Your Kid Couldn't Paint That

Sorry Karen, but your kids couldn’t paint that. The pros make it look easy, but it takes more than a smartphone and a false sense of expertise to actually be good at things that require skill. Or does it? It turns out that there’s a growing community of people generating digital art using artificial intelligence [insert “always has been”].

One of the more intriguing developments on this front is text to image generation which is exactly as it sounds: you type in some words and the robot painter starts working on your masterpiece. After messing around with some GUIs that were hit or miss, I located ThE sAuCe which is a Python codebook. No programming required, just have to change one parameter, i.e., the text that you’d like to generate an image of. This may seem too obvious, but I went with “Picasso Power Ranger with glitter” for my first run. Here are the results.


This is some truly horrifying stuff right here and a giant Harold burger based on other results I’d seen using this technology. Luckily, this was only the first image it produced and I quickly realized that it wasn’t finished. By default, the code is set to spit out an image every 50 iterations.

download (1)

(100) Still a horror show.

download (2)

(150) Ok now that’s pretty interesting and artistic. Maybe we’re going places.

download (3)

(200) We’re going places.

download (4)

(250) Possibly the Louvre.

download (5)

(300) Not better or worse, just different. It’s riffing effortlessly.

download (6)


download (7)

(400) Could be imagining things but I’m starting to see a Power Ranger take shape?

download (8)

(450) This is where I got spooked. From this point up to #2300, the changes between images are subtle, so I’m going to jump to the final image in the sequence.

download (46)

(2300) And there it is. Picasso Power Ranger with Glitter. I am completely mesmerized and humbled by this robot’s ability to create elite art.

*Edit: If you’re new, try one of these two simple GUI interfaces. You really just need to change the text prompt to whatever you want it to paint:

You can upscale your image to a higher resolution with AI here:

There are many forks of this now, but the original implementation was by advadnoun and then this Google Colab codebook by Katherine Crowson in the link below. *Those early versions may be a bit dated at this point depending on the type of image you’re trying to generate. The VQGAN+CLIP method seems to be inferior to CLIP-guided diffusion (posted later in thread) for photorealism.

You need to run the sections of code in order using the play button, and some of them take a while. Under the SETTINGS FOR THIS RUN section, change the text between quotes in prompts = ["the first day of the waters"], to the text you’d like to create an image of.


Tips: Using an artist’s name or writing “in the style of [artist]” seems to work well if you’re going for a certain look. If you pick something too esoteric, it may get stuck and not have a reference image for your thing (some of my other attempts were huge fails). Some people report that adding “unreal engine” to the end can generate rendered scenes with more realism. Also, there are additional operators you can use like | to separate fields but you should read documentation on those.

Here is a list of more implementations of the algorithm combo that produces the images:

Let’s see if anyone can beat Picasso Power Ranger with Glitter in this thread.


download (3)

“Hillary Clinton mustache painting in the style of Magritte”


This is pretty fascinating. Trying to do it on but getting errors. Will definitely give it a shot tomorrow on my pc though!

‘Nobleman in front of his cottage’

Iteration 250:

Not much cottage or nobleman to speak of…

Iteration 350:

Now a figure seems to appear center bottom; looks more like a cleric than a nobleman though…

Iteration 450:

In an unexpected turn of events, the cleric turns out to be the legs of the nobleman!

Iteration 650:

This is clearly Henry IX! :face_with_monocle: Does this symbolize him rising above the Pope?

Iteration 1950:

Now he looks like Friedrich II, “Der alte Fritz”! :smiley: My greatest heroes in one painting!

Iteration 3100:

Decided to quit here; does the program stop on it’s own at some point? Anyway, the cottage looks more like a gate house, but the nobleman mostly looks the part except of the turkishness going on with his legs. All in all not bad for a soulless machine! :slight_smile:

1 Like

Lol wow. The best part is how it seems like it will be nothing then makes a major turn. Afaik the program does not stop on its own. It keeps trying to minimize that loss function which has extremely diminishing returns at some point and you have to stop it yourself. However, there’s a limit to how much you can drain GPU on the free Colab account, so it will stop you if reach that amount I suppose.

“Ecce Homo by Lovis Corinth, but with a clown instead of a doctor.”



The singularity has been breached.

Seriously can somebody make one with that title?

I bought the Pro subscription so I’ll try it next, probably on one of the other implementations.


The Singularity Has Been Breached. (2021)


The Singularity Has Been Breached #2 (2021)


I would use this but honestly I’d just make a bunch of really weird porn.



Wow. I like that a lot.

1 Like

I meddled a bit with those AIs where you enter a sentence or two, and they try to continue the text in some way. If you slipped just a whiff of something raunchy in the seed text, like “nipple”, or “flesh”, oft than not, if would go into high eye brow area. And if you started with more explicit text, it would produce the weirdest sex nightmares.

1 Like

Ha, I didn’t see yours and am trying one “by Rene Magritte”

Can’t wait to see it.

1 Like

Can someone ask it to ape a painter who’s really strong on form and colour eg Modigliani?
Or one who’s strong on mood eg Turner?

These seem too tough for AI imo.

I ran out some Basquiat portraits of Sinema a few minutes ago:


That middle one is great because it put the Medusa snakes on her head and I didn’t even tell it to. The computer KNOWS.


Didn’t look particularly cool. Now it’s starting over, for some reason. (I don’t know what I’m doing with this, but one of the implementations Lawnmower_Man posted did a good job of holding my hand)

I am currently using the one called AI Art Machine.

This thread reminds me of this NFT collection, which seems like it was doing AI art based on CryptoPunk images. Its price is back in the gutter now but it was trading at like $1.5k-$2k floor for awhile.