Hacker News new | comments | ask | show | jobs | submit login
What convolutional neural networks look at when they see nudity (clarifai.com)
428 points by rcpt on Apr 19, 2016 | hide | past | web | favorite | 155 comments

this could be used as a YouTube thumbnail generator for clickbait videos. take the image from the sfw video that has the highest nsfw score, crop down to the area with the highest nsfw score, and bam, instant automatically generated clickbait

Google actually already does something along these lines: http://googleresearch.blogspot.com/2015/10/improving-youtube...

Why doesn't youtube show future frames when you scroll over a video's thumbnail? This feature has existed on many porn sites for quite some time yet youtube lacks it for some reason. Or is there a way to change this in settings?

Perhaps they expect one good keyframe to draw your attention, then the title to help you decide whether to watch. I suspect most YouTube watchers are not in it for the aesthetics, but rather for the content, so a title like "Jelly Bean Factory Tour" can be quite sufficient vs. something like "Amazing amateur video." Your interest in jelly bean factories is probably not predicated on whether the presenter is hot, or whether the beans are a certain color.

> Your interest in jelly bean factories is probably not predicated on whether the presenter is hot

TV networks seem to disagree with you... or at least place an immense amount of emphasis on finding hot presenters, for some reason.

Without being involved in the industry I would imagine this is a case of supply and demand.

There are plenty of people willing to be a presenter for the amount they are willing to pay. So they get to be more selective - attractiveness is one of the most obvious things to select further on.

They could pay less. Then they'd have less attractiveness in the presenter, which hypothetically doesn't matter, and more money, which we know matters. They don't seem to want to do that.

No. The skillset for being a presenter is a subset of other skills used in business. They have to remain competitive other wise people won't work for them - there is a price floor.

Maybe somebody has a patent on it (and the porn companies can afford not to care about that)?

This is exactly the kind of content I love to see -- interesting technical topic, completely outside of any area of expertise of mine, clearly presented, interesting, and pretty obviously useful. Please post more!

Machine learning, at a high level, is incredibly accessible. The basic technology has been around for a long time and has been taught over and over again.

Don't you have to have math knowledge ? Any good resource you'd recommend ?

There's "math knowledge" and then there's "math knowledge". I'm still learning the beginner stuff but so far all that it seems to need is basic calculus (ie. "do you understand what a gradient is and how to calculate one numerically?") and some linear algebra ("do you know what matrices are and how to multiply them together?")

And then there's math knowledge.

Social scientist here, and I have no clue of what you're talking about. Calculus destroyed my engineering aspirations.

Basic calculus doesn't seem that hard though. You understand asymptotes, right? When something gets closer to a certain value, but doesn't quite touch it, like 1/x with 0. That leads quite naturally to limits, where you say "ok, at a close enough value, we'll just ignore that little bit and say it touches".

Then come derivatives. So you have a function graph, and you're trying to determine the slope at a certain point. It's really simple for say y = x, where you just take any two points, but what if the graph is curvy? You'd take the point and then a point close to it, and calculate the slope for that. If the function is fairly smooth, that's fairly close to the actual slope. So you push the other point closer to the original point, and the slope you calculate gets closer to the real slope. You edge closer and closer, and suddenly it sounds a lot like a limit, where your initial point is x, your second point x + h, and h becomes smaller and smaller, so h -> 0.

Here's what's cool though. Since you didn't use 5, or 7.23, but x, you can put any point into this, so you've got a function that maps x onto the slope at x of the original function (at least if I remember correctly). You play around with a couple of different kinds of functions and arrive at various rules for differentiating things without having to do the whole limit thing, like x^2 -> 2x, x^3 -> 3x^2, e^x -> e^x, etc.

And derivatives are really handy. Say in kinematics, velocity is the derivative over time of position, acceleration is the derivative over time of velocity, and suddenly a whole bunch of things make a lot of sense.

Integrals are pretty neat too. So you want to estimate the area under a function from point a to point b. so you draw a box, and there's an area you're missing, or an area you're including but shouldn't, but it's close to the actual area underneath the function. So you think...hrm, I could use two boxes of half the width, that'd be closer to what I want to find. Then three, ten, a hundred, with the sum of the area of boxes getting closer to the actual area. Said more mathy, it'd be a to b, with a width of b-a, and each box gets a width b-a/h, and h gets infinitely large. Then we make that little limit jump again, and get a function. We do that for a couple of functions, deduce some rules, so we don't need to go through the entire process whilst integrating simple functions. a -> x + C, x -> 1/2x^2 + C for example.

With some understanding of the FTC (fundamental theorem of calculus), which roughly states that derivatives and integrals are inverse operations of each other (isn't that neat?), it again helps us make sense of things. For example, look at the equations in physics for constant accelerations: v = v_i + at, x = x_i + vt + 1/2at^2. See how v is the integral of a, and x is the integral of v? For the first one, "v_i" is the constant factor C, and "at" is what comes from the factor a. For the second one, "x_i" is the constant factor C, "vt" is "v_i" integrated, and "1/2at^2" is the integral of "at".

Calculus is super neat.

Thanks for writing all of this. I enjoyed it. It makes no sense to me how calculus makes so much sense. It all just fits together. It's an eerie high point in the "unreasonable" effectiveness of mathematics.

I enjoyed your writeup too. If I had to choose between being able to low level do everything without having a high level understanding of how they fit together and on the other hand having this high level understanding with no low level applying skills I would definitely choose the high level view. It is IMHO far more valuable to know how they fit together.

Very nice. I know it's pretty dated by now (mid '80s, I guess) but the "The Mechanical Universe...and Beyond" [0] videos really opened my eyes on the matter.

[0] http://www.its.caltech.edu/~tmu/

Go show someone how to calculate a gradient vector (assuming they can calculate basic derivatives) and multiply two matrices, and then try to teach them how gradient descent works or the SVM derivation. That's not going to end even slightly well for your average person. There really is no silver bullet for this stuff. You either need some mathematical maturity, or you need to put in an exorbitant amount of time to understand things.

In theory this is true. But it's very hard to find good instruction that is wilfully light on math. Andrew Ng's course for one requires you to be very comfortable with lin alg and convex optimisation.

I just realized my version of accessible is in fact... Inaccessible. Sorry, yes the math needed is moderate.

In general, if you're an user of ML (as in, apply already known ML methods to your particular problem instead of researching better conceptual ML methods) then you'd be able to work on a higher abstraction levels where all the math is inside the libraries and you're not really touching it, much less implementing the formulae.

Not really, you can go far playing around with off the shelf components. IMO, the math is less helpful in NN than other areas because your trying to create a useful aproximation not get a specific correct answer. Thus knowing why something works is only so helpful.

Hey thanks!

Did clarifai have a different company name a few years ago? I built a realtime threaded discussion app with images a few years years ago (like secret or yikyak but in 2011) and used a REST API that was capable of recognising actual body parts.

It generally worked well, with the exception of false negatives for cartoon nudity and false positives for pastrami.

Hrm it looks like I was using recogmission / pifilter, which now seems defunct:


Huh. Even after having used that classic Lena image in the header for multiple visualization projects, I never knew she was naked in the full version.

This paper uses Fabio instead of Lena for that reason:


Because sexy men are less NSFW than sexy women?

It was an expression of distaste for the practice of using a playboy shot as a scientific standard.

Did the author tell you that?


> “My coauthor (asked): ‘What image do you want to use? How do you feel about the Lena image?’ And, I said ‘Definitely not.’” Her coauthor wholeheartedly agreed. And then Needell half-joked: “We should go against the grain and use a Fabio image or something, a male model.” And from there, they both decided to use the image and “take a stance.”

> “I don’t know if the Fabio image will take off but I think what it will do is… stick in people’s minds. And when they think about using the Lena image, they might think (about) using at least a neutral image.”


Alexander Sawchuk estimates that it was in June or July of 1973 when he, then an assistant professor of electrical engineering at the USC Signal and Image Processing Institute (SIPI), along with a graduate student and the SIPI lab manager, was hurriedly searching the lab for a good image to scan for a colleague's conference paper. They had tired of their stock of usual test images, dull stuff dating back to television standards work in the early 1960s. They wanted something glossy to ensure good output dynamic range, and they wanted a human face. Just then, somebody happened to walk in with a recent issue of Playboy.

The engineers tore away the top third of the centerfold so they could wrap it around the drum of their Muirhead wirephoto scanner, which they had outfitted with analog-to-digital converters (one each for the red, green, and blue channels) and a Hewlett Packard 2100 minicomputer. The Muirhead had a fixed resolution of 100 lines per inch and the engineers wanted a 512 x 512 image, so they limited the scan to the top 5.12 inches of the picture, effectively cropping it at the subject's shoulders.

Cited from https://www.cs.cmu.edu/~chuck/lennapg/

Yeah, that was my first thought, "Wow...that's the rest of that image?"

After reading many of these articles I never thought that either. No body ever seemed to mention it.

Unfortunately, whenever someone mentions that Lena was a nude model (perhaps to suggest objecting to the image's use in academic settings as an appropriate testbed), everyone gets angry. Negative reinforcement, I suppose.

That negative reinforcement yields good results though.

To clarify: the only meaning "good results" could refer to is the continued use of a cropped piece of pornography as a standard test image for image research for over twenty years.

Oh, but aren't all SFW pictures cropped pieces of pornography at the end of the day?

If you expand the field of view enough eventually there's going to be people having sex in it.

Not necessarily. If the camera and primary subject are in an enclosed room, all you're ever going to get is more wall.

Of course not, and its really worrisome that you think that. Most of the set of grey scale 50x50 pixel images are not crops by the simple fact that they contain no correlations at all, that's simple combinatorics.

Is that actually already pornography?

Well, from context its pretty clear it is intended to be used as pornography and I think that's good enough to label it as such.

I just meant, that western society seems to have become more accepting since the photo was shot, and these days people might classify it as an 'artsy nude' instead?

Which is why I think it would be really interesting to run a set of Robert Mapplethorpe's photos on it to see how they are classified. Some of his photographs are much more explicit while at the same time are much more specifically art. He photographed a lot of posed homoerotic BDSM Large Format photos specifically staged to develop tension, but not necessarily or even usually sexual tension or disgust tension, between the viewer and the person posed in the photo as a person and as a form. Alternatively he photographed large format calla lilies or orchids that are explicitly erotic. The photos were almost always black and white, and were s seen as controversial when he was alive and then again right when he died.

Then, in contrast, run some Japanese woodblocks in the Ukiyo-e style, since many of those are actually explicit. (including straight up porn woodblocks) Most Japanese woodblocks are also highly stylized as well, and it can take a moment to realize what is going on, including in some of the pornographic or hinting at the explict/could be explict/explict personalities.

I'm not going to say all Robert Mapplethorpes are safe for all work environments, nor all Japanese woodblocks, but I am curious to know if it can see the difference of person in homoerotic outfit who looks physical and gets you to notice his form, vs a Suzuki Harunobu being a bit more sly but more more explict in he is telling a story and his story is sex

I don't think you can take something which was once meant to be erotic pornography and add time to make it an 'artsy nude.' But if you can point to an example from the history of art, that would make your claim more reasonable.

I've always thought that was part of the joke.

It's a Playboy centerfold...

Then you probably did not know she's more attractive in the full version ;)

This is very exciting as I toyed with this problem around 2000-2001. I speculated it should be easy to recognize porn because the images had recurring patterns. I made a list of a bunch of them that should at least reduce a moderator's workload. Not going into details but the best foresight I had was recognizing the Got Milk commercials would throw the classifications off. It did lol.

In any case, we just had things like fuzzy logic, SVM's, primitive ANN's, and so on. I could describe and illustrate the patterns with ease but couldn't encode them into an algorithm for the life of me. That the modern tech created in the image set some of the same exact patterns my brain spotted is pretty amazing. The researchers also have a ton of training data to better illustrate NSFW and SFW. That they're throwing in lots of SFW is very smart as it's easier to bias this topic than most.

All I can say I keep up the good work. Been neat to see an insurmountable-on-Pentium2 problem get stomped by the CNN's. I still want to see them applied more to hunting source code defects, host forensics, and intrusions.

True story: Then we throw before and afters of Prophylactic Bilateral Mastectomies with various types of nipple reconstructions among previvors who found each other and were part of a semi-private group on facebook, and hell breaks loose.

A classifier isn't human judgement by itself, and you need to be aware of how people view something in context. Good training images are also critical (one of the reasons why hell broke loose above were some people mastectomies were more likely to be censored than others)

Though I will admit I am very curious what happens if you throw a Robert Mapplethorpe at it.

Then everyone got censored, then the group got banned, then facebook got slammed by cancer groups and cut a deal with them to rehost cancer support groups as long as everyone stopped sharing mastectomies directly to facebook because Facebook could not specially moderate and alter feeds for those thinking about or dealing with prophylactic mastectomies. However, this issue still pops up ALL THE TIME for them on Instagram

True story: an argentinian video for breast cancer autoexploration had to be filmed with a fat man because otherwise it would get banned.

Warning: In spanish and surely NSFW for someone: http://verne.elpais.com/verne/2016/04/19/articulo/1461073073...

Heh, this SFW picture of a room (or is it a cardboard box? hard to tell!) is 75% likely to be NSFW, according to the API.


To be fair, at first glance my brain thought that too.

And because your brain (and my brain) parsed it that way, it is likely that others will too which may mean that it is not really a picture you want on your screen at work. I.e. even though it is technically OK, it may be, for practical purposes, NSFW.

Note, if you actually care about whether links you click are SFW or not, that link is NSFW as to your coworkers in the room, it will be indistinguishable from you looking at graphic porn.

Mostly only if they're porn consumers themselves. Plenty of folks don't make the suggestive links for images of this kind.

Reminds me of an MMORPG guild I used to be a member of, called Fist of Innoruuk. The forum had a large background image of a fist in a (star? wreath? can't recall). One of the members was reading the forum at work when his boss walked by and asked why he was looking at a fisting forum at work: "What's a fisting forum?".

> Mostly only if they're porn consumers themselves.

That's actually a very, very large percentage of people

> Plenty of folks don't make the suggestive links for images of this kind.

Well, not out loud.

Apparently apple pie has been a significant source of false positives in NSFW detection -- it mimics both flesh tones and contours. (Or so I was told by a then-major image hosting company pitching their filtering tech several years back.)

Combining one of these nudity trained classifiers with a project like neural-style (https://github.com/jcjohnson/neural-style) could create some serious mischief.

I will give $20 USD to the first person to set this up and run Donald Trump through it.

That hashname is disturbingly close to top Linux contributor and ext2/3/4fs maintainer's name Ted Y. Ts'o...


How shall I arrange payment? Please post the source as well, surprisingly I don't see much of a difference.

Yes it's not mine, I found it on that tumblr. Here is the original image: http://41.media.tumblr.com/7b0a64ba10fb31b206c2284d01cd86a8/...


It could be used as an NSFW content identifier, it also could be used as a fine-grained porn video classifier. At least that's what I want to use it for.

Yes, they have a pointer for work like that, too.

> If you're interested in using convnets to filter NSFW images, check our NSFW API documentation to get started. https://developer.clarifai.com/guide/tag#nsfw

Highly recommend this video on the Deep Visualization Toolbox to anyone interesting in understanding more about how convnets work through visualization:


author here: let me know if you have any questions

How well does your solution handle non-white-people NSFW images?

I know it's only a matter of training with other inputs, but most solutions I've seen haven't been trained with enough non-Caucasian images to give good results. Since this at least seems to be detecting shape regardless of skin tone it has more potential, but it'll only be as good as its training set.

It was trained on non-white-people NSFW images to (eg. non-white people, hentai). However, I don't have per genre numbers right now.

I'm sure it's easy to find "training data" for this project, but FYI 4walled dot cc is probably a GREAT source. It's an archive of thousands of 4chan images, and most are already classified by SFW / Borderline / NSFW. Since it came from 4chan, it includes lots of non-pornographic NSFW things as well (gore, swastikas, etc.), and lots of duplicates with various cropping, reencodimg and other processing done to them.

Is the idea that people would submit all images to your pre trained API / program, or would they also be generating a training set themselves? How sensitive is the algorithm to incorrectly tagged images in the training set?

Right now we've got the fully-trained convnet deployed here: https://developer.clarifai.com/guide/tag#nsfw

I have not quantified how robust this net is in the face of noisy labels.

How well does it work on grayscale images?

EDIT: Just created an account and tried it out on a test image, and it returned "nsfw" with prob 0.998, which is pretty amazing.

EDIT2: It's much less strong against oblique views, even with full color. http://40.media.tumblr.com/cfeea64370e3b1228906472ebd8f344e/... (NSFW, obviously!) only returns 0.541.

Awesome, fwiw you can hit the demo here without getting an API key http://clarifai.com/?model=nsfw-v1.0&probs=1

Could you do a deep dream run and generate porn images from the training set?

I don't think the result could be called "porn" in a productive sense for anyone other than the Great Old Ones.

How did you come up with 64x64 sliding windows?

Just tried a few heat maps and that one looked good.

Why did you have no male examples?

Look closer at the last figure.

As an aside, this might be the first time I've seen Lena used in a completely uncontroversial manner.

So true. Here the use fits the application. In past, people objected to using a Playboy centerfold as standard image in science, software distros, etc... I remember Lena on my Amiga 1000...

Recently watched the Machine Learning Conference and learnt about Clarifai there ( https://www.youtube.com/watch?v=oHE0DfBqwHo ) And its pretty cool to see that they are actually expanding along various verticals and understand every image and even videos. And its great to see it here today. Please keep posting more crisp, concise and technical posts here! Kudos for forevery!

ok, somebody go train a generative network now

That sound like it could go horribly horribly wrong. I wonder if it can go to an area where no single person is interested anymore. But then again, if the internet has taught us anything.. hm.

I believe you just called Rule 34 on The Thing. :S

I love the choice of Lena at the top.

The image has a surprising history. See https://en.wikipedia.org/wiki/Lenna for the basics.

How fun - their convnets have determined that Facebook is an incarnate deity.

https://www.facebook.com/images/fb_icon_325x325.png is classified as "cross", "no person", "Resurrection", "spirituality", "symbol", "religion", "god".

In all seriousness, I found this pretty interesting, as I've been toying with using RNNs to classify NSFW images on my icon conversion site. The biggest issue that I've run into is that "icons" are squarely in a different class that photographic images. For some reason the *NN networks that I've toyed with break down on classifying icons because they aren't able to easily correlate a stylistic interpretation with a real item. It kind of makes sense given how they work, but there's got to be a way to work around it...

Could you link to it?

EDIT: I see, https://clarifai.com/#demo.

This is great. Hadn't heard of deconvnets. Thanks for posting!

in case the blog goes down there is a backup here http://ryancompton.net/2016/04/19/what-convolutional-neural-...

I've been wondering, since there have been white papers on reconstructing occluded faces based on face detection algorithms[1], and we have these nudity detection algorithms - do you see where I'm going - when will someone make a neural network that can automatically make any photo showing skin into a nude photo...

[1] EDIT: I did not find my original link, but here is a similar paper http://www8.cs.umu.se/education/examina/Rapporter/NaeemAshfa...

Deep dream on this might be visually interesting.


What a time to be alive

Those aren't the best politicians to choose. You need to do Michael Dukakis, the Democrat who ran in 1988. He has a big long nose and bushy eyebrows already.

Those pictures look lame because the algo is trying much too hard.

Is there some way to locate areas most suitable for "enhancement" and process only those?

The only one that's really recognizable without any real cues (such as Hulk Hogan's bright yellow clothing) is Ted Cruz.

I find this hilarious for reasons that are off topic on HN and politically inflammatory.

can you imagine the ai bots that get to work on this assignment? once computers become more like the AI in the movie ex machina, would a bot be "turned on" by the nudity? it would affect them. The bot would be distracted by it and have a hard time concentrating on other things when it's around. Sounds human to me.

Why would it be "distracted" by it?

There certainly can be a bot that pays more attention to any feature the programmer instructs it to, but to what purpose? What is the bot supposed to do with this particular stream of bits that was just promoted to its attention?

If the answer is, "stop other productive work and waste CPU cycles" that says a lot more about what "turns on" the programmer than about the program itself.

There's no reason an AI would attach any particular significance to human nudity unless it was specifically designed for a sex bot.

Things that turn on and distract an AI would be things found in their positive training data. Nudity wouldn't work, but if you train an AI for long enough, it might develop some interesting 'fetishes'.

Maybe the bots would develop some curious interests if training progress required sexual interactions with humans and there was brutal competition between bots ;)

>if training progress required sexual interactions with humans and there was brutal competition between bots

Most realistic scenario: A new startup releases a 100% neural net based OS. Microsoft and Apple are implementing their own NNs but they struggle to keep up - until Cortana finds out that sex sells...

Why would "computers become more like the AI in the movie ex machina"? What are you basing that on? Are you knowledgable about how the field of AI is progressing or how these bots function? Do be careful to not confuse fiction with reality.

I'm just saying, it's funny to think once computer have "feelings" and approach human level, the ones trained to do this, are the lucky ones. And, if the world is a simulation, maybe a very similair algo was once programmed in us.

Surely if you were making a "human level" AI with "feelings", you'd make it attracted to cleaning; or building cars; or whatever task it was built for?


great post - where did the training data of 500K images come from ?

500k were for testing, we trained on several million. Can't say where I get them.

hand picked? hard work ;)

someone handpicked 500K images ... come on now :) no way

Interesting! I wonder how it performs on this vase, which is obviously inscribed with dolphins:


The about the author section describes him as a "NSFW enthusiast". Aren't we all...

Is it only me, or the top bar (which re-appears with scrolling up) is extra-annoying?

Accurate recognition of breast & vulva mimics the brains of men.

I don't think though that men can create neural nets that simulate the brain of woman ;-)

I wonder how this performs on black-and-white images.

Can you drop the scrolljacking on your site? I spun my mousewheel and I went halfway down the page. Usually I instantly jump ship in these cases (even when the header is a Playboy centrefold. Priorities, priorities...)

We detached this subthread from https://news.ycombinator.com/item?id=11530336 and marked it off-topic.

Oh come on HN! The author of this great article asked if you have any questions and you go for the scrolljacking comment?!

I really want to see dang doing the "detached and flagged" thing on these pissant comments about scrolljacking, JavaScript, contrast ratios, etc. especially as the top comment to an author making themself available to discuss a technical article on convnets!

If my comment were the only one allowed, then I would agree with you. Thankfully, there is a comment box available for anyone else who would like to comment on his article. So, my comment will be simply one of many.

I agree, and belatedly did so. In the future you should bring stuff like this to our attention by either flagging the comment or emailing hn@ycombinator.com.

You can point people to dang's post here: https://news.ycombinator.com/item?id=9238739

> A reader emailed to complain about how this and other HN discussions often become derailed by off-topic carping about blog design. I agree completely. Could there be a more classic form of bikeshedding? It would seem parodic if it weren't sadly real. This has become more of a thing on HN lately. It needs to become less of a thing.

We got a more readable alternative link to use and the author's now aware that the original was difficult for some people to read. What exactly is the problem?

+1. It was difficult to scroll on my laptop, even the smallest scroll gesture I can make moved the page down more than I wanted.

I don't know why some websites hijack scroll events. What are they trying to achieve? A better scrolling experience? Almost every time, it makes it a lot more frustrating.

Hey what browser are you using? None of us here know what's going on with this.

Not sure what's going on with the weird scroll speeds, but for me the problem is motion sickness from the smooth bouncy scrolling.

Browsers have options to disable smooth scrolling for a reason. Please don't force it back on.

Ugh. I still don't know what thing in WordPress set this off.

fwiw I have the same content on my personal blog which shouldn't do anything with your scrolling http://ryancompton.net/2016/04/19/what-convolutional-neural-...

Looks like it's part of the UNCODE theme[1], and can be disabled by dropping smooth-scroller from the body class.

Your personal blog is much more readable, thanks :)

1: http://themeforest.net/item/uncode-creative-multiuse-wordpre...

Wherever it comes from, I'm guessing the "smooth-scroller" class on <body> applies it.

I use Chromium. Just tried Firefox on a desktop, and it seems to be okay.

If they had taught me this in high school I'd have never dropped out.

You forgot to mention the boobs. Everyone likes those, even if it's just pixely thumbnails.

Please don't do this here.

We detached this comment from https://news.ycombinator.com/item?id=11530734 and marked it off-topic.

Could you make it faster if you had no ego?

We detached this subthread from https://news.ycombinator.com/item?id=11530336 and marked it off-topic.

What does that even mean?

Maybe pareci is considering whether the paper's author is a convnet.


That's past uncharitable and into personal attack. You can't comment like this here.

We detached this subthread from https://news.ycombinator.com/item?id=11530336 and marked it off-topic.

Porn must flow...

> During her pornographic career, which lasted from 1993–1994

She managed to get out early. Others don't:


These are pre-internet porn careers, the world has changed

For the worse. Skyrocketing demand, skyrocketing exploitation.


I can't comment on the film.

but if demand is strong then exploitation should go down by economics.

Exploitation is rife in all society, the US is unusual in that the welfare state encourages exploitation of the poor. Restaurant staff getting paid $5 an hour in Florida for instance. [1]

But back to your question :

> Have you ever wondered if the women that you're classifying have any reservations about being exposed for your titillation

I provided you with direct evidence that the answer was "no". You seem to have shifted the question to "exploitation" not "reservation".

JJ famously called out the industry but she seems to have few reservations about going back.

[1] http://www.paywizard.org/main/salary/minimum-wage/florida

You showed me one person who appears to be happy, but also has a huge incentive to appear to be happy even if she isn't: if she doesn't appear happy, she doesn't make money. I showed you a few people who appear to not be happy doing it, even though while they were doing it, they claimed to be happy doing it.

This really does not make you question at all how much they want to appear to be happy in porn even if they really aren't? How can you possibly tell apart those who are happy from those who aren't? And if they are appearing in porn without really wanting to, how can you jerk off to them with a clear conscience?

I was with that Dworkin style "porn is rape" feminism, then I met some sex workers and learned that for plenty of women it is something they like doing.

So my conscience is as clean as buying a cup of coffee from someone who hates being a barista.

I am sorry for people who are exploited in all walks of life, porn just happens to be one of those.

How can you use a computer ? Or any of the plethora of stuff made in the third world.

Porn is here to stay. Hounding and demonising it makes the situation worse, not better.

Heh, the internet does not like having its purpose questioned. Porn must flow...

Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact