Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI real-time human full-body photo generator

generated.photos

Post reply on HN
Print view View original

271–280 of 421 posts

Re: AI real-time human full-body photo generator

#271

Post by brucethemoose2 »

CDSlice wrote: post #258
wmeredith wrote: post #255

Earlier quoted context omitted.

I agree it’s gross, but I’m not sure how to articulate it. It’s making real something people have done in their minds-that is imagine people with no clothes on. These pictures aren’t the actual subject naked. There’s nothing being discovered or disclosed. It’s a pure fiction. But it still bothers me.

It is fiction yes, but if it is lifelike enough does the difference matter? Even without the ick factor of making porn of someone without their consent, it would be so easy to destroy someone's career or relationship by making these deepfakes and then spreading them around. Especially once the tech gets more life-like and loses the current AI gen tells. And the same would apply to doing this the old fashioned way in…

> it would be so easy to destroy someone's career or relationship by making these deepfakes and then spreading them around.

If nude pics can get you fired, work culture needs to change.

Same with relationships.

Basically, people need to learn not to trust digital media at all without some kind of authentication, and to be a little more tolerant of nude human bodies when they do pop up.

Re: AI real-time human full-body photo generator

#272

Post by a_t48 »

pezezin wrote: post #235

I know that currently AI is the hot topic, but if you want to generate pictures of well-known objects like humans, why not just render a 3D model of a human? Something like Poser (the software) is way more efficient than any GAN or diffusion model. But Poser is 28 years old now, I guess it doesn't attract sweet VC money.

This exists! https://www.unrealengine.com/en-US/metahuman

Re: AI real-time human full-body photo generator

#273

Post by jfoster »

CDSlice wrote: post #258
wmeredith wrote: post #255

Earlier quoted context omitted.

I agree it’s gross, but I’m not sure how to articulate it. It’s making real something people have done in their minds-that is imagine people with no clothes on. These pictures aren’t the actual subject naked. There’s nothing being discovered or disclosed. It’s a pure fiction. But it still bothers me.

It is fiction yes, but if it is lifelike enough does the difference matter? Even without the ick factor of making porn of someone without their consent, it would be so easy to destroy someone's career or relationship by making these deepfakes and then spreading them around. Especially once the tech gets more life-like and loses the current AI gen tells. And the same would apply to doing this the old fashioned way in…

Potentially the opposite. It may become more difficult to use such images to harm someone's career or relationship. Perhaps nothing is believable in the future.

Re: AI real-time human full-body photo generator

#274

Post by bravura »

gwern wrote: post #9

If you're wondering how it's so fast and cheap and they can generate variants so easily, it's because they're using GANs (see the footer). GANs are way faster than diffusion models because they generate the image in a single forward pass and their true latent space encoding makes editing a breeze. (And if you're wondering how it can look so good when 'everyone knows GANs don't work because they're too unstable', a wi…

Hi gwern, asking earnestly and without snark about this:

"And if you're wondering how it can look so good when 'everyone knows GANs don't work because they're too unstable', a widespread myth, repeated by many DL researchers who ought to know better, GANs can scale to high-quality realistic images on billion-image scale datasets, and become more, not less, stable with scale, like many things in deep reinforcement learning."

My interpretation from how this is written is that you are saying: "Researchers say that GANs are unstable at inference time." Is that the correct reading? If so, where have you seen this sentiment? More commonly I've heard that people criticize GANs at inference time for mode collapse and monotony (which comes from training), not instability.

Or are you saying that researchers say GANs are unstable during training, which is a common criticism. Don't you feel this is the case? A lot of different tricks are used for getting generators that are roughly balanced with more powerful discriminators so that the adversarial game is balanced, like TTUR and encodec's weight balancer etc. In this case, are you saying that GAN training is as straightforward as diffusion training?

In my experience, GAN training involves an unholy number of heuristics and best-practices are still quite murky.

I am eager to hear your response.

[edit: my experience with GANs and diffusion models isn't about taking image problems with existing working models and scaling or refining them, but about applying GANs and diffusion models to different domains (audio) and on novel problems with different kinds of conditioning. I would love to see more controlled experiments comparing weak and sophisticated generative backbones as the experimental control and varying the training regime across different GAN + diffusion flavors.]

Re: AI real-time human full-body photo generator

#275

Post by QuantumG »

rgmerk wrote: post #265
CDSlice wrote: post #258

Earlier quoted context omitted.

It is fiction yes, but if it is lifelike enough does the difference matter? Even without the ick factor of making porn of someone without their consent, it would be so easy to destroy someone's career or relationship by making these deepfakes and then spreading them around. Especially once the tech gets more life-like and loses the current AI gen tells. And the same would apply to doing this the old fashioned way in…

I’d also add that the knowledge of what’s possible diffuses to abusers a lot quicker than it diffuses to, say, the grandparents of victims.

Through all that pain, humanity adapts.

Re: AI real-time human full-body photo generator

#276

Post by sillysaurusx »

GaggiX wrote: post #253
sillysaurusx wrote: post #244

Earlier quoted context omitted.

There are so many holes in this, but to pick just one: > for example your buggy BigGAN implementation The bug was that the batchnorm gamma was initialized to zero instead of one, so all the model params were being multiplied by zero instead of one. Literally any model that uses batchnorm is susceptible to this bug, and it certainly has nothing to do with GANs. I think it’s too easy to spout dogma, and calling gwern d…

It has to do with GANs as the added complexity makes the model more difficult to debug, the other day a friend of mine had a problem with the normalization layers of his diffusion model, it was fixed almost immediately, if something goes wrong you have less to worry about. Also no I don't take "he is gwern" as an argument ahah

Bookmarking this for when time proves you wrong after a couple years. I’ll drop by to say hello.

Re: AI real-time human full-body photo generator

#277

Post by Mistletoe »

14 wrote: post #75
philote wrote: post #73

Earlier quoted context omitted.

Yeah, this was my first attempt: https://generated.photos/human-generator/64e648819c8c0400088... Not quite right. I am, however, impressed that the fingers are generally "mostly" correct.

my first attempt created some kind of gym monster lol https://generated.photos/human-generator/64e64a2f412bec0009b...

Hermione’s been in the restricted section again…

Re: AI real-time human full-body photo generator

#279

Post by dontupvoteme »

brvsft wrote: post #256
dotancohen wrote: post #155

Earlier quoted context omitted.

I don't know if the thing in the crotch is a penis or a scrotum, but it is definitely NSFW: https://images.generated.photos/0wV1dBnZ15hGneEfqfZT7SdEIill... My prompt was simply "Standing in front of a rocket.".

Hah. I've been playing with StableDiffusion locally and have gotten a few abominations like this myself.

Whatever you do don't take all those long prompts that include (((aesthetic))) and such and put them in the negative prompt with "human" in the positive prompt.

It's not nightmare fuel it's a nightmare engine

Re: AI real-time human full-body photo generator

#280

Post by jfoster »

hn_throwaway_99 wrote: post #98

I get that most of these are hilarious (this is my favorite comment on HN in some time, https://news.ycombinator.com/item?id=37239909 ). But still, I find this incredibly frightening. These are only going to get better. Does anyone doubt that in a couple years time (if that) we'll be able to put the image of any known public person into whatever generated photo we want, which would be indistinguishable from reality?…

It's really interesting how many people see a product/service with an impressive result (previously seemed impossible), albeit with flaws that humans find funny/obvious, and then are unable to imagine how it might be once improvement leads to the issues being ironed out.

A few years ago generating images like many of these was unimaginable, and now this does it, but perhaps 50% have some silly flaw in the image. Unless there were some reason to believe that further improvement isn't possible, as you say, it is likely to become indistinguishable from reality in the relatively near future.

Post reply on HN