.
>what is an hypothetical
clearly you're the dysfunctional Black person here
i accept your concession
>obviously impossible.
kek
remember, 5 years ago what we have today was thought impossible
so that's it, that's your cope?
gg, was fun arguing with a moronic kid
>the context window runs out and the LLM breaks
the only thing you're demonstrating is that the human brain is more efficient than machines, something no one disagrees with.
It's demonstrating that your mentally ill statement is false no matter how you slice it. Human artists learn through a different process and produce art through a different process. This is not up for discussion unless you are willing to stand by your psychotic assertion that human artists ingest 5 billion examples of other peoples' art to learn how to draw.
they still ingest images and produce images, unless you're willing to define bounds and the meaning of said bounds, there's no difference between 5, 500, 5 billions, or 5 quintillions. learn to logic
>they still ingest images and produce images
They don't ingest 5 billion examples of other people's work and they don't produce regurgitations that are by definition entirely defined by other people's work.
you still haven't explained why you're that attached to that 5 billion, i won't entertain you any further until you clarify why you think that's an important factor into discerning plagiarism from original creation
>you still haven't explained why you're that attached to that 5 billion,
Because it conclusively refutes your mentally ill assertion that human learning is analogous to corporations gobbling up mountains of other people's work.
in what way? is there a number of training images that would be ok?
>in what way?
In the way that it firmly establishes that the process of leaning is vastly different. That the process of "generation" is different is completely trivial, a given and not even up for discussion.
i don't have anything to argue with someone unable to deal with hypotheticals
>In the way that it firmly establishes that the process of leaning is vastly different.
The process of learning isn't going to be same unless you invent AGI and put it into a robot, it's still learning and that's the only thing that matters.
>That the process of "generation" is different is completely trivial
So ? Things don't have to be made a certain way that is defined by you to be considered "real".
>The process of learning isn't going to be same
Tell this to your psychotic buddy here who thinks it's the same due to unspecified "hypotheticals" the voices whispered to him.
>So ?
So again, tell your psychotic buddy. lol
you're threatening people with violence because your patreon didn't take off, i think you're the one with psychosis here.
>the psychotic spiteful mutant is arguing with imaginary characters in his head
Tell me more about my Patereon. Is it in the room with us right now?
le in the room with us response implying he is crazy lmao so funny!
upvoted
What about the people who threatened you with violence? Are they in the room with us? Do they know where you live? Are they in your walls? What did they say they will do to you? :^)
>Human artists learn through a different process and produce art through a different process.
doesn't mean the process isn't comparable. off course it's not the (exact) same process.
>This is not up for discussion unless you are willing to stand by your psychotic assertion that human artists ingest 5 billion examples of other peoples' art to learn how to draw.
How is it a "psychotic" assertion ? Humans can learn, computers can learn (albeit less efficiently).
Also, 5 pictures/second for 30 years gets you close to 5 billion. Your math is off.
>>
they can jack each other off at galleries. These are the people that are gonna buy ai art. I'm gonna keep drawing my shitty meaningless sketches and appreciate the little things while eating for free at a gallery while they burn their money on garbage
AI stands for Artist Immitator.
If only making friends with any AI was possible... Don't matter the algorithm or network, you can only see a fricking mirror to someone else's autism/insecurities.
The "People are NPCs either way" phrase is complete BULLSHIT.
and there goes the spiral into full schizophrenia
You can keep spamming and mumbling to yourself(selves?) but AI art still doesn't exist. lol
>that watermark
Maybe I'll start using AI now
Meds! Now!
That seething AItard is clearly beyond meds.
ai art is just better
it just is
sorry drawgayoids your time is up
get a real job like everyone else
>muh frick capitalism
lol
lmao
this is so funny because how do you think the "ai art" even exists in the first place
because actual art was forcefed into it and regurgitated after
top kek there are no words to describe this post
Graffiti isnt art
>Frick AI 'Art'
I wish I could...
*does a 360 & dadjokes the frick out of here*
you have to understand your opinion doesn't matter and I am not going to stop
People do. Everyday.
Powerful
God, why?
@98355123
>doesn't mean the process isn't comparable.
You have not show any aspect in which it is comparable and you never will. You and the rest of your corporate hivemind need to be put out of your misery.
>put out of your misery.
An unemployed twitter user will put me out of my misery ? you can't even get a job, let alone carry out an assassination.
Lol this homie is still going at it. I love ai for absolutely mind breaking art troons
Reminder that your Silicon Valley corporate handlers are all currently building bunkers and private islands. Tick-tock. Tick-tock. :^)
I will be fine, I think. I have a real job. You should try it sometime
>be artist
>press pen on tablet
>digitizer reports coordinates, angle and pressure information to the software through hundreds of levels of abstractions between the bus, hardware, os, driver then final software
>software uses that information to transform a preset of brushes onto a layer that then composed with other layers with each one of tens of possible composition modes
>then I can use post composition filters that use algorithms based on complex number, matrix decomposition and shit
hmm yeah, feels good to be a skilled traditional artist, it would make my cavemen ancestors proud. everyone owes me. I MADE THIS
>be proompter
>type a single word
>change a single weight somewhere by 0.00001
>output is completely different
>have to traverse a hyper dimensional space of variations to find exactly what I envision
>finally reach it
proompter: AAAAAH IM GONNA COOOOM IM COOOOMIIING, phew that was wild. here's it is guys, i hope like-minded individuals will find it pleasing
"artists": NOOOO YOU CAN'T DO THAT IT'S THEFT YOU OWE ME YOUR CUMMIES. PEOPLE WILL THINK I DREW AND CONDONE THIS
proompter: i don't even know who you are
thigh sex from the front with Miruko
I think image generation just goes to show how imitative and redundant pop internet art and graphic designers truly are.
Only fine art will survive the modern age as a means to a job. Otherwise it'll just be a hobby and that's okay because it doesn't get much better than that today either.
fine art you say?
>company rushes to feed it to the model so that their GAI remain relevant in the market
>the artist is no longer relevant because bots are generating anything and everything you can think of in their style
Ok, generate some van Gogh and sell it for a few millions then. Anyone can do it, right ?
Ah, no, the old-fashioned art forgery isn't the main game here. The average man typically doesn't buy actual paintings anymore. They buy prints and posters. Endlessly reproducible.
The main game is generating whatever might be marketable in the style of famous artists and then selling them as prints, posters, t-shirts, mugs, whatever. Sure, you profit maybe a few dollars per item you sell, but if you make a good enough app where people can quickly prompt a hoodie design and order it from a print shop in china, you can make big bucks.
People buy merch of bands, anime and TV shows, not drawings in the style of "famous" furry artists.
>print shop in china
they sell truckloads of the aforementioned things on ali, you can also just send anything you want to any print shop.
jfc i'm not gonna bother when your smooth ass coomer brain can't think further than the furry porn you're addicted to.
Fine art or 'collectible' art (putting it in the language of dnd might help you) is mostly about the artist and the work's history rather than any objective value you illiterate.
A supra is cheaper and technically superior to a 250 gto, which one is a 100 times more valuable than the other?
Black person, that's not what "fine art" means. You're thinking of "traditional art" or "physical media".
That market is corrupt as frick and virtually meaningless to people below the top 1%. Your "fine art" is either historical artifacts by long-gone artists, or it's contemporary nonsense art that's been marked-up to commit tax evasion.
aislop forgings of mona-lisa is obviously not what's at stake here. it's the fact that new artists will have a harder time emerging or have the benefit of being able to practice their craft and create art full-time.
ITT:
artists find out that coomers don't care about them (or the potential devastating effect it might have on the job market and economy), they only care about cooming. where the porn comes from has no bearing on it's coomability.
I just installed stable diffusion and I have been playing with it.
>frick AI art
You say this because you don't get it.
There's no "AI art"
This is a program that generates images.
It's not artistic. It's 90% random and you depend on someone else wasting a city's worth of electricity to make a checkpoint so you can make something remotely good looking.
I'm downloading my 3rd anime checkpoint right now because the other 2 don't work well with the style of the artist I'm trying to rip off.
How do I rip off the artist? I downloaded a LoRA someone else made based on that artist's drawings.
Do you see a pattern here?
I keep downloading stuff other people made and I don't know how any of this works just in order to rip off an artist.
It's not "art". I have almost zero input in this.
I don't know where the LoRA tokens come from. I'm guessing a booru. It's not good enough for me. I'll need to train my own LoRA with my own image descriptions just to have a tiny bit of creative control over this machine.
And I still wouldn't call it art then.
when you keep just downloading stuff others make your control is extremely limited yes, that is so far from pushing the limits of ai art
Do you have anything to do with your time, besided this weak trolling on BOT?
I will not bother (you)ing your messages, you know who you are. Though I am afraid you might miss this still, low intelligence is why you went into the arts rather than to STEM, Law or philosophy to begin with
I wouldn't mind if people weren't flooding the boorus with untagged slop. Its like looking at ESL writing that no one has bothered to edit.
They call me an 'AI art apologist'. It makes them sound moronic because I have never once apologized.
I am banned from over 400 subreddits for posting ai art, SEETHE jannies
>400
I cannot even name that many websites let alone subreddits
I'll all seriousness it is sad so many artgays do not see the potential in further democratization and streamlining of art in general.
But if everyone can make mediocre art most of them will be forced to find real jobs, it is not fair
your're'r move human
hand's the wrong way round
AI makes labor and therefore communism obsolete. Who needs to seize the means of production when you can just copy the means of production.
All aspiring artists working at Starbucks will soon have no recourse but to live the remainder of their short lives high on tranq in the ghetto.
>this is what NPCs actually believe
Imagine growing up in a culture and environment that makes you believe people are equivalent to a simplistic statistical regurgitator. Now imagine believing in this so strongly that you become one and spend your time shitting out dehumanizing talking points coming straight from a corporate marketing team.
no one says they are equivalent, just the way they can combine things once seen in the image, in that sense the AI models "learn" in the same way humans do.
>denies doing a thing
>immediately proceeds to do it again
Also bravo on being utterly delusional and knowing nothing about the subject of how machines "learn".
>nothing about the subject of how machines "learn".
its called machine learning because it mimics human learning, there is no other way you can explain how a 4GB model file can create basically any image you want for example otherwise
You know what, let's go with the most charitable interpretation possible of your corporate meme:
>machines learn from human art in order to create new art
>humans also learn from human art in order to create new art
Ok. Therefore what? How is this relevant to the discussion, when the two then diverge widely in how they create art, with completely different implications?
what
human artists are just organic diffusion models that consume other human artist's art and then spit out variations of it, that's why we professionals call it machine learning, because it's analogous to human learning
oh
Frick entartet art
If an artist isn't dumb as a coffee table, they should understand that they can only believe one:
>art created by a real human bean provides some unique value (philosophical, ideological, aesthetic, money laundering - doesn't matter) to other people, unlike AI art
>AI art isn't a thread to their craft
OR
>human-made art DOES NOT provide anything unique over AI-generated art
>AI is going to steal their jobs
YOU CAN ONLY BELIEVE ONE
All of your points are completely irrelevant and immediately stomped by the following observations:
1. Your corporate overlords have no right to gobble up other people's work regardless of any reddit arguments
2. Their endgame is to deprive professional illustrators of their livelihood using their own works against them without permission
3. You and your corporate masters are creating a completely cancerous culture
>have no right to gobble up other people's work
>using their own works against them without permission
A copyright respectoooor on BOT, what a time to be alive. I thought we're the cool kids who enjoy web scraping with selenium, ripping music, cracking software, and downloading torrents?
I can do whatever I want. Your corporate overlords are not people and don't have any rights, least of all the right to abuse other people's work in order to harm them.
>Your corporate overlords are not people
They are, in fact, people, and have more power and legal budget than you, so they actually can do whatever they want.
I dunno why you think I'm on "their side", I'm just stating facts.
No one cares about your babble. Here's your daily reminder that your owners are so afraid they are building bunkers.
>No one cares about your babble
You do.
>2. Their endgame is to deprive professional illustrators of their livelihood using their own works against them without permission
>3. You and your corporate masters are creating a completely cancerous culture
this. these are two parallel issues: one concerns the attempted destruction of people's source of income, and has little to do with "art" per se; the other concerns the propagation of a degenerate mindset that devalues actual art culturally, as well as professional illustration financially. the most obvious damage to art and culture comes from the intersection of the two: people who could previously hone their craft WHILE making a living doing professional illustration would have to use whatever time and motivation they have left after a day of soul-draining wage-slavery, limiting their progress if not demotivating them entirely.
>destruction of people's source of income
Why would an artist's income be affected if their work is superior to AI art and is necessary to others?
>devalues actual art culturally
Why would the value of the actual art be affected if AI art doesn't provide the same value?
Black person, what corporate masters? Stable Diffusion is FOSS, noone has it under control, don't give a shit about other services.
Artists kept whining about "muh stolen data", so guess what? Getty Images used their giant library of stock photos they legally own to make their own image generator, this is good, right? They have the full rights to use the images, after all.
Well artcels still bawl their eyes out about how it's not fair, should've thought about that before giving up the rights.
So no, I don't feel bad for them, because
1) They can't even stick to their own requirements and keep moving the goalposts
2) They, or anyone else are not complaining about other jobs being automated, why should they get special treatment while other jobs become obsolete?
3) The only people creating cancerous culture are the artists themselves, they are the ones being toxic to AI artists, insulting people regardless of why they do AI art. They are the ones accusing eachother of using AI, because of their giant ego's. They are the ones refusing to use AI as a tool that could make their jobs easier (since they constantly complain about crunch and too much work), despite Stable Diffusion having 0 entrance cost (It's understandable to be pissed at proprietary tools)
4) They are moronic enough to think a giant dataset like Laion 5b is still required.
Any huge corpo, like Disney owns enough assets and can join up with other ones, like above mentioned Getty Images, to train their own private model, which was trained only on data they legally own, completely shitting on artcels biggest argument.
Stick to the single-line corporate zingers.
>corporate overlords
>in FOSS
the absolute state of morons lmao.
AI is more creative than you are
>its called machine learning because it mimics human learning
This is what normie no-coders with no ML knowledge actually believe.
>cant explain how a 4GB model file can generate any image they want
concession accepted luddite, thanks for playing.
>its called machine learning because it mimics human learning
>i know this because Poop Diffusion is X GB and can generate many images
This is how normie no-coders actually reason.
>bot cant respond so it starts seething
already accepted your concession, no need to have a melty paintpiggie.
>it's called machine learning because it's just like heckin' human learning
>humans use gradient descent to learn
>humans need to process billions examples of other people's art to draw something
lol ok
>now he starts goalpost shifting and strawmanning while still being unable to respond how a AI model can create images or write coherent responses despite its small size
poor guy, imagine how this loser looks like irl, seething on anonymous forums online after getting BTFO while being a luddite at the same time, kek
cheers kiddo, good luck with that degree you got btw KEK
>tfw there's like a dozen of small art galleries around me
>only go there because an indie gallery is the best place to meet slender art hoes with septum piercings, chokers, and bob cuts (I love art hoes so much it's unreal)
>I've seen hundreds of photo sets, paintings, drawings, installations, improvised light shows and projections
>THEY'RE ALL EXTREMELY SHITTY. IT'S ALL BAD, ALL MODERN ART IS ABSOLUTELY WORTHLESS, THERE IS NO REDEEMING QUALITIES IN IT, NO INNATE VALUE
>continue to go there because it's easy to discuss and saying more than 2 coherent sentences in a row impresses zoomettes
I don't know what do you people mean when you say this shit has 'value' and should be protected from le AI.
Keep the standardized corporate zingers coming.
The two reasons modern art industry exists:
1. The scene is gatekeeped as frick and is gives people the sense on belonging to an elite group.
2. It allows the artists to inflate the prices.
Post good modern art.
>Post good modern art.
Why? I didn't mention modern art. Your generic corporate zinger #3 did.
So are you going to show us what the evil corporations gonna strangle or not? I'm not sure we should feel sorry for the poor artists.
>I'm not sure we should feel sorry for the poor artists.
>we
I don't consider you and your "we" human. I wasn't addressing you or your subhuman sensibilities in the first place.
No one cares what you 'consider', anon, since you're operating on emotions and assumptions and fail to support your point of view.
No one was addressing you, animal. Why do you keep barking?
Oh, another AI-generated response
I honestly don't know why your mindless horde thinks this is some kind of debate and owe them explanations. Whenever I hear you screech standardized corporate zingers claiming that art is shit anyway, that the people on whose work these models are based are shit and deserve nothing, that humans are just glorified diffusion models etc. there's really nothing I have to do except draw real people's attention to this, so that they learn to recognize your horde, understand what agenda you're serving and distance themselves from you no matter what they think about the usefulness of AI. You try to blend among midwit tech enthusiasts but you are something much more toxic and cancerous than they are and you need to be pointed out, circled and isolated from real people who are merely stupid or misguided.
>that the people on whose work these models are based are shit and deserve nothing, that humans are just glorified diffusion models etc
Never said that. But modern art is shit and if you insist it isn't you can show some examples of modern art (we're talking visual kinds right?) being good. Or you can ignore me and call a dog but then I don't understand why you even maintain the discussion.
>the rest of the text
Sounds like you're on a so-called 'heroic mission'. It's one of schizo symptoms so if you're still self-aware I'd suggest to look for help.
Why are you barking in my direction again, rabid animal? Someone should kindly put you down. That would be an act of mercy, especially towards the humanoids you call your family.
Oh, so the self-awareness train has departed already. Have fun then!
>shart-post.exe --corporate-db --genericity-level 100
>goalpost strawman loser irl seething anonymous online BTFO luddite kek cheers kiddo degree btw KEK
sorry but AI is cute and based
the melody reminds me of white wedding
Ai isn't going away no matter your kvenching. Better have a nice day now
AI art isn't really the problem. The problem is the shitty corporations replacing their human employees with Stable Diffusion.
I use Stable Diffusion myself but still occasionally try to support artists whenever I can, and actively practice Drawing as well, just because some computer program can draw so much better than I can and can learn it at a much faster rate doesn't really stop me from enjoying drawing myself.
Reasonable take, I also have mixed feelings about how certain groups are pushing AI prompt users as a new breed of artist (Usual argument is how it levels the playing field for people or something) which feels like a way to legitimise themselves.
But I start thinking if you consider photographer is a kind of artist and their work is based around using a "tool" - I've seen the term synthographer bounced around for artists who primarily use AI as their medium bounced around lately. But it still makes me feel a bit out of sorts.
>But I start thinking if you consider photographer is a kind of artist and their work is based around using a "tool" - I've seen the term synthographer bounced around for artists who primarily use AI as their medium bounced around lately. But it still makes me feel a bit out of sorts.
The discourse is as old as art itself. There's an easy way to separate all artists in 2 broad categories:
Classic artists, who create using simple tools and techniques known for hundreds of years, put in a lot of effort, and produce objectively beautiful (yes, beautify is objective) results.
Assisted artists, who create using increasingly complex tools and modern tech, and produce whatever they want. Photographers, electronic musicians, wacom furries, etc.
AI assisted artists fit perfectly into the latter category.
>beautify
*beauty
There are very important distinctions you are missing here.
Even digital mediums (visual and musical) has required a great understanding and practice of how to use your tool in order to create something excellent. Just because you have a Wacom tablet and a pirated copy of photoshop doesn't mean you can paint like Craig Mullins. Just because you downloaded FL Studio and know the difference between a major and a minor chord doesn't mean you're ready to work as a professional producer. Just because you have the best camera phone on the market, or even a proper DSLR, doesn't mean you're ready to go work for National Geographic. There is a lot of work and expertise that goes in to a high quality piece of art using these tools.
The direction AI art is pushing towards is "just make a wish!". There's no level of expertise required here, and in a corporate setting they don't even care about generating enough options to find the one that looks or sounds the best. Just hurry up and pick one. No style, no taste.
>There is a lot of work and expertise that goes in to a high quality piece of art using these tools.
I agree that there are great photographers and producers that possess exceptional mastery. My point is it's a different kind of art and you can't compare, say, Aivazovsky's Ninth Wave and a 48MP pic of a sea, even if it's a great shot.
It's a different kind of expertise.
>The direction AI art is pushing towards is "just make a wish!". There's no level of expertise required here
I disagree, SD requires a lot of technical knowledge and intuitive understanding of the way those programs process your requests. Even anons in the DALL-E thread used to demonstrate some pretty creative ways to force particular outputs. It all reminds me the other kinds of what I called "assisted art" all too much.
Even after a few more iterations, when the models' interpretation of the prompts' improves drastically and everyone will have access to an uncensored one, you will still have to "paint" your output with words, and the more capable you are in describing all the important details, the better is your result. "Prompt engineering" is a cringe term, but there's definitely a cohort of people who are more competent at this kind of writing the the rest.
>you can't compare, say, Aivazovsky's Ninth Wave and a 48MP pic of a sea, even if it's a great shot.
You're right, they're very different types of art that require very different skillsets and years of practice. The same isn't true for AI generated art since you get 80% of the way there by typing "big wave in ocean" into the prompt, and then you can take a couple of minutes to refine it further. And this goes for *any type of imagery*, plus other types of media.
>SD requires a lot of technical knowledge and intuitive understanding of the way those programs process your requests.
This is still very early days of this technology and these AI are only going to get better at interpreting prompts. That's a key factor to what's going to determine market leadership in the future. Again, you're already 80% there by typing the shortest summary into a text prompt. Aivazovsky was not 80% there when he drew the first pencil stroke of the sketch under his painting.
Furthermore, you have to understand that this is going to have a huge detrimental effect on the demand of handmade media. AI generated work is going to be normalized, everyone who needs a picture is going to prompt it in a couple of seconds rather than commission/hire an artist to do it, fewer people are going to bother learning anything when the computer can just do it for them, and only a fraction of those who remain are going to be able to do art for more than a couple of hours a week.
This isn't just a moment like the introduction of the automatic loom, this is more like a shockwave of multiple Industries suddenly getting their equivalent of the automatic loom *plus* they realise that they can outsource their work to chink slaves rather than domestic workers, all at once. This risks being a huge tank on the job market and economies around the world. This goes way beyond you hentai and furry hobbyists charging for commissions.
Oh no people will no longer be able to make the bare minimum slop to offload to corporations anymore and will actually have to pursue being a good artist if they want to have a hope of making money from their passion. What a fricking tragedy.
This has nothing to do with quality.
Corporations just want fast and good-enough results.
Vast majority of people just buys whatever the corporations push out onto shelves.
"Collectors" just want bullshit to over-evaluate so they can evade taxes.
Have to go and can't continue with the discussion, but you make a lot of valid points.
>AI assisted artists fit perfectly into the latter category.
You're no more an "AI-assisted" artist when you shit out a 10 word prompt and get an image than you are an "artist-assisted" artist when you comission something and have someone else draw it for you.
AI prompters absolutely aren't artists, I don't call myself an artist because I can type words into an image generator, i just call myself a degenerate who makes porn with it, however since anons said that art needs to have the artost have some idea on what the output is and to express a certain type of style, then by that logic Stable Diffusion fine-tuners and lora makers ARE artists, but not the people using their tuned models
>ML isn't real learning like a human! ML AI can't be anywhere close to a human!
>why yes, I simultaneously believe AI ML models are stealing MY work exactly the same way a human would, no it's not the human prompter it's the AI that gains sentience only in that specific scenario and NO STOP ASKING I won't post a single example of my art being blatantly ripped off, it just... it just exists, ok?!?
NOOO I can't sell my shitty yiff art for 2000 dollars anymore! I have to get a real job 🙁
@98365273
>sharts out corporate zingers about "digital artists"
>uses models trained overwhelmingly on their work
These malicious corporate drones aren't people and they shouldn't be indulged beyond getting them to regurgitate their corporate lines and expose themselves.
You do know that BOT has largely had it in for artistic types for years right? AI art is simply the nail in the coffin for the artist faction because we can finally own them, and even more ironically it can be done with their own work.
The fact that the STEM gang have beaten the fruity art school dropouts at their own game is the ultimate coup de grace.
>this corporate talking point again
They replaced the PR team with GPT bots, clearly.
How the frick is a legitimate rebellion over the drawpigs stranglehold over digital art something corpo?
We have local models dude, no corpo services are needed.
>yet another standardized corporate zinger
It's like clockwork. You will never hear them stray from the same 4 PR talking points. You will see the same 4 repeated dozens of times in every one of these shill threads. It's uncanny.
Shill thread? Some art Black person started this thread on a technology board in order to argue. Just stop whining and draw.
> Some art Black person started this thread on a technology board
No, it's your horde of corporate shills and their brainwash victims that starts this thread every single day.
What are the shill points anyway
Here's their entire program:
If the target knows nothing about ML, humanize the machine.
If the target knows something about ML, dehumanize the human.
If the target doesn't care about ML, but cares about art, slander art and artists.
If the target doesn't care about ML or art, but hates soulless corpos with their replacement agenda, start rambling about the New Renaissance and how AI empowers individuals.
If none of the above work, accuse the target of drawing and selling furry porn.
I guarantee that 99.99% of the shit this horde spouts follows this template to a T.
All true and valid points
>i confirm that my horde operates according to this program
Ok, thanks.
anti AI tards are all ignorant morons. i'd know, i've tried to engage them countless times now.
they refuse to understand the most basic things about AI.
and when you try to explain how it works they'll basically say
>it doesn't matter/why does this matter how it works? i believe what i believe, and i believe that AI is stealing souls
this is really the gist of it. there's nothing else to their thoughts.
they are fundamentally denying that the AI is capable of doing what it is OBVIOUSLY doing (which is learning), and choose to believe in fearmongering and being outraged instead.
even when you tards talk about "statistics" and "curve fitting", do you realize what that means in this context?
curve fitting means it is trying to fit into the pattern of the dataset. to be INDISTINGUISHABLE among the pattern. that's the goal and that the AI is trying to achieve.
Threadly reminder that statistical models of other people's work shit out output that is entirely defined by other people's work. It's actually the purest form of plagiarism ever invented.
there is a difference between copying the datapoints and inferring new datapoints.
this is what patterns recognition and learning is about.
there have been papers talking about how you can remove 10 thousand images of a specific concept from the model and the model will still output the concept in almost the exact same way.
that's because it isn't copying or cobbling together data, it is generalizing. it's using the data to "learn".
You did not dispute anything I said. Your "AI" is a statistical model of other people's work. Every output it shits out is completely defined by other people's work. It is the most perfect form of plagiarism ever conceived. A human specifically trying to plagiarize would inadvertantly add something of their own into the mix, meanwhile the bot is 100% determined by the work of other people.
>Every output it shits out is completely defined by other people's work.
>You did not dispute anything I said.
read my post again you braindead moron
Nothing in your post disputes this. In fact, no one working in the field disputes this. It's a trivial premise of the entire field.
>"Every output it shits out is completely defined by other people's work."
>removing 10k images and still giving the same output
how "defined by those images" was the AI here? clearly not very much.
obviously we're talking about very large concepts there with tons of data, but the point is the same.
generalization.
>Your "AI" is a statistical model of other people's work.
i already adressed this.
yes, and that means it is trying to fit into the dataset. not copying, not plagiarizing, but being it.
again, it's the same shit. you morons refuse to read and just stick to your misinformed beliefs.
what exactly do you think is not disputed here? generalization? even i'm not disputing the curve fitting nature of AI. i'm just explaining to you what that means.
>how "defined by those images" was the AI here? clearly not very much.
100%, in every aspect, by definition, since this is output from a statistical model of a bunch of images drawn by humans.
>it is 100% defined by an image that is removed from the dataset, i.e. not in the dataset anymore
just fricking lol.
again proving my point that you subhumans just stick to your delusional beliefs and don't actually read posts.
Who are you quoting? Are you losing your mind with rage? Sorry, but a statistical model of other people's work can only ever generate output defined by other people's work. There is no way around it. At least learn the basics of ML before you gape your filthy, animalistic maw at people who know vastly more than you.
>no counterpoint
>literally repeating your post despite me pointing out your complete and utter moronation (which you don't even seem to understand)
>trying to go offtrack with "u mad?" bait
>ad hominems
>appeal to authority (on topics which you don't really understand yourself)
same old shit every time.
i learned that it is a waste to engage with you subhumans. you basically argue with your emotions and you refuse to learn or even engage the facts.
you're tech illiterate, science illiterate homosexuals who love to be outraged.
i really did try for a long time to get people to understand. but you people literally don't want to. you'd rather stay mad.
>there is redundancy in the dataset therefore the model is not defined by the inputs
Imagine being this much of a moronic ape.
the model is obviously defined by the inputs. but that is different from it copying or plagiarizing the inputs. as my example clearly shows.
generalization is a accepted concept. and it is the thing you morons never want to talk about.
it has a GENERAL understanding of its tokens and concepts, inferred from seeing patterns in the dataset.
it can't do anything with a single image other than copy it. but if it sees a thousand cats, then maybe it will start to grasp the general features and shapes, textures and everything that represents a cat in general.
>the model is obviously defined by the inputs
The model is entirely defined by the inputs to the training process (aka other people's work) plus some statistical modeling assumptions. The output is entirely defined by the model. The output is defined by other people's work. Q.E.D. If your image completely depends on other people's work and is entirely defined by other people's work, you are a perfect plagiarist.
with this kind of logic, you can even say that people are plagiarists of real life.
you might think this is a leap of logic, but this is EXACTLY what you're saying. that output is forever theft of the input, regardless of what happens inbetween (learning or copying)
>b-b-but humans are not machines!!! machines are not humans!!!!!
yes, that'll be the same moronic point you subhumans always bring up.
yes we aren't the same. we are way better at this than machines. and we have way more inputs than just images and text.
and? that does not refute anything i said. not unless you're denying my notion that the machine is learning.
which you probably are, because like i said, that's the basic fact you are all delusional about.
>you can even say that people are plagiarists of real life.
>If the target knows nothing about ML, humanize the machine.
You tried it, it failed. Now you've moved on to:
>If the target knows something about ML, dehumanize the human.
Clockwork. Perfectly predictable. You are running a standard corporate program.
and there we go. nothing to say anymore.
waste of time. same as always.
go kill youself.
The model is entirely defined by the inputs to the training process (aka other people's work) plus some statistical modeling assumptions. The output is entirely defined by the model. The output is defined by other people's work. Q.E.D. If your image completely depends on other people's work and is entirely defined by other people's work, you are a perfect plagiarist.
>b-b-b-but humans are also statistical models of other people's work
See
You are now executing corporate programming routine #2:
>If the target knows something about ML, dehumanize the human.
first off, it is completely laughable to say that you understand anything about ML.
and this is not dehumanizing the human.
this is me explaining how learning works.
(notice how this is an emotional argument in the first place, not a real counterargument)
you and me, we can NEVER learn about cats without getting some input about them. we can NEVER learn to draw cats without getting some input about them. do you really understand that?
you are saying that if we do get input of cats, we are plagiarizing/copying that input.
You have already conceded that the output of the statistical regurgitator is completely defined by and dependent on other people's work. Now your only recourse is to dehumanize humans and claim that they are also nothing more than statistical regurgitators completely defined by a set of inputs consisting of other people's work. Good luck selling this one. lol
i already explained to you what curve fitting actually means.
and it's a perfect example for this whole thing. because it's not about other people's individual work, it's about the overarching "curve", the pattern in the entire dataset.
is me saying that machines are learning humanizing the machine, or is it just a simple fact at this point?
is me saying that humans need input dehumanizing?
is saying
>you need to see a cat to draw a cat
dehumanizing?
or is it just a simple fact?
if you've never seen a cat, do you think you'd be able to draw a cat from a different angle by only looking at a single image of it?
what if you also never seen mammals or other animals in general? how would you be able to do it then?
is this entire argument dehumanizing to you? or are you just an illiterate moron?
i'm done here. keep living in your angry delusion.
>psychotic patient babbles about curve-fitting
This guy is actually losing his mind. Either way, he's already conceded that every output of his "AI" is entirely defined by other people's work, so why does he keep posting? Is this an actual, literal bot?
>derailing into inane bait because you'd rather do that than think about what i said
yep, same thing as always.
exact same shit every time. this is why i don't bother anymore.
exactly like i said, you'd rather just believe what you believe, even when challenged.
What you said is fricking moronic and conclusively proves that you have zero technical background, but the thing is that it's not even relevant, because you have conceded my point several times. It's really funny to watch your primitive mind tie itself in knots in an attempt to reframe the simple facts you've conceded as something other than their direct logical implications.
and yet you have not conceded that humans are doing the same thing.
you didn't even provide a counterargument.
your "argument" is that it is dehumanizing. but that's more like an emotional outcry about how humans are supposedly more than this. after all, they must be!
did you know that artists LITERALLY and i mean LITERALLY copy other artists and real life in order to get better at their own art? it's called a study, i've done tons of them.
it's not unsimilar to what AI models do during training. (creating an output that matches the input, keep in mind this is only during training, during inference it doesn't try to copy anything)
>muh technology!
hilarious. you barely understand what curve fitting means. you think it would help if i typed out the specifcs of how diffusion models work?
i used to do that. it's even more of a waste of time than explaining what i'm doing right now.
if you don't want to believe in the concept of generalization, then you just won't.
again, like i said, it is an established idea in ML. one that you subhumans always like to gloss over.
A diffusion mode consists mainly of a variational autoencoder that transforms pixel data to and from a regularized latent space, which allows for easy interpolation of image features by interpolating between their corresponding points, and a denoiser, which can take an arbitrary point on the far outskirts of the main body of the distribution of training data in latent space, and take you back, in small steps, to the most likely starting point inside that main body, under the assumption that you got to your noisy point through a random diffusion process. Prompt-guided denoising just conditions the probabilities for each denoising on the prompt's tokens. There is also a cross-attention mechanism that helps mutually-condition parts of the image that are spatially distant, which helps coherence in terms of the color scheme and lighting. Calling this "curve fitting" is fricking ridiculous, and the main reason I even entertain your moronation is to highlight a striking difference between a diffusion model and a human artist: the former starts from noise and always try to make its way back to the bulk of the distribution, where most of the training data lies, so as to make "art" that is statisically normal, or at least as statistically normal as it can possibly be under the prompt conditioning. Artistic exploration does the opposite thing: it will start from something representative of "normal" art and make its way towards the outskirts where truly original works necessarily lie.
By the way, I didn't read a single fricking word of your botlike shart.
yep, of course you didn't.
i'm actually convinced intelligent people have all moved on or are pro-AI.
to be anti-AI, you either are misinformed or you have to be a moronic mouthbreather like you.
one that screams LALALA I CANT HEAR YOU.
>a striking difference between a diffusion model and a human artist: the former starts from noise and always try to make its way back to the bulk of the distribution, where most of the training data lies, so as to make "art" that is statisically normal, or at least as statistically normal as it can possibly be under the prompt conditioning. Artistic exploration does the opposite thing: it will start from something representative of "normal" art and make its way towards the outskirts where truly original works necessarily lie.
that's nice and all, but how does this confirm or deny your point in that AI is plagiarizing?
there are obvious differences in the process. many that come from the mechanical nature of progressively making art. what the AI is doing is more akin to just the imagining of something, without the putting-to-paper part and all the accidents and inaccuracies that can happen during the process.
>Calling this "curve fitting" is fricking ridiculous
"curve fitting" is what people mean when they call the AI an "statistical model".
or what do YOU think it means to be a "statistical model"?
and now you're going to deny this because it's inconvenient to your misinformed narrative?
again, everything you said about statistics is ultimatively about this. curve fitting.
it chooses the color that fits - based on the PATTERNS it has seen in the training data
it chooses the shape that fits - based on the PATTERNS it has seen in the training data
it chooses the details, the textures, EVERYTHING - based on the PATTERNS it has seen in the training data
it's not stealing images, it has LEARNED from them.
that's really cool and all, but most art is normal art, that's what makes it normal
if anything artists can use ai to learn what is normal, and use that to try to find ways to make something truly new, if that is what they want to do
Very nice explanation. I'll think about this while I'm prooompting.
holy shit, imagine getting mogged like this
samegay
There's no point in arguing with someone who thinks creativity spontaneously combusts from the ether.
>spouts standardized corpoarte zinger
Every time without fail.
Filtered by epistemology.
>reverts to generic reddit one-liner
Every. Single. Time.
>call out the brainlet for the incompatibility of his argument with the basic requirements and limits of ration
>r-reddit!
Just another NPC. Worse yet you're the same fool who degraded himself into pretending to be a pedophile. Truly disgusting.
it's practically impossible for a human to create something that isn't to some degree influenced by things other people have created (you'd have to have grown up in a forest away from any other people your whole life or something), so this argument doesn't really mean anything in the end
>but it doesn't have anything to add itself
how about literal random noise and the prompt given to it, it's influences also by randomness and the operator, so not 100% the training data
>but muh humans are influenced by things
This doesn't dispute, let alone refute, anything I said.
>it's influences also by randomness and the operator, so not 100% the training data
The denoising is 100% defined by the training data.
>This doesn't dispute, let alone refute, anything I said.
i agree that the output of AI is, at some level, derived from it's training data
my point is that this is practically true of human output as well, because humans also learn from the same kind of inputs and each each others' work
or put simply, what's the difference anymore? how can one be good and the other bad when they're become so indistinguishable?
>The denoising is 100% defined by the training data.
doesn't refute what i said
>i agree that the output of AI is, at some level, derived from it's training data
On every level. You start from a semi-random point in a fricking space entirely defined by the training data. Then you use a denoising process entirely defined by the training data, that relies on trying to guess where a point came from if it was from the distribution of the training data, which is defined entirely by the training data. Fricking lol. None of these details are even relevant, since it is immediately apparent from a high-level mathematical perspective that the outputs are defined by the model and the model is defined by THE DATA IT MODELS. How much harder can you fail? The fact that you keep aruing in and of itself cements my point about the nature of your cult.
Humans are defined by a posteriori as well, it's funny to me how hard this is for you to accept. The only thing you've stated is that the process through which AI and humans derive originality is different, which is true, but also doesn't explain why muh AI bad.
You are now executing corporate routine #2:
>If the target knows something about ML, dehumanize the human.
It always goes back to this with your crowd.
>It's a corporate routine because I can't argue against it
It's sad to see such an intelligent person unable to cope with the fundamentals of rational skepticism.
>muh artists are also just statistical models of other people's work!!!
>you can't heckin' argue with this!!!
I don't want to argue with this. I just want to draw people's attention to your corporate-issued anti-human rhetoric and decide for themselves if they want to be associated with this.
Ascribing to a specific school of thought from epistemology doesn't make some "anti-human" or corporate, humans have debated these topics for millenia and their concepts have existed long before capitalist mega-corps ruled. If you founded your argument less on ad-hominem you would convince more people of your ideas, or at least come off as less deranged.
>A human has facilities for analogy and symbolism. A human has artistic intuition. A human has an aesthetic sense.
I fail to see why a human that guides a diffusion model with conditioning to produce a work of art they subjectively enjoy is not employing these same artistic sensibilities you have described.
>A human has the capacity to explore the margins of art in a rational fashion and to trace a sensible path through it into something good, thereby expanding humanity's shared model of art
Most artists end up tracing that path to poorly drawn furry porn and other degenerate slop. What is it you fear exactly? That great artists will no longer be able to do this?
>epistemology ... debated .. argument ... ad-hominem ... convince
My eyes glazed over. Find new buzzwords. Those are permanently in my mental filter. I could mog you about philsoophy just like I mogged your buddy about ML but I'm not that petty.
>I fail to see why a human that guides a diffusion model with conditioning to produce a work of art they subjectively enjoy is not employing these same artistic sensibilities you have described.
They may or may not be doing so. I never claimed otherwise and it has no bearing on any of the points I made.
>Most artists ... poorly drawn ... furry porn ... degenerate ... slop
My eyes glazed over again. You are executing corporate routine number 3.
>slander art and artists.
I'll skip over the obvious guff.
>I never claimed otherwise
So, what exactly is your problem with AI art then? Displacement? Clearly it's not that there's no human sensibilities or intuition involved. Is it that those sensibilities are too far removed?
>what exactly is your problem with AI art then?
1. It's not art (granted, this is not a problem with the images themselves but with the cancerous culture surrounding them)
2. It's thinly-veiled plagiarism
3. "Capitalist" economics will ensure that this new form of plagiarism will be used to harm professional illustrators
4. While most professional illustrators are not producing genuine art, harming them will also harm the minority of them who do produce genuine art
>It's not art
>this is not a problem with the images themselves
Okay, so it is art. You just don't like the people making it. Understandable, many of them are twats who have no appreciation for why models like SD can exist in the first place.
>It's thinly-veiled plagiarism
I think that its ability to produce something which isn't perfectly normal from a statistical standpoint through condition means that it isn't inherently plagiarism. I won't deny that it has the potential to plagiarize, however.
>harming them will also harm the minority of them who do produce genuine art
I think if the minority leveraged diffusion models in their own processes, once they become more advanced, they could produce genuine art in higher quantities, which would benefit everyone. Perhaps, one day, a model which is geared specifically to integrating itself with the human process you've described will be created.
>it is art.
Nope. It is a sample from a statistical model of preexisting art, generated entirely by a machine with no artistic vision or intent.
>I think that its ability to produce something which isn't perfectly normal from a statistical standpoint through condition means that it isn't inherently plagiarism
I think something that can objectively and accurately described as being 100% derived from, dependent on, defined by other people's work, is plagiarism, pure as can be. Most normal human beings with normal human sensibilities would be inclined to agree IF they understood that the premise is true, but most people have no technical knowledge and are easy to mislead by superficial appearances.
>I think if the minority leveraged diffusion models in their own processes, once they become more advanced, they could produce genuine art in higher quantities, which would benefit everyone
Ok, let's see... I mentioned the abuse potential by soulless corpos and you did what?
>If the target doesn't care about ML or art, but hates soulless corpos with their replacement agenda, start rambling about the New Renaissance and how AI empowers individuals.
You see? It's like clockwork. You are executing corporate routine #4.
>generated entirely by a machine with no artistic vision or intent.
Guided by a human through conditioning which leads to works that aren't perfectly normal, which implies some amount of originality and hence can be considered art. I came to that conclusion after reading your explanations.
The rest of your argument is just grouping whatever points you cannot tackle into a boogeyman. If you think that will change people's minds, go ahead. I think the only product from that will be arising agreement from other people who already share your perspective.
>How is higher quantity inherently better?
A higher quantity of original art produced by the minority of good artists is not a bad result in my opinion. It's not like the reflection and passion of that art disappears just because they used AI in the process to help create it, as long as the human is still the dominant factor in the process of its creation.
>a simple commodity rather than an expression of something painfully human
This isn't a problem with AI, what you're describing is a product of unchecked capitalism. We already started treating art like this before diffusion models. Diffusion models are simply accelerating that process. If you want to destroy them because they accelerate, I can understand that perspective, but I don't think it is a proper resolution.
>Guided by a human through conditioning which leads to works that aren't perfectly normal
Ok.
>which implies some amount of originality
Tenuous.
>and hence can be considered art
Complete nonsequitur.
What else is art in this context if not a work created through the employment of artistic human sensibilities? It's only a non-sequitur if you completely ignore the context of everything we have discussed, which is just like you.
How many of humanity's greatest painters were aphantasic, by your estimation? How many enjoyable paintings have you seen created by the congenitally blind? Just how absurdly low are you setting the bar for "visual art"? The human is not creating anything under the "AI" art scheme. just nagging the machine and selecting. If you think your prompts encapsulate the "art" in any meaningful way, just post prompts next time instead of the images and call yourself an artist. Or better yet, commission a drawgay and say his work is yours. You'd still be taking credit for someone else's work, great artist that you are, but at least you'd be helping pay someone's bills while you're at it.
A much more subtle point, which will 100% be lost on you, is that learning how to draw is learning how to see. After learning how to see with your eye, you will learn how to see in your mind's eye, and only after learning to see with your mind's eye can you start about having an artistic vision. People who lack such skills vastly overestimate the coherence and substance of their "visual ideas". They will claim to have a crystal-clear picture of something, but when asked to nail down specific visual details even on the most basic level, they start to mumble.
>After learning how to see with your eye, you will learn how to see in your mind's eye, and only after learning to see with your mind's eye can you start about having an artistic vision.
Do you genuinely think writers have no artistic vision?
If you want to be a writer, be a writer. Don't write prompts and claim you're a visual artist. If you think your prompts are art, post the prompts instead of the images and see how much anyone cares about the only part of your "art" that is actually a fruit of your mind.
Okay so you don't need to learn how to draw to have artistic vision. Got it.
You're getting BTFO so badly all you can do now is essentially ignore the posts that stomp you and hope the thread dies.
Says the guy who stuffs every argument can't counter into a boogeyman.
not him, but i've been impressed before with how detailed and specific prompts can be
no matter what people say, it's definitely a skill in and of itself
How is higher quantity inherently better? If anything an increase in quantity, and the ease of creation, makes art as an end result a commodity like any other that is consumed and discarded. This mentality is inherently negative as it becomes the baseline assumption for what art is: a simple commodity rather than an expression of something painfully human. Excess of quantity by nature creates a culture of devaluation, and if AI art is perceived as art then that will become the assumed standard for what art is. We can look forward to a future scenario where "art" is so cheaply and easily created that it bears no capacity for reflection. Obviously the aforementioned won't matter to you if you're the type that believes art is purely the end result rather than a process, but I believe that mentality is itself a problem. Humans aren't really good at restraining ourselves, even when it would be to our benefit to do so.
Nice post but I beat you to it here:
I'm not sure I agree on the upside bit. As skills to create manually are lost and the reliance on "skill" shifts to prompting (or whatever future form that manifests as), the underpinning mechanisms will be so complex that they will undoubtedly be under the control of large corporate/governmental interests. So while the return to form you mention may happen since there isn't a profit motive, it will happen under a context of allowance by authority (something that is only going to get worse). Art under a heavy pretense of censorship. AI mechanisms will be in a position to steer thoughts and ideas passively in a way that has no accountability. I know this isn't really part of the "art" argument, but they're inherently connected.
I'm not sure I understand you correctly. You're saying that the technical expertise behind organic art will gradually go extinct, forcing people to rely on the controlled medium of machine-assisted illustration for visual self-expression?
What a deluded idea, to have so little faith in humanity and human expression to believe a diffusion model will snuff out the flame of creativity. That is a truly pitiful way to look at the world.
I wasn't talking to you.
way to ignore the main part of my post and focus on one particular way i worded something else
the "at some level" part was not intended to mean "partially"
So you're conceding for the nth time that the output of the model is completely defined by other people's work. Acknowledge this directly in your next post and we'll see where we can go from here.
>completely defined by other people's work
You already stated that conditioning prevents perfect statistical normalcy.
>conditioning prevents perfect statistical normalcy.
This doesn't change the fact that the output is completely defined by other people's work. It just nudges the process of sampling a model of other people's work. You've conceded that the model and its output are completely defined by the training data. Why do you keep going in circles? Be honest, do you have a diagnosed mental illness?
>Why do you keep going in circles?
I'm not the previous person you were arguing with. I don't disagree with that notion at all. I just wonder why you think human originality is not derivative of a posteriori knowledge and what basis you have to believe this. If you believe in free will, just say so.
Humans brains also model data, including what plausible images look like, obviously, but while a human artist can (very roughly) be said to HAVE a model of art, he cannot be said to BE a model of art. As I explained in a previous post, if you think about art in terms of some latent space, all the juicy and original stuff lies on the margins of the data-generating distribution existing art is hypothetically drawn from. In this sense, maybe most of the groundbreaking art of the future is actually included in the latest SD model, but the overwhelming majority of what lies on those margins is utter shit. The further you stray from the main body of the distribution, the deeper you go into a library of babel. The diffusion model, by definition, cannot tell apart barely coherent shit from strikingly original, genius works of art. It's all just low-likelihood points to it. A human doesn't work this way. A human has facilities for analogy and symbolism. A human has artistic intuition. A human has an aesthetic sense. A human has the power of self-reflection, to pinpoint what aspects of avant-garde trash appeal to him, so that they could be distilled and refined in a better form. A human has the capacity to explore the margins of art in a rational fashion and to trace a sensible path through it into something good, thereby expanding humanity's shared model of art. Etc. etc. etc. None of this is comprehensible to adherents of your corporate religion of automatonism, and it's """Rational Skeptic""" (tm) brand of spiteful mutations.
besides the random noise and human-provided prompt, yes, i agree
this is what i meant before, but hopefully this is more concise
my main point before is that how is this any different to how humans make art? humans also take in inputs, form ideas (internal models), then create something from that
there's a difference between simply copying something or a few things directly, like a clone or a collage, and forming a conceptual model of something out of many examples (such as the dog example, where you take many thousands of examples of dogs, and form a mental model of what a dog is that isn't a copy of any particular example you learned from) and making something from that concept
>how is this any different to how humans make art?
If humans are no different, then train your models of Google Earth photographs and on each other's output until you get Dali.
>so as to make "art" that is statisically normal, or at least as statistically normal as it can possibly be under the prompt conditioning.
So through conditioning AI can be limited from perfect statistical normalcy and thus provide some amount of originality. Interesting, anon.
>it will start from something representative of "normal" art and make its way towards the outskirts where truly original works necessarily lie.
The result of either process is through a posteriori knowledge, which is impossible to refute. It is impossible then for a work to be "truly" original, to say so would mean that work is not derivative of anything.
Get a load of this monkey absolutely losing it from rage and jealousy. Not only can he not draw, but he also can't into ML.
it's funny that all anti-ai tards act like you do.
it's eerie even.
i'm convinced that you have to act like this to stay strongly antiAI even when faced with all the facts.
it's genuinely pathetic.
total inkcel death
>puts into dataset tagged as blue hand, wrong amount of fingers, badly drawn hand, bad art, watermarks, etc.
Thank you, human.
You vill work a real job and let the ai do all the fun stuff
>SD requires a lot of technical knowledge and intuitive understanding
Get a load of this professional prompt artisan. Age of the talentless aphantasics is here.
>AI fans denying that the models try to model the data-generating process implied by the inputs
i like seeing people suffer even that is enough for me to support ai development
t. schadenfreude
ai art is as much art as a trans woman is a woman
aka an imposter that is clearly found to be fraudulent when peeking past the surface level a bit
open the box and its just jeet python code
>ai art is as much art as a trans woman is a woman
If only this was merely a shitpost and not profoundly true. You are watching the golem consoomer's entire framework of reality getting redefined in real time to consist of artificial ersatz substitutes.
It's only a matter of time until AI begins to reliably generate images that beat anything humans have ever made. It'll be funny to see how artgays cope with that one
>It's only a matter of time until AI begins to reliably generate images that beat anything humans have ever made. It'll be funny to see how artgays cope with that one
See
>If the target doesn't care about ML, but cares about art, slander art and artists.
You are now running corporate programming routine #3.
You don't think the technology will get any better?
Move on to the next routine, golem. Tell me about the age of AI-empowered human aphantasic no-draw-skills super-artists.
ok I can see that you're in denial. It's ok anon we'll navigate through this just fine
Looks like the corpo-bot decided to bail out of the AI-specific routine and execute 100% generic reddit deflection routines instead.
Artrannies get the rope
ah yes, the plagiarism so "pure" and concentrated that you literally can't link an ai generated image to any set of authors without.... being able to do the exact same shit with 99% of the digital art published nowadays
and what exactly is even plagiarized at that point?
can you plagiarize a color relationship? an angle?, a light setting?
look at pic related. it literally APPLIED its learned knowledge to the original sketch. changes could be made to the sketch and the AI would still be able to interpret it. it could do that to any sketch. to the best of its ability anyway.
what is being plagiarized here? do you think all these versions of that original sketch exist in the dataset, which the AI is supposedly plagiarizing from?
(not to mention that the AI has no link to the dataset after training, merely the learned representations it got out of it)
all these obvious questions that go against your narrative, and you refuse to think about it. refusing to accept the simple truth that the AI is learning, not stealing.
nooooooo you don't understand!!
the ai had an image made by scrimblo bimblo for a shitpost back in 2009 where he used a pastel brush for line art which makes some pixels around the generated lines 0.000001% noiser when "pastel brush" is mentionned in the prompt, but it still looks like actual pastel line art because there are thousands of other artists who did that, this is literally plagiarism and trans genocide
if AI art is so shit
why has no real ''artist'' done a pic that goes this hard