This was peak generative art from its time IMO. Was available as a screensaver for Windows 98 (can still download from archive.org and works on windows 10). https://www.kurzweilcyberart.com/aaron/static.html
Some samples from the article are similar to what is found on my favorite Generative Art site - Dwitter. It's a social network of coders where we can view, interact and remix live code. Everything under 140 characters of creative JavaScript code: https://www.dwitter.net/top/all
I was inspired by Christiane Baumgartner[0] at my printmaking class - her method isn’t (I suspect?) strictly generative - she takes film stills and then carves them as woodcuts.
So I started doing some prints where I’d use code to create generative art, then use this as the inspiration for my own lino and screen prints.
I really like the idea of code making a thing that is then put through a “human filter” to give different interpretations of the starting point - sometimes rougher and less refined than what I was generating. I think the idea of only having control over parts of the process also appeals to me.
I am a generative artist, but I write my tools in Swift, and the algorithms and math differ from those of other generative artists. In addition to the generative elements, I also include manual image manipulation and digital painting. I wrote an article on my art blog, https://artasartist.com/what-is-generative-art/, that has a different focus than this article and includes what I do.
Put them together with ffmpeg (which is free. You just do a command line command and specify the dir with the image files. You can also specify fps, quality etc)
This article has some good content, but to be honest, I expected more from the title. When discussing generative art, generative music and sound art should also be considered. I've done some research on generative music while developing Glicol.
Simple randomness can create some "embellishments":
Before next token prediction become the only meaning of “generation” generative art was a way of describing art automatically being generated. It never sounded good so we always thought of it as a visual medium.
My whole job is big-model GenAI music stuff and I'm still unclear what counts as "generative." Are loops in Ableton generative? What about that GarageBand drummer thing?
The traditional concept of generative music was that it was different each time and varied/developed/evolved while playing.
So it's generative in the way "generative design" is, but not oriented towards optimising for a goal, if that makes sense.
I don't know about Ableton loops but I've met some drummers like this but they are usually not all that popular. So I am guessing that is not what GarageBand is aiming at, in quite the same way. ;-)
Though perhaps that is generative in the sense of optimising towards a goal; I've not used it.
Eno's work was initially done with the Koan and I think he ended up in a sort of symbiotic relationship with the developers. It was a big deal for a while in computer music magazines in the 90s.
I love generative art and have been recently trying my hands at it (https://circuiter.abhijithota.me) but I'm not sure if I feel positive about it being generally tied to NFTs. I follow a lot of these amazing artists on Twitter and half of them promote their NFTs. I generally tend to ignore that part and focus on the nice patterns made by computers.
Is there no other way to make this a source of income other than NFTs?
It seems to me that the problem you're struggling with is that there is no market for generative art PNGs, which I'd generally agree with. There are a few exceptions, but for the most part NFTs are a gold rush in speculative investing coincidentally often coupled to things which inherently cost nothing to reproduce, like digital images, because for some reason artificial scarcity is an amazingly effective hack for humans.
Making money in art tends to boil down to making something that somebody with money wants, just like everything else. The most successful artists I know recognize it for what it is.
I wonder if there's a market for doing a high-end print-on-canvas of some generated art. But if you manage to add the kinds of margins you see on other traditional art it seems not much different than the NFT variant.
More practically, I bet you could build a decent patronage-style or subscriber-style distribution on top of good generative art, especially if you are releasing often. Incentives like being able to affect prompt or style details could motivate subscriptions or donating at higher tiers.
There are probably other options, like doing the design and tweaking necessary for pulling dozens of images with a consistent theme or character. There's a lot of real work that goes into that process that you could easily justify value/time on. A lot of overlap with existing design roles but more on the craft vs artiste side of things. Depends on what you want to do with it, I guess.
> Is there no other way to make this a source of income other than NFTs?
If you really enjoy generative art I think this shouldn't distract from that enjoyment. It's a way to create unique digital outputs that can be owned by collectors (ie its not just right-click saving PNGs), so in a generic sense it's a natural fit. You don't have to get caught up in a crypto frenzy or spend hundreds of dollars to collect NFTs from a generative artist whose works you really like. The author of this article has many wonderful projects available in a reasonable price range (for art) on both Ethereum and Tezos blockchains.
I wrote about this a few years back on my blog, but yeah, in my opinion the general problem with a lot of digital art (generative or otherwise) is the lack of physical artefacts that come out of it.
You can hone your aesthetic sensibilities and coding skills making art, then apply said skills to commercial work in various design fields, media, advertising etc.
Honest question: why do you feel conflicted about artists monetizing their work through NFTs?
Buyers know exactly what they're getting and artists get a huge cut of the sales (intermediaries usually take a 1-2% fee). They also get royalties on secondary market sales after the initial "print" happens.
I never understood the randomness argument against art. Wasn't there a lot of randomness in Pollock's use of paint on his canvases?
Also, one can extend this thinking to argue that gardens (or mazes) cannot be a piece of art since plants grow by themselves, and it is (to some point) unpredictable. Even in studio photography, there are myriads of random effects - the only way to make a computer image perfectly deterministically is to set RGB to each single pixel manually.
> the only way to make a computer image perfectly deterministically is to set RGB to each single pixel manually.
You don’t need to do it manually. A quick and dirty example is to have each pixel’s colour depend on its position on the canvas. Instant gradient, perfectly deterministic, yet you can still be surprised by the colours.
function setup() {
createCanvas(256, 256);
noLoop()
noStroke()
}
function draw() {
for (i = 0; i < 256; i++) {
for (j = 0; j < 256; j++) {
fill(i, j, i)
rect(i, j, 1, 1)
}
}
}
Play with the fill. Make it, e.g. `fill(i, j, 180)` or `fill(i, j, sin(j) * 200)`. Perfectly deterministic and automatic, yet the result can still be unexpected.
Great article with very nice examples, really tempts me to play around with some of these techniques myself!
I saw an exhibition of Refik Anadol's Nature Dreams last year in Copenhagen, can only recommend to see one of his video installations if there is one near you, they are quite mesmerizing:
https://refikanadol.com/works/
In generative art, randomness plays a crucial role, as it ensures that each piece of art is unique. The art is typically produced by algorithms that incorporate elements of randomness and autonomous systems. This can include anything from computer code to natural processes, mathematics, or even physical systems.
There is not such thing as generative art by definition.
If you describe an activity as an art, you mean that it requires skill and that people learn to do it by instinct or experience, rather than by learning facts or rules.
One thing that is missing from this article is using an image as input to modify how the image looks.
I always found random abstract generative art to feel a little meaningless. Once I added images as input and started modifying image pixels, such as color and positioning, my generative art become a lot more interesting.
It's a short conference video from Tim Holman, where he breaks down the "toolbox" you need. He shows lines, curves, randomness, repetition etc, and how you combine these simple concepts to make something cool. After that short video you basically have what you need to start experimenting.
I'm looking for resources on this too. I recently started working through this book [1], which might be a good place to start. In the introduction to that, the author also mentions this site [2] and this book [3].
Tyler Hobbs writes about generative art, including getting started and improving. Some of his tips are Clojure-specific, and others are more general. I found the essay on flow fields especially interesting. https://tylerxhobbs.com/essays/2020/flow-fields
Dwitter is also a cool place for generative art meets code-golf, which has been talked about many times before on HN: https://hn.algolia.com/?q=dwitter
Honestly I don't feel like that really fits the definition of art, at least how people imagine what the word "art" means.
Procedural generation, yes. But art? I don't know.
If a human makes a thing, would people consider it art? Like a car, a shirt? What if a machine makes something that looks like a painting, would that be art?
If we want to protect art and artists, that would be nice to have a stricter definition of art, in my view.
As a software developer I find artists' obsession with random() a bit baffling. As an artist (a photographer creating visual/fine art photography) I am missing a story and a movement in generative art. There is no human connection, no story, no (however wacky) artistic manifesto that would drive efforts of a group of artists. I am frankly embarrassed when artists talk about coding or the results of feeding the output of random() into an image generator.
Have you read Foley and Van Dam's "Computer Graphics, Principles and Practice"? I think the narrative you're missing could be the history and theory of computer graphics research, and the recognition of the incredible ongoing contributions of artist-scientists in this field.
Photography has a deep and storied history and terminology and humanity emerging from experiments with a medium... That's what gives it life, and gives us things to discuss as photographers. But computer graphics has that same kind of story - thousands of research papers, experiments, and the work of millions of artists using and advancing the tools.
Generative art is a subset of computer graphics and is intertwined in its history. It is far more than artists playing with random numbers, the same way that photography is far more than pointing a sensor and pressing a button. There is an endless history of past attempts to build on, and huge amounts of mathematical and artistic theories to build, in both cases.
If you need help to see this as an art, consider reading Foley and Van Dam, and work to see the development of computer graphics techniques - and their application - as a form of art in itself, where we develop and explore experimental media and subject, without the constraints of physical reality. I believe that by working with graphics algorithms, as an artist and coder, you will see how the artistic and technical beauty of these algorithms combine to create a beautiful and compelling form of art.
I hope that after this, you will get endless joy from contemplating computer graphics, the way you get joy from photography. And I hope it will give you enjoyment of the more abstract and experimental side of things, and even the work in image AI, which challenges our identities and livelihoods as artists. Imo it's a very amazing time to be alive as an artist!
That was one of the first books on computer graphics. I loved it.
I will admit that as a software dev and a former 3d modeller and animator I may be too close to how the digital paint and canvas are made to get excited. Another reason is the infinite number of works that can be produced by machines running the same algorithms, which has an unfortunate effect of turning artworks from an event to mass-produced objects (or renderings on LCD screens). That mass-production aspect removed barriers to entry to become an artist, see the deluge of "abstract art" on Etsy and hundreds of tutorials on how to produce it on YouTube. There just isn't anything special about it. I understand that this is not new, painters are still salty about photography and photographers are not happy about digital distribution of their works on a mass scale without compensation. And I know that mass production of artworks is not new; in the past, master painters or engravers ran workshops that produced works "in the style of" too. Still, there was some effort required to produce such works. I think the perception of generative art suffers from the same thing that makes it possible: ease of access to a lot of compute power for next to nothing.
I spent over 10 years of my life working with 3d graphics and finally went back to photography and recently to film photography. I feel more involved in the process of creation and there is a physical artifact at the end of the process. Once it is produced it needs no power to enjoy it unlike generative art that needs a screen to be seen (some of it can be printed in 2d or 3d).
As for AI, I see it as a threat to all who invested their time and effort into creating business models that support them in the digital ___domain. That value is being taken away from them now. Technology applied to the creative space has always been about improving efficiency of the creative process and offering new tools, but this time round there is an assumption that all that we create is fair game for the AI companies and that we should accept it while they monetize it. There is no model for the creator to make a living from his creations in the AI space and I think that's dishonest.
The model for creators to make a living after AI hoovers everything up is simple: it's called revolution. You pull up on the wealthy person, and you take ownership of their AI, and make it free and open source for all. Then the AI production system can provide for any quality of life needs, and you no longer have to do any art for any reason other than you enjoy it.
Remember, art is fundamentally worthless bullshit that we do because life is miserable and terrifying, and we're obsessive miserable bastards. There's no reason to make it a living, except to enslave artists to the whims of the wealthy. We can just take from the wealthy, instead.
Seize the means of production from capitalists, and you'll be free to be make art without the added misery of having to please bastards with too much money and no real love of the process of art. But continue defining art as having to produce revenue and support the living of artists, and you're just fighting for continued slavery at the hands of the wealthy. I say seize the means instead, using extreme measures if necessary.
I find this view completely mystifying. The image generator that somebody piped random() into didn't fall off a tree - a human made it. They took pure thought-stuff, as Fred Brooks put it, and instead of shaping it into an image they made an image generator. Do you imagine that person had no human connection/story/manifesto/whatever? Or their human x/y/z disappeared because they used code instead of a paintbrush, or what?
(Aside: this view seems doubly odd coming from a fine-art photographer. If you've been doing it long you'll surely have met folks who scoff that "it's just clicking a shutter button". Are you not doing something similar here?)
I understand the feeling. i don't know how to feel about some contemporary art either. I'm sure there' some story behind this but it also looks like randomly generated art to me, even though I like it
I treat abstract art as a source of inspiration (colour palette, pattern design, composition, technique). I prefer when it is made by a human working with a physical medium, because that's how you achieve mastery. Yet another Processing app able to churn out an infinite number of random images does nothing for me.
Why? Some art is purely representative, some is abstract, some is narrative driven, some is not.
There is 'chaotic' abstract art - like Pollock - and regular examples - like Escher or Vasarely - and many points in between. Why should using a rand function and a line/point shape be fundamentally different?
It's definitely one of the best examples of well-executed generative art that stands well above the spray hose of crap produced on a daily basis. Thank you for sharing.
I mean, the person using random() would have to define how the output from it actually goes on to generate an image, thats their choice and their influence on the art
I remember early days building some of this stuff with Flash. Good times. Back then, you had to do quite a bit of work to make it yours, and it felt fresh and risky.
Now, digitally-generated images (including photography) is totally ubiquitous.
Generative art is awesome! But it isn't Art art.
My thinking has come around to the idea that Art art is an exclusively human-to-human transaction, and must be imbued with the soul of its creator; the intangibles of style and taste, the ineffable of lived experiences, the errors and happy accidents and imperfections.
Both are art. But Art art can never be replaced by machines, and that's a good thing.
But the creator does have influence on the art, they created it, they are just combining it with an extra hit of randomness. I've done some of this stuff in the past, and errors, happy accidents and imperfections absolutely come into it. They're not just sending computer input directly to output, they are having their own influence and control over it
I respectfully disagree. At the end of the day, generative art is the same thing as googling an image to search for the end user, in as your aren't creating anything, but rather, you are asking for something to be produced for you based on typed or entered criteria. You are a patron waiting for the engine to produce something for you.
If you want to say there is an art to prompting, that's fine.
To add a bit more precision to my previous reply: You're interpreting "generative" in the way that AI companies have tried to hog the term, which happened only in the last 2-3 years. It's the hyped definition, not the historically rich definition.
Generative art as a general concept has a long, long history, that we can date back at least to performance events like Musikalisches Würfelspiel in the 1700s.
I respectfully disagree. let's think from a first principle point of view, what is the definition of art in the first place? But whatever the answer is, art is just a way of structuring/representing a mix of multiple arts into one single art, and it's not limited to painting arts. historically people have been taking inspiration from nature, other people, the environment, and more. That's what an AI does as well, but with much more creativity than a human.
I did. It's trying to play off randomness as art, hence the generative part, and trying to subtly relate that to AI through use of abstract artists like Pollock.
It's also not new. If you've taken any modern art history class before, this is article/website is the same content. Every generation of art/artist tries to justify their existence to the last.
