I'm looking forward to the legal battles that will ensue from AI generated CP
Eventually someone's gonna fabricate evidence of rape too from a program like that which is gonna be its own unique problem. Just keep generating new content until it matches your face.
I don't know why everyone seems to think it would have to be trained on actual CP. It already knows what a clothed child looks like and it already knows what naked adults look like. Even though there are some differences physiologically it can still get close.
transformers still produce abominations with all the money and datacenters google and meta have, and the best datasets money can buy. i seriously doubt a current gen image generator could consistently make decent fappable material with a severely gimped training data, but i'd be glad to be proven wrong
You act like the GANs that produce this content are sentient and understand things like "clothed" and "nude" and "child" and "adult". They don't. The discriminator would need a collection of real CP in order to develop the capability to help the generator improve. Take a moment to learn how generative adversarial networks are trained.
something doesn't have to be sentient to understand that and the latest models aren't even using GANs dumbass. Get out there and go learn stuff yourself before you try to correct others based on your own imaginary scenarios.
>the latest models aren't even using GANs dumbass
This thread is about DALLE, which does, in fact, use multiple GANs. You be a retard somewhere else, anon.
8 months ago
Anonymous
nope
8 months ago
Anonymous
DALL-E uses a combination of gpt-3 text embedding, AttnGAN, CLIP, and STackGAN++. Just STackGAN++ is multiple GANs by itself. You are a moron.
8 months ago
Anonymous
nope
8 months ago
Anonymous
>If I just keep saying "nope", maybe people will think I was just pretending to be retarded
>the legal battles that will ensue from AI generated CP
at some point in the distrobution pipeline, a real child wouldve had to been photographed while naked, thus making any form of AI generated CP still depend on a criminal activity. as a shitty analogy, packing weed into a bong isnt by itself illegal, but at some point you wouldve had to buy or grow the weed, which is illegal, and it can be reasonably deduced that you would smoke it which is also illegal.
until an AI can dream up entirely new things on its own, there cannot be any ethically or legally sourced 3d dicky. remember that it isnt a victimless crime, even a "willing" child still cannot reasonably consent to the activity, and i shouldnt really need to explain that.
That reasoning could also be used to ban drawn CP. Even if you aren't using a real child as a model for your drawing, you still had to train your own neural net using imagery of real children or using imagery which was created based on real children, etc... Somewhere down the line. Why is it any different when it's an artificial neural net instead of a real one?
you cant reasonably prove that a e-boi artist has a stash if CP to reference from, especially since most of them dont. the human mind can say "i know what a child looks like, and what nakedness looks like, i can put those together".
no MLA we have right now can do that, because nakedness is a concept not a noun or verb.
were you a child once?
as far as im aware, using yourself and/or your own memories as reference material is not illegal provided you arent remembering a time you broke the law.
if these answers seem retarded, its because this is a retarded strawman and you know it. generating AI CP requires a stash of real CP to train from. drawing e-boi doesnt, especially since you can "train your model" on other e-boi art, or otherwise just imagine stuff. did all those people who drew aliens have real aliens to train from? do you think the 3d artists behind world of warcraft saw real orcs to use as a reference?
just as you viewing CP is contirbuting to the industry that is abusing kids, using a MLA to generate CP contributes, in fact even more so because hundreds or thousands of images wouldve been needed to train it.
MLA cannot make something entirely new, the human mind can (within certain limitations thay i shouldnt need to explain, >inb4 you cant think of a new color)
do you really want to argue about ethics, morality and human nureology/pyschology or are you just a pedo that cant admit to it? theres no need for shame, just dont bring this shit up if you have no idea what youre on about and dont really have anything meaningful to say
8 months ago
Anonymous
normal humans cannot remember exactly what their genitals looked like at that age, let alone those of the opposite gender.
8 months ago
Anonymous
bait
8 months ago
Anonymous
female vulvas look exactly the same at every age, save for the hair that appears at around 10yo
most men like them clean shaved for a reason
8 months ago
Anonymous
spot the larper, I think you forgot where you're posting.
8 months ago
Anonymous
>if these answers seem retarded
yes. artists don't become decent at human anatomy until around 16 yo. no such thing as precocious genius artists. and you don't learn art using your mind as reference, let alone a decade old memories, imbecile
8 months ago
Anonymous
>do you really want to argue about ethics, morality and human nureology/pyschology or are you just a pedo that cant admit to it? theres no need for shame, just dont bring this shit up if you have no idea what youre on about and dont really have anything meaningful to say
ironic that you're pulling this card while accusing someone else of being a pedophile based on nothing except a fantasy in your head
8 months ago
Anonymous
the scientific term for what he's doing is called projection
8 months ago
Anonymous
>were you a child once? >as far as im aware, using yourself and/or your own memories as reference material is not illegal
I don't think most people have vivid enough memories of their own genitals as children to perform this task with accuracy. >did all those people who drew aliens have real aliens to train from?
They most likely did not base their drawings on real aliens, but they still based them on imagery they trained their neural net with. Whether they are trying to make something unlike what they are familiar with or trying to make something similar to what they are familiar with, they still need a data set to get started. >just as you viewing CP is contirbuting to the industry that is abusing kids
I don't think this is objectively true. Paying for it would certainly be contributing. Producing it would be contributing. Distributing it would be contributing. Leaving comments like "Great stuff. I'd love to see more of it" in CP forums would be contributing. Downloading in situations which display a public counter of how many people have downloaded it might even be considered a form of contribution. However, I do not think the act of viewing the images is actually unethical.
8 months ago
Anonymous
>I do not think the act of viewing the images is actually unethica
what about fapping to them?
couldn't the cia agents watching your webcam take it as some kind of support to the industry?
8 months ago
Anonymous
>couldn't the cia agents watching your webcam take it as some kind of support to the industry?
that's just a reddit teenager thing, no cia agents are actually watching your webcam.
this being said, there are millions of magazine photos of child models for underwear brands. while personally i think that shit shouldnt be allowed, it is technically legal and in most cases done with parental consent (another thing that shouldnt really be allowed, "yes please photograph my child mr not-a-pedophile, im sure she wont regret it in 15 years") so you could AI generate sinilar images.
but again it still isnt really ethically viable as you are still needing real kids to train the model on, which is a big no-no
There are already AI models that detect CP. So you can set up and adversarial network, as the VP model already exists in a sanitized state. It's like clown room design.
you're going to wait a long time, nobody is going to spend $400,000+ to train a model on something that will have them pay 50 times that in legal fees.
ok, here's the danbooru image database including nsfw results.
feel free to train a model using it and then release it to the world
https://www.gwern.net/Danbooru2021
people have used it for normal anime, see:
https://thisanimedoesnotexist.ai/
or the more "artwork" like from https://twitter.com/_aixile/status/1542287395776876544 which is going to be a paid product after they're done shilling it for their pointless game.
they probably have access to the boobies but they will not release it to the public because they don't want their names associated with it.
I hope some horny graybeard sets something like [email protected] especially for that. Millions of coomer GPUs training the most important piece of AI ever created. Then put it on blockchain or P2P, so it can't be ever destroyed, and the world is a better place.
Before any company can work on the generation of sexually explicit material there is the issue of getting access to sufficient training data.
There are pornographic data sets but those are not freely available; to access them you need the backing of an institution with a detailed explanation of what you want to do with the data.
I've been thinking you could maybe just scrape videos from redgif and extract the frames but then your data would maybe have too much autocorrelation.
There are also legal issues.
Many jurisdictions do not bother to differentiate between live-action pornography and synthetic pornography - or they explicitly criminalize certain kinds of pornography even if they are synthetic.
If someone manages to generate something that looks like illegal pornography using your product, does that mean you distributed illegal pornography?
I would argue that distributing a pre-trained model for generation is more like distributing software like GIMP but I wouldn't want to be the one that tries this defense in court.
>If someone manages to generate something that looks like illegal pornography using your product, does that mean you distributed illegal pornography?
No. You are retarded. The person who used your product would be guilty of distributing illegal pornography (assuming they actually distributed it).
Honestly I can see someone making the case that the information regarding what makes up illegal pornography is encoded in the weights of a neural network, and that therefore a pre-trained model contains illegal pornography.
From a technical standpoint I would disagree that this is the case since generative models typically also take random noise as input but trials are unfortunately decided by laymen.
Technically you can generate images completely at random and still have a small probability of generating CP. Also, if you had some monkeys typing for an infinite amount of time, they would eventually produce the steamiest pedo novels. So really, the truth is that we need to ban monkeys, typewriters, and random number generators.
That's still stylized.
Also generating humans is way harder. We are much more sensitive to uncanny valley when it comes to human faces, proportions, skin, hair, etc.
I don't think you understand that Dall-E intentionally had face generation gimped because they wanted to make sure they could control celebrity faces and such.
i swear i'm being completely honest, transparent, and candid
please explain me with as much detail as possible your porn prefferences, so i can understand your viewpoint, and we can both together make the world a better place for future generations
8 months ago
Anonymous
Just plain ol solo female amateur porn. It does not matter to me at all if one of her eyeballs is slightly more teardrop shaped than the other.
8 months ago
Anonymous
Take this image for example. It doesn't matter to me that her nose is a little bulbous and blurry or that her hand is mangled. This is content generated by a computer in whatever style or scenario you want and we're still in the relatively early stages with incredible breakthroughs in computing power and training quality coming out every few months.
8 months ago
Anonymous
a "slightly off" face is interpreted as sick for humans. you don't want to fuck a sick whore, or even be close to her. it's genetics
https://i.imgur.com/0q4zCbR.jpg
Take this image for example. It doesn't matter to me that her nose is a little bulbous and blurry or that her hand is mangled. This is content generated by a computer in whatever style or scenario you want and we're still in the relatively early stages with incredible breakthroughs in computing power and training quality coming out every few months.
a bitch with that nose would make you puke irl. i guess you could scroll her face and imagine she looks like prime jenniffer o connelly instead, but why would you do that when there's plenty of porn with 10/10 qts on the internet?
is like you want to punish yourself unconsciously because deep inside you see sex as something sinnful and immoral or some shit
8 months ago
Anonymous
I take back
I can't tell if you're trolling or not.
it's incredibly obvious now.
8 months ago
Anonymous
how rude. who's gonna defend fapping to ugly deformed bitches now if you can't even articulate some bullshit argument?
A company would never do it because the porn industry is too profitable. Content produced by an ai is not copyright protected so they can't stop piracy
I'm looking forward to the legal battles that will ensue from AI generated CP
Eventually someone's gonna fabricate evidence of rape too from a program like that which is gonna be its own unique problem. Just keep generating new content until it matches your face.
only the CIA has enough CP to train an state of the art image generator, and they've been banning competing collectors for decades
I don't know why everyone seems to think it would have to be trained on actual CP. It already knows what a clothed child looks like and it already knows what naked adults look like. Even though there are some differences physiologically it can still get close.
transformers still produce abominations with all the money and datacenters google and meta have, and the best datasets money can buy. i seriously doubt a current gen image generator could consistently make decent fappable material with a severely gimped training data, but i'd be glad to be proven wrong
You act like the GANs that produce this content are sentient and understand things like "clothed" and "nude" and "child" and "adult". They don't. The discriminator would need a collection of real CP in order to develop the capability to help the generator improve. Take a moment to learn how generative adversarial networks are trained.
something doesn't have to be sentient to understand that and the latest models aren't even using GANs dumbass. Get out there and go learn stuff yourself before you try to correct others based on your own imaginary scenarios.
>the latest models aren't even using GANs dumbass
This thread is about DALLE, which does, in fact, use multiple GANs. You be a retard somewhere else, anon.
nope
DALL-E uses a combination of gpt-3 text embedding, AttnGAN, CLIP, and STackGAN++. Just STackGAN++ is multiple GANs by itself. You are a moron.
nope
>If I just keep saying "nope", maybe people will think I was just pretending to be retarded
>the legal battles that will ensue from AI generated CP
at some point in the distrobution pipeline, a real child wouldve had to been photographed while naked, thus making any form of AI generated CP still depend on a criminal activity. as a shitty analogy, packing weed into a bong isnt by itself illegal, but at some point you wouldve had to buy or grow the weed, which is illegal, and it can be reasonably deduced that you would smoke it which is also illegal.
until an AI can dream up entirely new things on its own, there cannot be any ethically or legally sourced 3d dicky. remember that it isnt a victimless crime, even a "willing" child still cannot reasonably consent to the activity, and i shouldnt really need to explain that.
That reasoning could also be used to ban drawn CP. Even if you aren't using a real child as a model for your drawing, you still had to train your own neural net using imagery of real children or using imagery which was created based on real children, etc... Somewhere down the line. Why is it any different when it's an artificial neural net instead of a real one?
you cant reasonably prove that a e-boi artist has a stash if CP to reference from, especially since most of them dont. the human mind can say "i know what a child looks like, and what nakedness looks like, i can put those together".
no MLA we have right now can do that, because nakedness is a concept not a noun or verb.
no the human mind can't
child genitals are very different from adult ones.
what if i want to fap to adult bodies with little girl faces?
would that be legal and nice?
maybe? I have no idea how that's related to the previous two comments at all.
sorry, i didn't read the whole thread before quoting your shitpost, retardo
quite the influx of redditors lately
it takes one to know one?
were you a child once?
as far as im aware, using yourself and/or your own memories as reference material is not illegal provided you arent remembering a time you broke the law.
if these answers seem retarded, its because this is a retarded strawman and you know it. generating AI CP requires a stash of real CP to train from. drawing e-boi doesnt, especially since you can "train your model" on other e-boi art, or otherwise just imagine stuff. did all those people who drew aliens have real aliens to train from? do you think the 3d artists behind world of warcraft saw real orcs to use as a reference?
just as you viewing CP is contirbuting to the industry that is abusing kids, using a MLA to generate CP contributes, in fact even more so because hundreds or thousands of images wouldve been needed to train it.
MLA cannot make something entirely new, the human mind can (within certain limitations thay i shouldnt need to explain, >inb4 you cant think of a new color)
do you really want to argue about ethics, morality and human nureology/pyschology or are you just a pedo that cant admit to it? theres no need for shame, just dont bring this shit up if you have no idea what youre on about and dont really have anything meaningful to say
normal humans cannot remember exactly what their genitals looked like at that age, let alone those of the opposite gender.
bait
female vulvas look exactly the same at every age, save for the hair that appears at around 10yo
most men like them clean shaved for a reason
spot the larper, I think you forgot where you're posting.
>if these answers seem retarded
yes. artists don't become decent at human anatomy until around 16 yo. no such thing as precocious genius artists. and you don't learn art using your mind as reference, let alone a decade old memories, imbecile
>do you really want to argue about ethics, morality and human nureology/pyschology or are you just a pedo that cant admit to it? theres no need for shame, just dont bring this shit up if you have no idea what youre on about and dont really have anything meaningful to say
ironic that you're pulling this card while accusing someone else of being a pedophile based on nothing except a fantasy in your head
the scientific term for what he's doing is called projection
>were you a child once?
>as far as im aware, using yourself and/or your own memories as reference material is not illegal
I don't think most people have vivid enough memories of their own genitals as children to perform this task with accuracy.
>did all those people who drew aliens have real aliens to train from?
They most likely did not base their drawings on real aliens, but they still based them on imagery they trained their neural net with. Whether they are trying to make something unlike what they are familiar with or trying to make something similar to what they are familiar with, they still need a data set to get started.
>just as you viewing CP is contirbuting to the industry that is abusing kids
I don't think this is objectively true. Paying for it would certainly be contributing. Producing it would be contributing. Distributing it would be contributing. Leaving comments like "Great stuff. I'd love to see more of it" in CP forums would be contributing. Downloading in situations which display a public counter of how many people have downloaded it might even be considered a form of contribution. However, I do not think the act of viewing the images is actually unethical.
>I do not think the act of viewing the images is actually unethica
what about fapping to them?
couldn't the cia agents watching your webcam take it as some kind of support to the industry?
>couldn't the cia agents watching your webcam take it as some kind of support to the industry?
that's just a reddit teenager thing, no cia agents are actually watching your webcam.
>t. CIA agent
People know what naked children look like because they were children once too. A pedo AI wouldn't and would have to be fed images of children.
the ""reasoning"" could be used to ban children, because at some point they must be naked irl
this being said, there are millions of magazine photos of child models for underwear brands. while personally i think that shit shouldnt be allowed, it is technically legal and in most cases done with parental consent (another thing that shouldnt really be allowed, "yes please photograph my child mr not-a-pedophile, im sure she wont regret it in 15 years") so you could AI generate sinilar images.
but again it still isnt really ethically viable as you are still needing real kids to train the model on, which is a big no-no
There are already AI models that detect CP. So you can set up and adversarial network, as the VP model already exists in a sanitized state. It's like clown room design.
you're going to wait a long time, nobody is going to spend $400,000+ to train a model on something that will have them pay 50 times that in legal fees.
Just make it hentai
ok, here's the danbooru image database including nsfw results.
feel free to train a model using it and then release it to the world
https://www.gwern.net/Danbooru2021
Why hasnt anyone done it?
people have used it for normal anime, see:
https://thisanimedoesnotexist.ai/
or the more "artwork" like from https://twitter.com/_aixile/status/1542287395776876544 which is going to be a paid product after they're done shilling it for their pointless game.
they probably have access to the boobies but they will not release it to the public because they don't want their names associated with it.
Because it's hard, labor intensive, and expensive and to top it off you can't even put it on your CV.
I hope some horny graybeard sets something like [email protected] especially for that. Millions of coomer GPUs training the most important piece of AI ever created. Then put it on blockchain or P2P, so it can't be ever destroyed, and the world is a better place.
>you're
TELL ME YOUR AN IGNORANT WITHOUT TELLING ME YOUR AN IGNORANT!!!!!
LOOOOOOOOL!!!!!!
Before any company can work on the generation of sexually explicit material there is the issue of getting access to sufficient training data.
There are pornographic data sets but those are not freely available; to access them you need the backing of an institution with a detailed explanation of what you want to do with the data.
I've been thinking you could maybe just scrape videos from redgif and extract the frames but then your data would maybe have too much autocorrelation.
There are also legal issues.
Many jurisdictions do not bother to differentiate between live-action pornography and synthetic pornography - or they explicitly criminalize certain kinds of pornography even if they are synthetic.
If someone manages to generate something that looks like illegal pornography using your product, does that mean you distributed illegal pornography?
I would argue that distributing a pre-trained model for generation is more like distributing software like GIMP but I wouldn't want to be the one that tries this defense in court.
>If someone manages to generate something that looks like illegal pornography using your product, does that mean you distributed illegal pornography?
No. You are retarded. The person who used your product would be guilty of distributing illegal pornography (assuming they actually distributed it).
Honestly I can see someone making the case that the information regarding what makes up illegal pornography is encoded in the weights of a neural network, and that therefore a pre-trained model contains illegal pornography.
From a technical standpoint I would disagree that this is the case since generative models typically also take random noise as input but trials are unfortunately decided by laymen.
Technically you can generate images completely at random and still have a small probability of generating CP. Also, if you had some monkeys typing for an infinite amount of time, they would eventually produce the steamiest pedo novels. So really, the truth is that we need to ban monkeys, typewriters, and random number generators.
how many more dalle threads do you plan to make today, you shilling spazz
it's literal advertisement at this point, but mods are so fucking useless
All current porn is CGI. You've been lied to.
Go to /aids/ on /vg/.
Image generation is going to be utter shit for long time. But GPT based AIs are already amazing for generating smut and not.
>utter shit for a long time
uh anon? what year are you living in?
Generating heavily stylistic art is way more forgiving than realistic porn that will not freak you out.
>he doesn't know
That's still stylized.
Also generating humans is way harder. We are much more sensitive to uncanny valley when it comes to human faces, proportions, skin, hair, etc.
???
human faces were perfected years ago https://thispersondoesnotexist.com/
>https://thispersondoesnotexist.com/
creeptastic
you may not like it but that's the ideal woman form
I don't think you understand that Dall-E intentionally had face generation gimped because they wanted to make sure they could control celebrity faces and such.
want to play a game? https://twitter.com/triplux/status/1542529379485396995
i hope so
(◍•ᴗ•◍)
you ever seen a kid eat a frog?
>noooooooooooo the ai can't generate humans!!!!!!
well it can't. there is not a single image where it got _all_ parts of a face right, in most cases there is something wrong with the eyes.
I don't often make eye contact with the subjects in porn, I know some people are different and do that but it's really not a big deal to me.
face is 99% femenine beauty
no idea what kind of degenerate crap you're into, but for most males a deformed face is kind of a deal breaker
I can't tell if you're trolling or not.
i swear i'm being completely honest, transparent, and candid
please explain me with as much detail as possible your porn prefferences, so i can understand your viewpoint, and we can both together make the world a better place for future generations
Just plain ol solo female amateur porn. It does not matter to me at all if one of her eyeballs is slightly more teardrop shaped than the other.
Take this image for example. It doesn't matter to me that her nose is a little bulbous and blurry or that her hand is mangled. This is content generated by a computer in whatever style or scenario you want and we're still in the relatively early stages with incredible breakthroughs in computing power and training quality coming out every few months.
a "slightly off" face is interpreted as sick for humans. you don't want to fuck a sick whore, or even be close to her. it's genetics
a bitch with that nose would make you puke irl. i guess you could scroll her face and imagine she looks like prime jenniffer o connelly instead, but why would you do that when there's plenty of porn with 10/10 qts on the internet?
is like you want to punish yourself unconsciously because deep inside you see sex as something sinnful and immoral or some shit
I take back
it's incredibly obvious now.
how rude. who's gonna defend fapping to ugly deformed bitches now if you can't even articulate some bullshit argument?
>DALLE, make Elsa tentacle porn, as shot by Stanley Kubrick
What ever happened to deepmindy and the other coomer reverse ai face searches? I could never find a solid replacement
Can't wait for dall-e for furry porn.
A company would never do it because the porn industry is too profitable. Content produced by an ai is not copyright protected so they can't stop piracy