post pictures your computer has generated ( algorithmic art )


is not like other people
"generative" music is a bit of a buzzword in my opinion. you set out some rules and the system makes music according to those rules, it's quite difficult to get the system to rewrite or change those rules. in other words, if you set your generative patch going and go out for a few hours, when you come back it will still be recognisably playing the same piece. so what's been generated? maybe autechre are that gifted though i dunno.
related? i've just played some autechre for my mum and she says it sounds like a computer game without getting to play the game.


Well-known member
i read "the three stigmata of palmer eldritch" by pkd while on holiday - seems pretty relevant:

“we’re going to spend the translated period listening to and watching Pat’s new Great Books animator–you know, the device they’ve just brought out on Terra… you’re surely more familiar with it than we are, Barney, so maybe you ought to explain it to us.”

Barney, dutifully, said, “You insert one of the Great Books, for instance Moby Dick, into the reservoid. Then you set the controls for long or short. Then for funny version, or same-as-book or sad version. Then you set the style-indicator as to which classic Great Artist you want the book animated like. Dali, Bacon, Picasso… the medium-priced Great Books animator is set up to render in cartoon form the styles of a dozen system-famous artists; you specify ”

“which ones you want when you originally buy the thing. And there are options you can add later that provide even more.”

“Terrific,” Norm Schein said, radiating enthusiasm. “So what you get is a whole evening’s entertainment, say sad version in the style of Jack Wright of like for instance Vanity Fair. Wow!”

Sighing, Fran said dreamily, “How it must have resounded in your soul, Barney, to have lived so recently on Terra. You seem to carry the vibrations with you still.”

“Heck, we get it all,” Norm said, “when we’re translated.” Impatiently he reached for the undersize supply of Can-D. “Let’s start.” Taking his own slice he chewed with vigor. “The Great Book I’m going to turn into a full-length funny cartoon version in the style of De Chirico will be–” He pondered. “IJm, The Meditations of Marcus Aurelius.”

“Very witty,” Helen Morris said cuttingly. “I was going to suggest Augustine’s Confessions in the style of Lichtenstein–funny, of course.”

“I mean it! Imagine: the surrealistic perspective, deserted, ruined buildings with Done columns lying on their sides, hollow heads–”

I've also having a go with the VCGAN + CLIP generator thingy, but broadly got disappointing / scary results... I've been trying to feed it input and output images, as well as text prompts and styles.

Here's one of the tree photos i took a while ago, which i thought looked like a leg.


So I put that in, along with the text prompt "human leg" and set it going for 750 iterations and got this video back:

Also tried it with one of my drawings:


the prompt was "man | impressionism"

and i got this:



θερμοδυναμικός καπιταλιστής
Definitely some rich potential here for synthetic intelligence (natural/artificial) poiesis.


Well-known member
this is one of the images of a pavement i took:


i've already been using these as the basis for paintings / drawings and doing my own iterations (full sequence for this one)

so as well as putting the input image in, i added in a "target image" ie something for the nn to work towards, which was the last painting of my own, the final iteration for this paving stone:


and then put in this text prompt:

hills and moons landscape | goya

and got this:



Well-known member
this is the video generated by inputting the tags from the "where is the avant garde" thread ie:

an expensive car driven by a fool | beyond psychedelia | corpsey's law | dribbled by jeremy lin | focus on the nipple | it's fucking vr porn | it's not fucking vr porn | its fucking memes | living information | please abandon this train of thought now | politics is downstream of culture | sublime decay | the g-o-d function | the new gothic | the people | the plasmate | the warrior-philosopher



Well-known member
the exposed stomach on the left in @catalog's video above looks familiar - once you've generated hundreds of these images you can start spotting common biases of the AI

here's one that has no artistic value whatsoever, but you can see the same bloody organs


elm fuck


Well-known member
The other thing is the noses and eyes coming out together. I like the videos cos you see it trying loads of stuff. I'm doing 750 iterations as standard, might go for a lot more.


Well-known member
The other thing is the noses and eyes coming out together. I like the videos cos you see it trying loads of stuff. I'm doing 750 iterations as standard, might go for a lot more.

When I was doing "elm fuck" / "treeshagging" prompts I noticed a lot of disembodied fingers

I've been letting some run for as long as 20,000 iterations...


Well-known member
@linebaugh but seriously, what "title" do you want to render | in the style of Philip Guston?
No idea. Whatever the AI seems to pull off easily. When I tried the render I put 'three men in a car' with 'in the style of Philip guston' and then uploaded a references pic of a guston painting of three men in a car and it still didnt get close. Oddly enough 'glass face on the operating table' was closer


Well-known member
been working on trying to get my face to appear superimposed onto a tree. thanks to @william_kent for helping with it. not quite there yet, but i did get this interesting result which i like cos of the "switch" about halfway through - it's like those videos you see of cells splitting and then there's suddenly some flesh. like the neural net has made this sudden decision.

Added the music from michael j blood's ii tape, cos was listening at the same time


Switch image: