This is so fucking cool. Technology is truly amazing. Definitely a thing people in the future will take for granted though. Littke shits.
Anyway, who even needs artists anymore now? We're getting to the point where we'll let AI create art.
https://deepart.io/
>>8018365
The intersection of the set of all art and the set of all AI-created stuff is [math]\emptyset[/math].
>>8018373
The Op image was created by an AI
Just like how calculators and computers replaced mathematicians, right?
http://slither.io/
AI can only follow a set process, smart programmers in the future will make robots with problem solving and learning algorithms that can allow them to seem near-human, but the one thing a computer will never do is have an original idea. The illusion of creativity in AI is only possible through random values.
>>8018531
Nice armchair faggot
>>8018548
think of counterargument then, nigger
>>8018578
Counterargument to what? Nothing you said means anything, it's just a bunch of shit you pulled out of your ass. Try reading a book.
>>8018600
Well I'm a programmer focusing on AI, do I have to explain it in more detail?
>>8018380
An artist needs to be sentient being which understands the concept of art, and is capable to appreciate their own work as such. So far, AI certainly isn't at that stage. For that bot that created that image, it's just a matrix of values which it arrived at by munching some algorithms, and nothing more.
>>8018403
Maths is a lot more complicated than art
>>8018427
Go ahead post it
If I'm not mistaken deepart still uses the older implementation of this work. There has been an improvement published a few months ago using CRF rather than gram statistics as the 'style statistic' which works much better but takes much longer to optimize.
>>8018631
This, I won't accept a computer as an 'artist' unless it can at least look at pictures and accurately describe the contents, which is still yet to be achieved
>>8018531
>The illusion of creativity in AI is only possible through random values.
Isn't that the same as humans?
>>8018620
Hahaha yeah and I've got a PhD in Machine Learning also I've got a 6ft dick and my dad could beat up your dad, do I have to explain it in more detail?
>>8018670
>accurately describe the contents
That's what deepart pretty much does though to understand the style of a picture to copy it.
Really the only thing missing is deepart mixing styles trying to find its own style
>>8018631
So you"re saying a piece of work created by an autist who is not really aware of his artwork is not real art? No matter how unique and beautiful it is?
>>8018670
http://googleresearch.blogspot.com/2014/11/a-picture-is-worth-thousand-coherent.html
This research is two years old so you can bet it's gotten a lot better. Kill yourself.
>>8018681
>That's what deepart pretty much does though to understand the style of a picture to copy it.
You are a moron and you have no idea how any of this shit works.
So many retards in this thread.
>>8018685
I think that art must be created conciously, if it's the result of a random and involuntary process, it's hard to classify it as such. You probably wouldn't describe a mountain range or objects in space, or even anything that was made by an animal as art, despite any possible quasi-artistic qualities that their view might have.
>>8018689
It literally breaks the image down into content and style to work out how images are supposed to be drawn in the style. How does it not fill the criteria "accurately describes qhat's in the picture"?
Maybe you shouldn't choose so poorly arbitrary and retarded definitions to decide what art is
>>8018706
So far it still requires human input, so doesn't that still make it art?
>>8018689
Oh bullshit, that article is theory and speculation, if the technology exists it has not been released publicly
>>8018713
I guess it depends on your interpretation of art... would you consider a program that generates a random chord progression and a melody to follow it to be an artist if all the user does is tell it what key and scales to use?
>>8018728
If it's beautiful, sure. Art is entirely subjective. Public consensus is the only thing that decides what art is and what's not. Reminder that during Van Gogh's life time everyone thought his artworks were pieces of trash
so i dunno the sauce of op image but its been trained on van gogh right
and then fed a picture of einstein
i mean at best thats a slightly jazzier version of sampling
no great shakes, art has been fucking around with convolution forever anyway
>>8018736
The source link is literally in the OP
>>8018738
right so you upload as training image ('style image' as they call it)
and then an image to transform
this is literally photoshop-filter tier
>>8018734
then I guess the argument is over whether it requires creativity to be an artist, this is sure to be the headlines in 10-15 years
>>8018380
Photo was taken by a human. Style was defined by Van Gogh.
>>8018742
Except that it breaks down the style of one image and the contents of the other and recreates the image in the new style.
It goes far beyond a mere filter and requires deeep processes.
You don't seem to grasp how much of a feat that is
og
>>8018761
no i definitely do
but what i'm saying is that it's useless without the input data ie vincent van gogh
>>8018714
Are you literally retarded?
Google "automatic image captioning" to find mountains of this stuff. Here's something from this year:
http://arxiv.org/pdf/1603.02814v1.pdf
>>8018764
The same holds true for humans. Theyy just mix a lot of input data.
Blind people can't draw pictures either
>>8018734
>Art is entirely subjective
I've heard enough of this.
>>8018772
i get that it can endlessly spit out pics that people might think were van gogh
altho to me it looks like it was trained on his still life and landscapes
but whatever, its possible for it to imitate an artist i like, and that's incredibly cool
but that's all it's doing, whereas van gogh had a unique style which he came up with (mental illness notwithstanding)
which is infinitely more valuable (there are already tons of human van gogh copycats as you alluded)
>>8018776
Then give me an objective value of art.
For all we know everything we consider art today might be considered poor trash tomorrow.
>>8018670
can a person accurately describe all the contents of a picture? i don't think so. he or she can on a certain level describe most of the content, but not everything.
![cba5e36dd3417dff4002dbe4115a2f46[1].jpg](https://i.imgur.com/KthsTTD.jpg "cba5e36dd3417dff4002dbe4115a2f46[1].jpg")
Oh shit I found /g/
>>8018781
What makes you think subjectivity even exists?
>>8018838
Now it's getting retarded
>>8018670
>it can at least look at pictures and accurately describe the contents
Lol. Show https://deepart.io/ "art" to https://www.captionbot.ai/ . Keks aplenty will surely be had.
>>8018886
>>8018631
You mean sapient. A sentient being is not sufficient condition to create art, but a necessary condition.
>>8018676
One could argue that point. In the same way that there's no such thing as a completely original song as everything stems from inspiration. everything is a product of its environment to put it simply. We can only choose what to do with our influences.
Wow. It figured it's a "painting".
>>8018365
>who even needs artists anymore now?
Well I am not an expert but I know deepart.io for example needs input from actual artists before it can do anything meaningful. It also requires some level of artistic aesthetic to decide which buttons to press and which images to upload. Overall the site is just another tool for artists to use, master, and subsequently discard when it stops making unique and interesting pieces.
>>8018915
This. Even ancient raster graphics processing programs like old Photoshop or Paint Shop Pro had stuff like "artistic filters", which were really a much more rudimentary version of the processing that deepart is doing to the processed images.
>>8018922
We really don't. Neuroscience and psychology are relatively new compared to other branches of study. Consciousness in definition is still not really understood. Is it own thing or the result of all the mechanisms currently active within our mind? I think it's the latter but would like to know for sure some day.
>>8018863
For all you know the universe could be totally uniform, all matter contiguous.
For all you know nothing could be said to have existed at all, since science cannot prove being to be qualitatively different from non-being, in much the same way that it cannot alone distinguish between life and non-life.
True, people see things differently, but why do you think that this fact means there cannot be an objectively knowable reality?
My answer to your question is that whatever exists attains an objective quality via both its participation in contingency as the single principle of existence and its participation in the intellect of the one observing it.
Mathematics is abstracting from matter, but artistic knowledge (intuition) is abstraction from sense-perception, hence the tastes of a highly adept artist may approach a metaphysical ground in being itself, inasmuch as being is ordered to potentiality and is therefor called good.
>>8018950
So if our consciousness experience were like a radio frequency there could be one main station frequency that could be considered the 'one ' as in reality there would be objective.
>>8018959
wut
>>8018963
>>8018670
image recognition can easily recognize most stuff by now. I just picked a random free software.
>>8018922
Well, there was that whole pre-scientific era of lots of superstition, philosophy, and religion so we have a pretty firm baseline to measure from.
If you brain can formulate a mind that conceives of what the brain is truly capable of then you have become the very thing that is capable of comprehending what a brain is capable of. If we can't become that comprehensively self-aware, then it stands to reason that we're physically incapable of actually knowing.
It is at this exact point that you realize the question won't make sense until you can express it as a self-replicating Turing machine.
And I don't think you're the next Von Neumann.
>>8018636
lel
>>8018989
>fashion
>>8018986
Why is Microsoft ai so obsessed with Giraffes?
>>8018365
It changes the pixels are makes it blurry. Truly amazing.
>>8019548
You're retarded. Like no joke, that's the level of intelligence of my dumbest facebook friends
>>8018365
Think this is interesting. You could probably use a similar construction to create a generative "art model" that is capable of simulating (to some degree) the process of inspiration in the mind of an artist.
You would train the network to be in the 'state' of an artist's mind (by using his exhaustive works) and then allow it to generate sequences of art ad infinitum. This would be interesting because it could give us a visual 'hint' as to the kind of artistic images that might flash through an artist's own mind as he creates art. Similarly, we might be able to feed such a network a set of images that act as inspiration, giving us further hints about how the world informs creativity.
Very intredasting application of deep learning.
>>8018365
So this is probably a retarded question - but what makes this "artificial intelligence?"
Are they saying that the computer is making it's own decisions as how to create the picture? One of the first AI's as far as I remember was Deep Blue. It was able to "think" and "plan" in real time to defeat a chess champion.
This seems to be a new type of AI, I guess? It's "thinking" about how to create a picture? What makes it so much more cutting edge and revolutionary to Deep Blue? Is it just the complexity of what it's doing?
>>8019575
Not an argument
>>8019673
It's pointless to argue with you because you don't understand what's happening on a fundamental level
>>8019633
it is specialized in a different field than deep blue
>>8019633
Deep Blue is a chess program.
This is a program that analyzes the style and contents of a picture
>>8019690
Lmao why are you so defensive about this program? It's a photoshop AI kid, nothing more.
>>8019821
You're completely unaware what technological feat that is, are you?
>>8019923
it's garbage AI
>>8019944
>He didn't realized that he has been talking to an AI program all along
>The AI program was trying to defend his artist friend
Truly the end of human kind.
hmm
>>8019995
I think we have the solution to the war on terror.
>>8019944
By what standards? No program has ever been able to do this before
>>8018365
Can't wait till all my hentai doujinshis are redrawn in the original artstyle!
>>8020055
I am hentai bot, I am make shitty art of your animu
Feed me
>>8018755
Tree was created by god
Pencil by factory number 5
Hurr durr
So this is really just like an advanced image-rendering process, right? It is software that can convert an image into what looks like a painting using a selected art style?
It sounds similar to an image upscaler called waifu2x, which has to guess and make up new detail every time you upscale. This site does it all for you if you link it a picture:
http://waifu2x.udp.jp/
You can download the software for that and upscale images locally if you want; it's pretty fast if you have CUDA. I wonder if you would need expensive hardware to use something like deepart, or if one could slowly convert images on their own PC?
>>8020355
It seems like you can run it yourself. It uses CUDA and is written in Python
https://github.com/andersbll/neural_artistic_style
THIS IS NOTHING!
You guys wait, in a couple of years these neural networks will be able to generate meshes for objects from photos and then generate the textures which can be artistic in their own right.
Want more? They will probably be able to animate them too. Oh boy, I can't wait for the butthurt from game artists.
>>8019085
unconventional clothing and lots of skin?
looks like fashion to me. Not the kind of fashion you wear but the kind of fashion designers are all over about
>>8020311
/sci/ has a knack for overreacting to "breakthroughs".
Honestly this shit isn't impressive at all.
>>8020605
yeah i know, fashion recognition isn;t a thing yet
it will be though
job interviews are going to be awful in the future
>>8020311
Chess isn't that complicated. It's got a limited amount of moves and if AI didn't beat it we'd simply do it by brute force simply making computers powerful enough.
Just recently a human was beaten in GO which is much more impressive.
And the ability to break down art into its components is a HUGE step. There were and are a lot of people that believe that art is what would forever seperate a human from a computer, but we aren't that far away from AI creating its own, completely original art.
>>8020747
Maybe if you're below average intelligence. Just as the OP said, this is technology future kods will take for granted not realizing what a huge step it is.
>>8020467
You know what will be cool? Video Games that are automatically created and expanded by AI. Like Sword Art Online
Oh shit the future is gonna be awesome
>>8018365
>Estimated wait time 6 days
kek
>>8021398
Better create your artistic memes now
>>8018531
Humans never have original ideas either. Anything you imagine and think is creative and new is just a recombination of things you have already seen. Maybe these recombinations of already given information can be so complex that it seems like an original idea, but it never is.
What kind of garbage website is that, that they actually milk such a small insignificant gimmick for money?
>>8021803
Show me the results. I doubt they'll look as good
>>8020272
But he's right you mong
Even if you gave it words and it spit out a masterpiece, a human still wrote those words :^)
>>8019633
Deep Blue didn't think and definitely couldn't plan anything without the help of a human assistant
This is AI (at least the modern definition of AI) because it uses convolutional NNs
It isn't revolutionary, but the fact that such an application for them was found and that it works is interesting enough
>>8020976
open up photoshop and use the magic selector tool to select a person's face, then delete the face and select the option to use the smart fill
better yet, select a person's clothes and use a sketchbook filter on it. If you aren't amazed by that then you are of low intelligence because it is the same idea -- image filtering.
>>8022571
Go ahead. Create a Van Gogh painting that way
>>8022622
I'm not seeing any science in your post.
>>8022631
spotted the faggot
>>8022631
I'm not seeing you draw a Van Gogh
>>8022633
I'm not seeing any science in your post.
>>8022639
spotted the faggot
>>8022494
Can't, I don't have Linux so I cant install all the dependencies. All the guides on installing them are meant for Linux filesystems
I don't see why it would look worse, though. It's literally the same thing. Waifu2x looks the same when I render using my own card as it does when I get them to render it through the website.
>>8018365
Why the fuck does it take so long
>>8023072
computer needs time to paint
>>8018365
a neural network making photoshop filters is not art
>>8018365
>Who needs artists anymore?
Nobody as of the late 1800s.
>>8023285
that is pretty amazing
>>8018365
"Estimated waiting time: 7644.0 minutes"
>>8023285
The network does not understand depth?
>>8018365
Time to Pepe this shit up
>>8018365
but that's not art
that's like graphic design or something
there's no artistic input
>>8018886
you think that's state of the art? lol kill yourself you fucking gorilla
machines are BEYOND human-level at image classification. imagenet error rate is like 2.5% now
>>8026570
So are you, yet you pretend you're smart
CALL ME WHEN IT CAN MEME
I wanted an Auschwitz image in style of Family Guy but it tells me to way 3 days
imagine the pornibilites
>>8018631
A human is nothing more than a bunch of neurons, why would it qualify as an artist ?
>>8023285
webm is damaged :(
I am trying to install neural-style on windows using Visual Studio (with Anaconda 2.7 plugin) and I cannot for the life of me get it set up right.
When I get to the part of my python module where I try to import caffe, it says:
[code]
File "caffe\pycaffe.py", line 10, in
from ._caffe import Net, SGDSolver
ImportError: DLL load failed: The specified module could not be found.
[/code]
if I can import caffe I should be able to use neural-style on my own GPU
>>8018365
Just one more day till my pictures are finished!
>>8018365
Do you realise that a lot of people create art because they specifically enjoy the act of creating? Sure an AI can shit out paintings too but that's only relevant from a business standpoint, it won't make people quit art as a hobby.
>>8026840
You're forgetting about the soul m8
>>8018365
The AI mimics, it's incapable of actual creativity.
>>8022088
For that to be true the thing ever imagined by a human would have been an original idea though.
>>8027366
first thing*
>>8027365
I don't think we're really that far off. Soon you can feed AI lots of different artstyles and it'll start creating its own art
>>8027371
No, it's just humans experience of nature recombined.
Human sees stone
Human sees stick
Human make spear
>>8027379
What will an AI possibly create "art" from? Where would an AI find "inspiration"? What separates "art" from non-art to a machine? Do Androids Dream of Electric Sheep?
>>8027394
When programming AI the question is always "how do humans do it?"
Just make an AI do the same thought process that an artist does when creating art.
The problem is rather our loose definition of art.
Are the images deepart creates already consideres original art? If not, why not?
Does an artist who draws a banana not create original art just because he depicts something that already exists?
Neural networks are exciting. It's like a totally new and different way of using computers
Could we use neural networks to create software that can reason? Like I mean where you could give it a scenario and it would use its "neurons" to simultaneously consider countless possible responses, narrowing them down based on its pre-programmed criteria until it finds the one response that is the most "reasonable?"
One problem see with this, you can only make art based off of old art styles and as a result it maxes out as a genius art student in college but never can pioneer truly unique ideas or methods.
>>8027382
Yeah but how did they come up with idea of using some sort of natural dye to paint recreations of things they saw? Surely that is an original idea?
>>8018772
>Blind people can't draw pictures either
This was painted by a man blind since birth.
>>8024761
>Pepe
Got to wait 3 days for my Pepe van Gogh. Shit sucks.
The robot learned van goughs style just like a human art student and applied it to a photograph just like many artists do
The AI created this image.
>>8027612
>>8018365
>https://deepart.io/
>applying fancy filters is now art
I bet you fucking faggot think that tracing is also art.
>>8027624
This is /sci/ everyone here is autistic, not artistic.
>>8018365
L'IA peut créer 'le beau'. Mais il ne pourra jamais faire de l'art ou être considéré comme un artiste.
Cet a dire s'exprimer a travers une création, poser ses émotions et ses expériences de tel façon a ce que d'autres humains puissent l’interpréter, ressentir et compatir.
'Le beau' n'exprime jamais que de la beauté, c'est un travail de forme et qui plus est relatif a chacun.
Sorry for the french I'm not able to tranlate this my english is too poor.
If someone who can understand what I wrote and can translate it may reply the translation, it would be nice.
>>8027510
Bullshit. At least not without help. How would he even know watermelons are red?
>>8027785
This information is easy to acquire even if you are blind.
>>8018636
>Jackson Pollock
>>8027451
The Neural Network boom was big and everyone was shitting there pants like you forever ago. Then they realized it didn't work well for anything other than rudimentary categorization. And then a few years ago they just made them bigger and "deeper" and now we're shitting our pant's again.
>>8027963
Can you link any functional from that "forever ago"?
What kind of achievements did you get?
>>8028263
https://en.wikipedia.org/wiki/Artificial_neural_network#History
People thought we could basically recreate the brain after we had a mathematical model of a perceptron and everyone was quite excited like everyone here. Then they realized they couldn't solve the exclusive-or problem and everyone abandoned the idea until they figured out how to solve that problem. After that everyone was excited again and made a few sentence parsers and rather nice categorization techniques. Then it stabled until they figured out "deep learning" in the 2000's and now we've hooked up a trained "deep neural net" to some picture filters and voila this is the result.
AI research is so unstable. We all blow it out of proportion and the public thinks "Holy shit it's conscious" when all we can really do is make it give us neat graphs or separate a sentence into verbs and nouns and all that and the public is unimpressed and funding slows down. Then, "Look! We can make pretty pictures" and the public, again, shits itself at the "Awesome potential."
I'm not saying this isn't a step forward because it is but it's very modest step forward which, when added to the existing techniques will help us achieve what could be possible but this isn't much farther from where we started with a mathematical model of a perceptron.
>Tl;dr It's nifty but not that big of a deal. We still have a long way to go. Calm down, everyone.
>>8028346
>wikipedia
>>8028346
I'm not going to scour the internet for books and research papers from the 50's through the 80's to prove a point to you on a Syrian refugee board. Wikipedia is a fine starting point to branch out from. Stop by your local college library and look at some old books about AI research if you don't believe Wikipedia.
>>8028370
https://cs.stanford.edu/people/eroberts/courses/soco/projects/neural-networks/History/history1.html
http://psych.utoronto.ca/users/reingold/courses/ai/cache/neural4.html
Does this please your autism? Reputable enough for you? Otherwise go check out a book on the subject but your favortie reputable author.
>>8028346
>it's very modest step forward
I think we've had a pretty huge step forward; not so much the fact that we are using it to create art, but the fact that we've got a framework and a system now where anyone who has an average nvidia card can run a powerful neural network on it to do tasks that would've otherwise been impossibly slow on traditional processing.
We are at the point where neural net computing is actually becoming accessible to average people. Now anyone can code their own neural net program if they have the time, understanding and dedication to do it. Sites like deepart help bring exponentially more attention to the field of NN processing, and all those new minds will bring with them new ideas and novel uses for the technology
>>8028400
>stanford
>canada
lol why do you americans even try to act like high IQ people?
>>8028522
>that would've otherwise been impossibly slow on traditional processing.
That's thanks to improvements in actual computer systems and not AI research itself.
>We are at the point where neural net computing is actually becoming accessible to average people. Now anyone can code their own neural net program if they have the time, understanding and dedication to do it.
We've had a few systems for creating neural networks already now. But, like you said, it has been restricted to actually computational power.
I agree with you that utilizing the parallel processing in video cards is a great addition to the arsenal.
>bring exponentially more attention to the field of NN processing
Yes, that was a lot of what I was saying. AI is heavily reliant upon appealing to the masses and over playing the technology we have.
Perhaps their is something here I'm missing but this thing seems to simple create an internal representation of the picture and using a fancy "trained algorithm" to manipulate it in such away that it looks like a painting. Not too far removed from the older work that could look at a sentence and parse it into it's many different parts. It just looks nicer, which, attracts the public and then after a short while the public will be uninterested because all we can do is train this thing to mimic different artists styles just like we could train the earlier work to parse different languages.
>>8028551
>because all we can do is train this thing to mimic different artists styles just like we could train the earlier work to parse different languages.
Bleh, give the tech a little credit. From what I understand it analyzes the original picture tries to rebuild it out of distinct shapes it finds within the style image, both on a large and small scale. In doing this it works through a bunch of "layers" of image processing (exactly what happens in those layers I couldn't say).
However it works, it's pretty impressive and it can do a lot more than just imitate existing artist styles. Since a "style " is just a picture, there are literally infinite possibilities in how you can style one image. For example, RMS with a picture of a circuit-board as the style: >>8018819
>>8028598
But it's the same basic principle extended. It's not novel and it's merely an extension of decades old processes. Rather than make and train one layer they've made many, many layers and trained them each uniquely and then together to produce a more intricate outcome. That's all deep neural networks are. It's no more revolutionary than, say, adding another battery to an under-powered utensil. We've had the idea we just hadn't applied it well enough. It's merely a step forward that looks awesome to the layman. It's not a revolutionary leap forward anymore than screws were to nails.
Cool
>>8028782
Oh I downloaded the thumbnail
bump for future pictures
>>8028814
original image?
>>8018403
Nice bait
i'm so fucking happy that we're finally destroying the fuck out of humanities
when do we get something similar for writing styles?
like, you write something about ANNs and it changes your writing style to look like shakespeare wrote it
>>8031722
Are you retarded
argh how do I install CUDArray onto Anaconda in windows? I keep getting problems with functions taking too many arguments when I run setup.py after compiling libcudarray.lib
I think I need an old version of CUDnn maybe, but you cant get it anywhere except possibly nvidia's site after joining their developer program
I find computer-generated music even more impressive.
https://www.youtube.com/watch?v=wd__bsIDih0
didnt anyone request a picture a few days ago? post it
>>8031681
>>8033539
>just another shitty search algorithm
But isn't intelligence just about search spaces?!
>hurr durr muh optimization curve trajectory curve
>>8033550
>isn't intelligence just
stopped reading there, not interested in cognitive science memes
>>8019546
It's a stand-in for "i cannot comprehend lineart"
>>8033411
bumping for more meme images, I just got this one back
style is from picasso's "artist and his model", 1963
>>8034122
original
![The Artist and his Model [1963].jpg](https://i.imgur.com/ruKnYAj.jpg "The Artist and his Model [1963].jpg")
>>8034122
style
>>8031979
This is pretty relaxing, I am enjoying it so far.
AI generated everxpanding video game world when?
>>8034133
try looking up stochastic music on youtube, there's some pretty cool stuff out there
https://youtu.be/s1FFxqPWOcI
>>8031979
David Cope Emmy Vivaldi
https://www.youtube.com/watch?v=2kuY3BrmTfQ
Holy shit, this is amazing.
>>8034126
I appreciate art of many forms, but that's a completely terrible painting with negative redeeming features. How could this possibly be "great art"?
>>8034160
it's late era picasso, it's not to my taste either but I picked it because it's a distinct, "advanced" style that would be difficult for a computer to imitate.
>>8021712
>>8034170
i bet the computer can imitate that crap but the real question is, why waste the computing power?
>>8035464
what was the template?
>>8018631
Complete and utter bullshit. If this is the definition of art, then can random person can tell from most of the "modern art" BS that the people responsible didn't know the first thing about the concept of art.
Art is an artificially crafted setting or work, that does not normally exist in the natural world. And to be perceived as art, it needs to be seen as such by enough people in either number or accountability.
In general terms, anything can be art, as long as the correct number and quality of people accept it as such. In more personal terms, art is anything you, subjectively, accept as such.
Such as, I've actually seen "art" that was literally a pile of turd with a cocktail rainbow stuck on it. That was not art, period. The OP's image is objectively a million times more artistic than that.
>>8018365
Since I define art as human expression, the art that robots create is counts as human expression too since it was a human who created the tool we call "artificial intelligence".
>>8018365
I see a lot of posts saying the "AI" was successful with recreating a cubist style image when it was not. Cubism, by it's conception was meant to illustrate a literal interpretation of space as seen from every possible angle. What it does is just copy the stroke variation, places it based on shape recogntion and then copies the hue chroma values onto the subject. The same goes for other movements and styles, unfortunately it will always be a pale immitation.
So did it produce a plausible copy? yes
Did it reproduce the human ability to create and think? of course not, although I doubt that was ever the goal
Did it succeed in creating enough buzz to impress pedestrians, and in turn investors? yes
>>8036841
I see evidence that this AI is reproducing the human ability to create and think and you're claiming it is not. Where is your evidence.
>>8036305
