Dreamy AI-generated geometric scenes mesmerize social media users
- Thread starter JournalBot
- Start date
Indeed. But claiming that analogy with photography is wrong based on some superior secret knowledge of generative AI on a tech forum is rather arrogant. It is a good analogy.
Upvote
7
(11
/
-4)
We're already at the point where people use ChatGPT to create Stable Diffusion prompts FOR THEM, so they don't even have to put effort in THAT anymore, either. I really don't see a way out of this mess. How many layers of separation between a human and the end result is tolerable?
What is this going to lead to? Well, human beings have all kinds of inherent limitations built-in. For starters, we have to SLEEP, and we can only do so much in a specific amount of time. Machines don't have any of those limitations, and can also be turbo-charged at infinitum. So it looks like it's going to be "turtles (=AI) all the way down" soon.
Where's the "solution" to this that doesn't trample on anyone who DOESN'T own a farm of H100 GPUs ?
Upvote
-1
(6
/
-7)
I suspect you're referring to Bryce for generating landscapes. I actually bought a copy of that through work just to study the UI which was very different from the norm in that era. Through a long sequence of software mergers and acquisitions it still survives as part of Carrara today.
And Kai's Power Tools? Wow, this brings back memories!
Last edited:
Upvote
8
(8
/
0)
I agree with you that people too easily jump to Jackson Pollock and Abstract Expressionism. However, that's because I think Dadaism is a better example. Dadaism explicitly condoned the Art world, so much so that they called themselves "anti-art". Many of their works can be seen as attempts to mock or destroy the pretenses of the art world at the time. A prime example would be Dadaist poetry, which was explicitly made to be random; one method was to randomly draw cut out words or letters from a bag, and let sheer chance create the resulting "art". Duchamp was notorious for his readymades, specifically Fountain, a porcelain urinal signed and put on a pedestal in Grand Central Palace. The meaning of such a provocation was not lost on the public:
"Whether Mr Mutt with his own hands made the fountain or not has no importance. He CHOSE it. He took an ordinary article of life, placed it so that its useful significance disappeared under the new title and point of view – created a new thought for that object."
-- Beatrice Wood
"The artist is a not great creator—Duchamp went shopping at a plumbing store. The artwork is not a special object—it was mass-produced in a factory. The experience of art is not exciting and ennobling—at best it is puzzling and mostly leaves one with a sense of distaste. But over and above that, Duchamp did not select just any ready-made object to display. In selecting the urinal, his message was clear: Art is something you piss on."
-- Stephan Hicks
Upvote
8
(8
/
0)
True. Knowing the technical aspects of photography is not the creative part anymore than knowing how to mix paint is the creative part of painting. First and foremost is composition, something the camera can't do for you. When a camera can compose a photograph then you have to ask who created the photo.
Upvote
1
(3
/
-2)
At the same time, what's the point? I can walk around covered in GoPros and create hundreds of frames/photos per second that are legally mine. They have no commercial value because nobody's going to look through 99.999% junk and the chance of anything being sufficiently similar to sue anyone for copyright infringement is basically none while the storage/processing cost would be huge.
Just hitting the same RNG so that your photo of "a dog" is the same photo as mine of "a dog" would be extremely unlikely. If your prompt has even a twinkle of creativity the odds of it being in a pre-generated database is starting to resemble monkeys writing Shakespeare. If they do happen to look derivative it's probably because you ran into some kind of fixed, uncopyrightable setting like "wedding couple on church steps" or "sports team photo" who look 99% the same only with different faces.
I mean, you can't even get a machine to play all the guitar riffs and copyright music for the rest of time. What are you going to do with images that have orders of magnitude more variation?
Upvote
6
(6
/
0)
Yes! Thank you, I was totally blanking on the name, that's exactly it.
Bryce was the first instance I can think of where people could generate "gee whiz" images easily, though obviously within a much more limited lens.
Upvote
0
(1
/
-1)
The thing about Dadaism is the work was rarely that interesting, the interest came from the conversation humans were having with the world.
One person dropping a urinal into an art show was a statement. 100,000 people sharing pictures of their toilet is inherently less interesting. Because the point in the end wasn't actually the toilet.
People have paid over a million dollars for 'authorized reproductions' of the urinal. Which is a wonderfully absurd thing.
Upvote
0
(4
/
-4)
It's not secret knowledge if you read meincmagazine.com
Only if we're acknowledging that the AI is the monkey in the analogy.
https://meincmagazine.com/tech-policy...onkey-cannot-own-copyright-to-famous-selfies/
Upvote
-8
(2
/
-10)
It seems to me that what @Aurich and @JoHBE are getting at is the tenuousness between the creative mind and the result. A written prompt even dozens of sentences, using tools like controlnet, using chatgpt to generate the prompt and then using the resulting image to feed the img2img prompt.
The creative input on one hand, while not zero, including picking an output from a group of results (perhaps to further process), and the end image involved such a astronomical number of variables not intentionally chosen by the user of the generative tool.
The creative input on one hand, while not zero, including picking an output from a group of results (perhaps to further process), and the end image involved such a astronomical number of variables not intentionally chosen by the user of the generative tool.
Upvote
0
(0
/
0)
All of this art was made by people, then stolen, tumbled and laundered by these tools.
Never exploring the world of art before doesn't mean this exact practice of using geometric patterns is novel. It's just novel to you.
Upvote
-3
(6
/
-9)
Wonder what Trent, the graphic designer, has to say about modernist artworks. Is he also equally irritated when Magritte, Ernst, Dali, Escher, Braque or Miro, just to name a few, put chimneys where they make no sense or don't get the shadows right? Or is he irritated because an AI managed to generate an image that would take him days to plan and execute, assuming he actually has the technical ability to do so?
Upvote
6
(6
/
0)
Artists, from the most renowned to the anonymous, have been - using your terms - stealing, tumbling and laundering art that was previously done by other artists for millennia.
And when they did so, but added their own personal touch or applied a different technique in the process, the result was oftentimes seen as a major anomaly by their contemporaries and critics instead of an innovation. Meanwhile, humans managed to develop (AI) tools that are now able to create art based on centuries of human output. And we are now in the point in time that everything that is produced by these new tools is considered an anomaly.
Upvote
8
(8
/
0)
Does there need to be a solution?
Anyone who wants to can still do art... I do lots of things because I like them, not because they make money...
Upvote
5
(5
/
0)
Nah, in the analogy, the prompt is the composition.It's not secret knowledge if you read meincmagazine.com
Only if we're acknowledging that the AI is the monkey in the analogy.
https://meincmagazine.com/tech-policy...onkey-cannot-own-copyright-to-famous-selfies/
Upvote
4
(4
/
0)
If it's only not impressive because a person made it using AI tools, then you're not being fair or objective.
Upvote
3
(3
/
0)
Sorry, but I feel like you're being really strict with the definition of 'create' if 'convert random noise into an image' doesn't count.
Upvote
3
(3
/
0)
How is creating a prompt through trial and error, and curating the results, while seeking a specific end result, lacking in intentionality?
Upvote
7
(8
/
-1)
Walks into front door of abode of 80% of the people on this thread to see their art
sees velvet Elvises, beanie babies, pet rocks, Dogs Playing Poker, Painter of Light, Hummels
slowly backs out and closes door
sees velvet Elvises, beanie babies, pet rocks, Dogs Playing Poker, Painter of Light, Hummels
slowly backs out and closes door
Upvote
-5
(0
/
-5)
That doesn't make any sense whatsoever. If I were a bit more arrogant, I would try and claim you don't really understand the comparisons people are making here. There are a few AI 'elements' at play here, the latent space of weighted vectors, generated by scraping and training models. Then there's the prompt and the software that uses the prompt to assemble the results from the latent space.It's not secret knowledge if you read meincmagazine.com
Only if we're acknowledging that the AI is the monkey in the analogy.
https://meincmagazine.com/tech-policy...onkey-cannot-own-copyright-to-famous-selfies/
The latent space is the 'landscape' equivalent of photography. The prompt is the framing of the landscape, as well as tweaks to the framing, 'exposure' settings, and essentially the 'shutter' button. The software interface that assembles the results is the camera.
I'll refine the comparison one step further and reference astrophotography. I can set up my telescope with a camera, and have it point (i.e. prompt) at different random parts of the sky (i.e. AI latent space), and have it algorithmically stack and denoise images (i.e. software interface). Then, I can take a look at all of those images, find one that looks good and tweak the scope position as well as some of the exposure, denoising, and stacking settings. Eventually I get a copyrighted image of something that I didn't 'create' in exactly the same way as not creating an AI generated image. When beginning the process, I would have no particular idea of what images would be coming out of any given telescope position, and it's only afterwards that I would focus on the interesting images and 'adjust' the settings to improve them.
Then, my friend can take a look at my image, decide he wants to tweak something in the scope position and/or image acquisition settings, and generate a nearly identical, also copyrightable astronomy photography with a minimal (yet de minimus) amount of human input.
This process is nearly functionally identical to the process of AI image generation, no monkey necessary.
Upvote
7
(10
/
-3)
What I find bitterly amusing about this is, that a photographer/artist has been replaced by an AI, but the person setting this up is acting as the customer, stylist and editor of the produced work. It feels like it is taking a dig at everyone else involved in the production of artwork that they are not meaningful enough to count as human. I'm wondering how editors are feeling about being being devalued in this process.
Upvote
0
(1
/
-1)
Trial and error is carrying a lot of weight here.
The reason it takes trial and error is because you are not, despite your prompt crafting, actually controlling the creation.
You’re asking for something, and getting something back. But the something you want and the something you actually get are fundamentally disconnected.
This is why the work for hire analogies are the only ones that really make sense to me.
You can commission work to your specifications but you didn’t make it still.
Upvote
-5
(3
/
-8)
This is a really well thought out example.
I just fundamentally reject comparing latent space and outer space.
Your process is your process. The universe is still there despite your efforts to not see it until afterwards.
Upvote
-3
(2
/
-5)
- Add bookmark
- Featured
- #146
It's fascinating, but ControlNet gives a lot of control to the human artist. So whoever said this is where it past the Turing Test should reconsider their interpretation. I'd say this one has more artistic vision from the human side than just prompt generating leading to an output and then selecting the ones that look nice.
It's not like they asked an AI to "make something that will wow me" and it produced this out of thin air.
It's not like they asked an AI to "make something that will wow me" and it produced this out of thin air.
Upvote
5
(5
/
0)
Dadaism is alive and well. Today it is called Avant Garde Art with such installations as a banana duct taped to a wall (replaced regularly due to bananas having a short life when used this way & was even eaten by a viewer once), but the art is the concept not the banana.
Another was a controversy over copying an art installation. Was an aquarium filled with basketballs a counterfeit because it wasn't assembled by the artist who displayed an entire series of variations on basketball(s) in aquarium in major art galleries?
Then there are the infamous urine and crap art installations that crop up regularly.
One I had to laugh at was an installation that displayed a dirty floor. A janitor got in serious trouble for destroying the installation not realising that the mess was Art. Not just one incident as I found more when looking for a link.
https://www.bbc.com/news/entertainment-arts-26270260https://news.artnet.com/art-world/janitor-throws-out-art-installation-347937
The robot janitor done did it!
https://reality13.com/post/janitor-robot-discarded-50m-art-installationThe artist Z who uses rubbish as his medium had this reaction.
The artist's choice of what to display & how is what makes literally anything Art with a capital A. It has nothing to do with the tools. A banana from the supermarket duct taped to wall, or basketballs in an aquarium, all bought in the local mall are Avant Garde Art and are copyright protected.
My personal favourite of all time has to be the "Nothing" art installation. It is an empty gallery and the viewers are invited to admire the concept of nothing.
The original "artist" later sued another artist on the basis of (unregistered) copyright for selling a statue that was invisible because it was made of "nothing". His concept art was copied & sold without his permission!
https://news.artnet.com/art-world/i...18300-literally-made-nothing-1976181/amp-page
https://www.wcjb.com/2021/06/28/creator-invisible-sculpture-sues-italian-artist-profiting-his-work/
And people are complaining that using newly invented tools makes an artist's concept and creative curation not-original? SMH
Upvote
9
(9
/
0)
I am strict. Because the algorithm doesn't make any decision. The outcome of a stable diffusion run is basically predetermined, by both user input and random noise input.
Real creativity isn't possible on such an algorithm.
Upvote
-4
(4
/
-8)
Editors are still needed to some degree, to fix problems in the output of the algorithm. There usually are some, most famously hands.
Upvote
1
(1
/
0)
The Irish Music Rights Organization [IMRO] (RIAA for Irish traditional music) is doing it.
The concern has also cropped up in Scottish, Welsh, & English Celtic circles as the notes, chords, and arrangement are tightly constraining. Any music composed today, is probably a recreation of what another musician played a century or more ago, but copyrights continue to be awarded for "new" arrangements.
The IMRO stance on copyrighting traditional music with centuries of history.
https://imro.ie/faq/is-there-copyright-protection-over-traditional-songs-and-airs/
A scholarly analysis in the Collection of Queen Mary University focusing on UK and Irish law and the effect on the traditional music community. (PDF)
https://qmro.qmul.ac.uk/xmlui/bitstream/handle/123456789/1325/MCDONAGHDoesThe2011.pdf
There is a slight bit of corporate overreach on YouTube (no surprise)
https://thesession.org/discussions/42376
Beware. In US there is a precedent for claiming copyright on the feelings induced by a work even when all else is original. This definitely hits traditional music, where genre is largely the feel of the performance.
https://blog.oup.com/2015/03/blurred-lines-copyright-infringement/
An explanation of how to get a valid enforceable copyright on music that has centuries of history.
https://thesession.org/discussions/22838
Upvote
2
(2
/
0)
You're obviously unaware of collage & mashups. Both are very simple, very common, & and when sold in an art gallery, very expensive.
The result of taking someone else's copyrighted work to tear up to use as your art material or simply defacing another work is most definitely eligible for copyright as "original art"
Design mashup Pinterest
View: https://www.pinterest.com/beyondthebolt/design-mashup/
Art mashup Pinterest
View: https://www.pinterest.com/loramorgenstern/art-mashup/
Collage art Pinterest
https://www.pinterest.com/search/pins/?q=Collage art
Collage: Mona Lisa reimagined (Pinterest)
View: https://www.pinterest.com/pin/69594756734006774/
Paraphrasing your claim to not be art.
"All of this art was made by people, then stolen, tumbled and laundered by these fools."
(Actually, most digital collage artists use tools)
Upvote
5
(6
/
-1)
Just ask any editor who works for a publisher or freelances commercially.
They have a reputation in their industry, but are rarely credited in the published work.
Much of their work is also automated by spell check, grammar check, style check etc. tools.
Even when advising on word choice, writing style, and phrasing, they need to be careful not to turn the result into the editor's version of the author's work.
An editor has done their best work when it is hailed as a beautiful example of the author's work. The best editors are invisible to readers.
As for being devalued when an author decides not to hire an editor? That's the author's choice and you will find these works anywhere self-published books are posted for distribution. Amazon Kindle, Royal Road, Scribd, etc.
Upvote
8
(8
/
0)
Real creativity is introduced by choosing the image library, composing the prompt, tweaking the multitude of settings that effect the outcome, and most important, curating the results to toss anything that is not acceptable to you (Yes, creativity is involved in choosing which images are not representative of your vision of the finished work)
Upvote
6
(7
/
-1)
The faces would still look creepy, I'm sure we'd say "uncanny valley".
Upvote
0
(0
/
0)
There is no any particular requirement of effort for picture to be copyrightable. I can stick a phone out of the window and make a picture not even looking outside. Tada! A copyrightable image.
"How many layers of separation between a human and the end result is tolerable?" - any. Why wouldn't it be?
"Where's the 'solution' to this" - to what exactly? Where is the problem?
"trample on anyone who DOESN'T own a farm of H100 GPUs" - Stable Diffusion works just fine on consumer video cards. Midjourney is cheap. There are places you can rent computer time. There are places you can generate for free, civit.ai is one of them.
It looks like you are very concerned about some non existent problems.
Upvote
7
(7
/
0)
It is probably too late, and “no true Scotsman” etc, but can we please stop calling this and ChatGPT-stuff AI?
Upvote
-5
(3
/
-8)
It is about as productive as saying "Google it".It's not secret knowledge if you read meincmagazine.com
Only if we're acknowledging that the AI is the monkey in the analogy.
https://meincmagazine.com/tech-policy...onkey-cannot-own-copyright-to-famous-selfies/
There is no perfect analogy. By definition it is a comparison of two different things. You don't like analogy between AI tools and camera - perfectly fine. But this is just your opinion, you can't support it with logic and you can't convince people.
No, reading Ars until your opinion magically turn into a fact is not an option. Ars is far from being a perfect place to learn about generative AI - the articel I linked above is a good example of it.
"Only if we're acknowledging that the AI is the monkey in the analogy" - no, not only. There is very limited set of pictures you can get from a monkey, AI tools gives user full creative control, artist's imagination is the only limit. Bad analogy. See, objectively bad, for actual reasons.
That's exactly why you have to "repeatedly made your point clear" - it is irrational and people disagree with it.
Here is a good analogy for you: AI is a tool, like Photoshop. Now that AI is integrated into commercial version of Photoshop it is not even much of analogy - it is a fact of life. Very soon most of the commercially produced images will have AI components. If you think the copyright law in this area is settled you are delusional.
Upvote
9
(9
/
0)
Anton Longshot
Ars Praetorian
Super interesting discussion.
I came to believe the lines between "created" and "generated" will (and often already are) get so blurred as to be impossible to determine.
And I have no idea how to deal with that on the legal level. I'm pretty sure more than a few artists use AI without admitting it.
I shot this picture two days ago. Or not?
I came to believe the lines between "created" and "generated" will (and often already are) get so blurred as to be impossible to determine.
And I have no idea how to deal with that on the legal level. I'm pretty sure more than a few artists use AI without admitting it.
I shot this picture two days ago. Or not?
Upvote
2
(3
/
-1)
This is not an example though, it is a refined analogy between latent space and outer space.
I agree, a really well thought out analogy. Which you "just fundamentally reject" for no reason.
That's why people tend to fundamentally ignore your opinion on this matter.
"The universe is still there" - what does it have to do with anything? Is it supposed to be deep? A random drop of unrelated wisdom?
Upvote
6
(6
/
0)
Upvote
4
(4
/
0)