It's super impressive to me, and getting this kind of fidelity is honestly brand new. You can include an image in an AI image generator to serve as an outline. In this case a blank image with bold text. It'll fill in the prompt using the reference as best it can (in this case, a bunch of kittens/puppies).
Prior to a few weeks ago, I'd only ever seen this using a posed figure--basically a stick figure--to generate the prompt's subject around.
I suspect these are using additional tools to guide the AI beyond a simple prompt. For example the spiraling medieval village was generated with stable diffusion and controlnet.
I think the prompt is not much other than "puppies" and "kittens". Major, middle and minor features of the image can be controlled individually in some AIs (they can be differentiated using a Fourier transform or Gauss convolutions and fed into different discriminators) so I think:
major features (scenery) are controlled by the prompt (grass or couch)
middle features (text) are a source image that the AI is punished for straying from
minor features (details) are controlled by the prompt (faces and fur)
Or it's just Stable Diffusion that starts with a text rather than random noise.
Not sure what this prompt was, since I didn't make this one. I linked the site I used above, and they're pretty simple to do but need a few tries to get a good one.
Stable Diffusion together with Controlnet. You basically feed it the text as a black and white image and provide it with a description of the picture of cats. It will then generate this output while using the black and white image as a base. It's fairly simple to do but it can take a while to get a quality result such as this one.
Is this the kind of thing you mean? https://files.catbox.moe/0c97kh.png Getting the text to be more subtle isn't as simple as telling the model, but changing some of the settings got me this and I think it's a lil better
It’s ULTRAVISIBLE as a small icon. If you’re using a phone app that has a compact mode with little preview icons, try looking at it via that. I only use compact mode, so any of these posts are spoiled immediately hahaha
I had a hard time with this too, I didn't even know there were supposed to be words. What finally did it was actually squinting, like so much that my eyes are nearly closed and I can just see a little between my eyelashes. Then it stands out clear as day.
I'm able to read it easily if I squint really hard, to the point that my eyes are nearly closed. Alternatively, it's clear as day when I just look at the thumbnail.
interestingly, the thumbnail gives it away immediately.
Feel like those cats are trying to tell me something.
Especially the two legged one
Would love to gift this to an homophobic person
Full on send it to a professional printer and have it framed, then claim it's a photo of Aintie Rhonda's kitties.
Those kittens have been hanging out with these puppies.
These have to be AI generated right?
Yep, has that signature AI image look.
I am a bit impressed they were able to integrate the message in. I wouldn't have thought an AI to be able to do this yet
It's super impressive to me, and getting this kind of fidelity is honestly brand new. You can include an image in an AI image generator to serve as an outline. In this case a blank image with bold text. It'll fill in the prompt using the reference as best it can (in this case, a bunch of kittens/puppies).
Prior to a few weeks ago, I'd only ever seen this using a posed figure--basically a stick figure--to generate the prompt's subject around.
I think the way stable diffusion works makes this kind of message imprinting quite easy to implement.
the wikipedia article has some good pictures
I would imagine it works by inserting the text on some of the first steps so the ai has the text as it's seed instead of just random noise.
It's that or some kind of next-level art project
What's the prompt to get this stuff done? This looks wild and interesting.
I suspect these are using additional tools to guide the AI beyond a simple prompt. For example the spiraling medieval village was generated with stable diffusion and controlnet.
https://arstechnica.com/information-technology/2023/09/dreamy-ai-generated-geometric-scenes-mesmerize-social-media-users/
I think the prompt is not much other than "puppies" and "kittens". Major, middle and minor features of the image can be controlled individually in some AIs (they can be differentiated using a Fourier transform or Gauss convolutions and fed into different discriminators) so I think:
Or it's just Stable Diffusion that starts with a text rather than random noise.
Not sure what this prompt was, since I didn't make this one. I linked the site I used above, and they're pretty simple to do but need a few tries to get a good one.
How do they make these?!
Stable Diffusion together with Controlnet. You basically feed it the text as a black and white image and provide it with a description of the picture of cats. It will then generate this output while using the black and white image as a base. It's fairly simple to do but it can take a while to get a quality result such as this one.
That sounds pretty complicated. I've just been plugging the text and description into this site: https://glif.app/@fab1an/glifs/clmqp99820001jn0f2xywz250
That's crazy
yoo i made a thing https://files.catbox.moe/egdau8.png
Good one lmfao
yo i'm getting better at it https://files.catbox.moe/og2bry.png
Nice. Ask it to hide it better. Right now you can read it without squinting your eyes.
Is this the kind of thing you mean? https://files.catbox.moe/0c97kh.png Getting the text to be more subtle isn't as simple as telling the model, but changing some of the settings got me this and I think it's a lil better
Yeah, much better.
another one i think went well :) https://files.catbox.moe/rcr9z2.png
That one is really good!
You could literally just Photoshop it and reduce the highlights to be darker which should work
Here's about as hidden as I can make the text https://files.catbox.moe/tkynvz.png
Not really readable unles you know what to look for
This one's amazing
First off you need a mummy cat and a daddy cat
Wrapped in paper mummy or a parent mummy?
A flower mummy.
Copyright infringement, just like how they make any other meme.
Even if this were a meme made of some IP, it would fall under fair use.
Not according to anti-ai luddites
There's valid criticism of the kind of AI made these days, but scraping data isn't one of them.
Lol
these images are so fascinating
These AI pictures look great until you zoom in. There's all sorts of funky stuff going on with the cats' faces.
Top row goes from horrifying to full derp.
Those are flower petals in the right spot to help form the letters.
I'm so saving this
Same, now I have something to reply my wife with, when she sends me smut memes, instead of just emojis like an idiot.
What sorcery is this?
I seem to have a really hard time at these.
I couldn't get one of them without knowing what it should say.
What does it say? I can one or two letters
"send nudes"
Try holding the phone (or computer screen) further away. That's what helps me.
Tried that. I only can see it if I know what words it should be
Try moving the screen around
Ok squinting and constantly scrolling up and down seem to work. Thanks
The lazy way, find a way to unzoom or see the thumbnail.
Can you try over-focusing your eyes? Basically focusing the background while looking at your phone.
You can achieve a similar effect by squinting.
It’s ULTRAVISIBLE as a small icon. If you’re using a phone app that has a compact mode with little preview icons, try looking at it via that. I only use compact mode, so any of these posts are spoiled immediately hahaha
And this is how you find that you are colorblind.
If you’re on mobile, tilt the phone away from you.
On my phone if I tilt it so the screen is at an extreme angle I can see it. If I look straight on I find it hard to see.
I had a hard time with this too, I didn't even know there were supposed to be words. What finally did it was actually squinting, like so much that my eyes are nearly closed and I can just see a little between my eyelashes. Then it stands out clear as day.
I'm able to read it easily if I squint really hard, to the point that my eyes are nearly closed. Alternatively, it's clear as day when I just look at the thumbnail.
The cats are cute and all, but why do I suddenly get the urge to tell you to watch "The Wolf of Wall Street"?
Can't tell if I should make a Wolf of Wall Street joke or a Futurama joke.
You should make both!
"Impressive. And I'm in a library, so no."
Keep up the good work.
Just let us know when you are not in the library.
I think the top left cat might be having a stronk
No that's just peak cat
i'm kind of a cat too, so please do send me also
Alright, but I've only got a couple bucks, so we're going to have to use the small fedex box. I hope you're flexible.
Check PMs
What is the stable diffusion prompt to make these
https://feddit.de/comment/3271240
So you like pad Thai?
Whqt q time to be qlive
The puppies were more subtle lol
Why does squinting your eyes work?! I wanna know!
Stop posting shitty AI generated images
I also hate just posting AI art, but this is really cool!
subliminal messaging is a art