It's crazy that Visa, Mastercard & other payment processors were debanking actual creators and artists for adult content recently and now OpenAI is doing this.
OK, and what's the problem with that specifically? I presume they will use more sophisticated methods for age assurance than "I'm 18".
Besides, people are using LLMs for exactly that for ages (including Claude and ChatGPT), so the only difference is that the non-NSFW behaviour will be better (less crippled).
In general I think it's perfectly fine to have uncensored NSFW chats. The fact that all of the mainstream LLMs are way more restrictive than basic web searches we've had for decades is kind of ridiculous.
In this particular case though I can't imagine it's a good idea to hand Sam Altman and his company chat logs of your explicit fantasies tied to your real identity.
That's the idea. For too long people have been able to free-associate and share truth online. Now they finally have a reason to end anonymity online. Due to the raw power of AI there is a very real threat these new tools provide to bad actors in the generation of misinformation and disinformation and now we need to ascertain the providence of all content created online. With "AI" being injected into every site, app, and OS we will be forced to associate real world ID's to any and all actions int he digital realm in order to "protect" the public from the wave of fabricated reality they will create.
Of course these tools will be used by those that control the moderation systems to craft and shape global negative on a scale we have never seen before. A new kind of Dark ages are here. Not one where dusty tomes are changed to the monastery walls, no, our books write and rewrite themselves in real time as needed until every thought is controlled by the over-mind.
And I saw one of his heads as it were wounded to death; and his deadly wound was healed: and all the world wondered after the beast.
Me too, but that was in the OpenAI developer console, which is a different product than ChatGPT, so I wouldn't be surprised if one would have to submit it again.
and if you don't send them ID the Chat will start responding "oh you little tike, that's something we need to ask your daddy if we're allowed to tell you"
I assume that the porn industry will not agree to consider the training material as fair use. And this industry has quite a lot of money for good lawyers. In this respect, we will see on what basis the quality improvement in the relevant content is to be achieved.
OAI is a big fat lawsuit target, plus they can restrict it and still make the money because odds are they'll also have it on API, and someone could verify for the API then offer a wrapper around OAI with less stringent ID verification.
OAI ultimately still makes a ton of money, and gets to wash their hands of much of the liability. It's a win-win.
I find these takes so odd. Porn is the early adopter of EVERY technology. When non-linear editing software was invented in the 90s by Avid, who was the first industry to use them? It might not be advertised, but it's true. Porn leads the way. lol. This is no different.
This isn't really porn adopting ChatGPT, it's OpenAI allowing porn. There was nothing stopping Pornhub from releasing their own LLM. If porn became a feature gated behind YouTube Premium I would probably say "wow they really need customers huh".
If so, it's a clever strategic move as they know that most of their competition will struggle to allow such a policy because they can't move so fast on such morally loaded topics like OAI.
He literally said on the August podcast that they wouldn’t do it because it would be a short-term gimmick. As I type, CNBC (who pumps them every hour of the day) is talking about how suspicious it looks.
If we assume that on the podcast he presented a well reasoned opinion based on competitive analysis and public sentiment and whatnot, which is a big if for this guy, then I'd ask what changed? If he's truly desperate now, why wasn't he 2 months ago?
Maybe he just looked at their analytics and saw that everyone is trying to have virtual sex with it?
What do you mean? It was banning porn that didn't work well for Tumblr. These models have always been very capable of writing erotica, and consenting adults using them for that purpose is clearly something that should be allowed by default, within the bounds of legality.
"Now that we have been able to mitigate the serious mental health issues and have new tools"
Huh? Quite the statement when there are still people out there having parasocial relationships with chatbots. This problem is going to get exponentially worse when the chatbots can start to get sexual.
Always a new account with some snarky neo-luddite reply, as if Google, or Facebook, or Microsoft, or Amazon, doesn't already know everything there is to know about you. I mean, come on. That ship has sailed at least two decades ago and this kind of conversation simply isn't productive imo.
This is a ridiculous attitude, exactly equivalent to 'if you've nothing to fear you've nothing to hide'. In practice different geographic localities (e.g.: EU, California) have far different (and improving) limits on the personal information that can be stored by companies and how it is used.
Various forms of online hygiene, from ad blocking to VPNs, to anonymous email accounts are likely utilised by the hacker news audience. Because we're all well aware of the value of our privacy, the value of our personal data, and the aggressive and often illegal efforts by major tech firms and data brokers to acquire both.
It's incredibly important to continue progress in the area, both in terms of informing the public (and politicians) of how our behaviour is observed and shaped by social networks and big tech: and by taking all obvious and easily available measures to reduce and remove the information they have and continue to gather.
One obvious way is by not sharing private sexual data with a major corporation, where not absolutely necessary.
> This is a ridiculous attitude, exactly equivalent to 'if you've nothing to fear you've nothing to hide'.
Quite the contrary. My attitude encourages one to actually solve problems that we are encountering today, and not to fight battles that were lost two decades ago, which was the main thrust of my argument.
Also, as someone born in the Eastern Bloc, please allow me to throw up in my mouth at a publication that unironically cites the Communist Manifesto.
"We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. We realize this made it less useful/enjoyable to many users who had no mental health problems..."
I dread to think where society is going to be in 10-15 years. We've had a generation or two raised thinking porn == sex and the next generation will all be in romantic relationships with ChatGPT. Businesses have just given up on having any social responsibility whatsoever.
Full quote is this:
"The questions posed by AI are open to all of us. And they demand answers. A book that is written to challenge all readers, no matter your worldview, 2084 shows how the Christian worldview, properly understood, can provide evidence–based, credible answers that will bring you real hope for the future of humanity."
Does 'the Christian worldview' invalidate any conclusions the authors may present?
Describing the "Christian worldview" as "evidence-based" suggests that a healthy dose of skepticism is warranted when evaluating the authors' conclusions, yes.
'The Christian worldview' is actually evidence-based, it's not based on childish stories of "ghosts" and "angels".
Please, read the book first before judging by the cover.
It's crazy that Visa, Mastercard & other payment processors were debanking actual creators and artists for adult content recently and now OpenAI is doing this.
OK, and what's the problem with that specifically? I presume they will use more sophisticated methods for age assurance than "I'm 18".
Besides, people are using LLMs for exactly that for ages (including Claude and ChatGPT), so the only difference is that the non-NSFW behaviour will be better (less crippled).
Win/win. No? Why?
In general I think it's perfectly fine to have uncensored NSFW chats. The fact that all of the mainstream LLMs are way more restrictive than basic web searches we've had for decades is kind of ridiculous.
In this particular case though I can't imagine it's a good idea to hand Sam Altman and his company chat logs of your explicit fantasies tied to your real identity.
I assume you will need to do a Blockchain Iris scan from his other company in order to prove your Age and Identity.
----TINFOIL HAT BLOCK START----
That's the idea. For too long people have been able to free-associate and share truth online. Now they finally have a reason to end anonymity online. Due to the raw power of AI there is a very real threat these new tools provide to bad actors in the generation of misinformation and disinformation and now we need to ascertain the providence of all content created online. With "AI" being injected into every site, app, and OS we will be forced to associate real world ID's to any and all actions int he digital realm in order to "protect" the public from the wave of fabricated reality they will create.
Of course these tools will be used by those that control the moderation systems to craft and shape global negative on a scale we have never seen before. A new kind of Dark ages are here. Not one where dusty tomes are changed to the monastery walls, no, our books write and rewrite themselves in real time as needed until every thought is controlled by the over-mind.
Happy Prompting.-----TINFOIL HAT BLOCK END-----
I think this is my fav HN comment ever.
Perhaps they intend to collect gov ID's?
They already do. If you want to use gpt-5 with the new responses API, I had to upload my id.
Me too, but that was in the OpenAI developer console, which is a different product than ChatGPT, so I wouldn't be surprised if one would have to submit it again.
and if you don't send them ID the Chat will start responding "oh you little tike, that's something we need to ask your daddy if we're allowed to tell you"
> Besides, people are using LLMs for exactly that for ages
Hey! How old are you!?
5+ years is definitely "ages" in LLM context. It was a thing way before GPT 3.5 (obviously, as this is humans we're talking about).
[dead]
Are you responding to something specific you're seeing in this thread or on Twitter (I can't see the replies on Twitter)?
I assume that the porn industry will not agree to consider the training material as fair use. And this industry has quite a lot of money for good lawyers. In this respect, we will see on what basis the quality improvement in the relevant content is to be achieved.
>> I presume they will use more sophisticated methods for age assurance than "I'm 18".
Why presume that? "Are you over 18?" is the age verification method used by 99% of the web.
OAI is a big fat lawsuit target, plus they can restrict it and still make the money because odds are they'll also have it on API, and someone could verify for the API then offer a wrapper around OAI with less stringent ID verification.
OAI ultimately still makes a ton of money, and gets to wash their hands of much of the liability. It's a win-win.
Anthropic: We are releasing reports on economic impact and evaluating model welfare just in case it experiences consciousness
OpenAI: We are... doing something else...
You have to pay some respect to sama. Imagine this move done by Google, Meta, Amazon or Microsoft: It won't happen.
> Meta
Indeed; Facebook didn't bother with the verified adults bit: https://www.theguardian.com/technology/2025/aug/15/meta-ai-c...
Can you imagine if Google search returned adult content?
Kind of reeks of desperation. And, given what’s been alleged about him personally, it seems odd. Is this something MS is on board with?
I find these takes so odd. Porn is the early adopter of EVERY technology. When non-linear editing software was invented in the 90s by Avid, who was the first industry to use them? It might not be advertised, but it's true. Porn leads the way. lol. This is no different.
This isn't really porn adopting ChatGPT, it's OpenAI allowing porn. There was nothing stopping Pornhub from releasing their own LLM. If porn became a feature gated behind YouTube Premium I would probably say "wow they really need customers huh".
> There was nothing stopping Pornhub from releasing their own LLM
There's probably a lot stopping them considering there's about 3 companies making models that aren't shit.
If so, it's a clever strategic move as they know that most of their competition will struggle to allow such a policy because they can't move so fast on such morally loaded topics like OAI.
Because AI bubble is based on companies spending a trillion dollars on AI, not porn.
Why desperate? There's clearly a demand for it.
He literally said on the August podcast that they wouldn’t do it because it would be a short-term gimmick. As I type, CNBC (who pumps them every hour of the day) is talking about how suspicious it looks.
If we assume that on the podcast he presented a well reasoned opinion based on competitive analysis and public sentiment and whatnot, which is a big if for this guy, then I'd ask what changed? If he's truly desperate now, why wasn't he 2 months ago?
Maybe he just looked at their analytics and saw that everyone is trying to have virtual sex with it?
Typically not a good sign when a company resorts to selling porn.
Didn't work well for Tumblr, not working well for Twitter, maybe AI is different.
What do you mean? It was banning porn that didn't work well for Tumblr. These models have always been very capable of writing erotica, and consenting adults using them for that purpose is clearly something that should be allowed by default, within the bounds of legality.
Have billions from investors, desperately seeks business model
A lot has changed in 2 months: https://www.reddit.com/r/artificial/comments/1o78ewb/sam_alt...
I've thought that the Sora videosharing app was dumb, but if it were NSFW at least it would be very popular.
Well, here we go.
I think they want to finally give their investors something stable in return and see how it works at OnlyFans.
To me this is a counter to grok.
Huh, all the replies are about wanting access to version 4o again.
I'll just leave this link here: https://www.reddit.com/r/MyBoyfriendIsAI/
The folks there will provide a detailed explanation about 4o and 5 differences.
Onlyfans models already paying poverty wages for people to pretend to be them in chat.
Presumably this will just be part of that offering.
Just wait until the Onlyfans models end in poverty themselves because Sora replaces them (after training on their data).
The G stands for gooner
the PT standard for "Porn Time"
"Now that we have been able to mitigate the serious mental health issues and have new tools"
Huh? Quite the statement when there are still people out there having parasocial relationships with chatbots. This problem is going to get exponentially worse when the chatbots can start to get sexual.
What do you mean start? It's been a thing the whole time
Especially given this[1].
>> "ChatGPT is trained to direct people to seek professional help," such as the 988 suicide and crisis hotline in the US or the Samaritans in the UK.
>> The company acknowledged, however, that "there have been moments where our systems did not behave as intended in sensitive situations".
[1] https://www.bbc.co.uk/news/articles/cgerwp7rdlvo
More discussion: https://news.ycombinator.com/item?id=45582874
The internet is good for two things. Sharing research and sharing porn.
It's also great for spreading conspiracy theories and fake news.
Another kind of porn?
InfoPorns
Cool, I've always wanted to share my intimate desires with creepy tech bros
Oh no worries, worst case it will be leaked once you run for office somewhere ;)
Not before they try to use it to leverage some concessions first.
What about the people currently in office?
Only few have the liberty to grab some on some tape in Moscow while going into some island by paying of some starlet ...
Always a new account with some snarky neo-luddite reply, as if Google, or Facebook, or Microsoft, or Amazon, doesn't already know everything there is to know about you. I mean, come on. That ship has sailed at least two decades ago and this kind of conversation simply isn't productive imo.
This is a ridiculous attitude, exactly equivalent to 'if you've nothing to fear you've nothing to hide'. In practice different geographic localities (e.g.: EU, California) have far different (and improving) limits on the personal information that can be stored by companies and how it is used.
Various forms of online hygiene, from ad blocking to VPNs, to anonymous email accounts are likely utilised by the hacker news audience. Because we're all well aware of the value of our privacy, the value of our personal data, and the aggressive and often illegal efforts by major tech firms and data brokers to acquire both.
It's incredibly important to continue progress in the area, both in terms of informing the public (and politicians) of how our behaviour is observed and shaped by social networks and big tech: and by taking all obvious and easily available measures to reduce and remove the information they have and continue to gather.
One obvious way is by not sharing private sexual data with a major corporation, where not absolutely necessary.
Edit: Oh, and since a reference to the Luddites is a regular thought terminating cliche on HN - https://www.currentaffairs.org/news/2021/06/the-luddites-wer...
> This is a ridiculous attitude, exactly equivalent to 'if you've nothing to fear you've nothing to hide'.
Quite the contrary. My attitude encourages one to actually solve problems that we are encountering today, and not to fight battles that were lost two decades ago, which was the main thrust of my argument.
Also, as someone born in the Eastern Bloc, please allow me to throw up in my mouth at a publication that unironically cites the Communist Manifesto.
[dead]
[dead]
"We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. We realize this made it less useful/enjoyable to many users who had no mental health problems..."
I dread to think where society is going to be in 10-15 years. We've had a generation or two raised thinking porn == sex and the next generation will all be in romantic relationships with ChatGPT. Businesses have just given up on having any social responsibility whatsoever.
Yeap, it's started... https://www.johnlennox.org/shop/24/2084-artificial-intellige...
A sad name collision :/ https://en.wikipedia.org/wiki/2084:_The_End_of_the_World
FYI: Christian worldview
Full quote is this: "The questions posed by AI are open to all of us. And they demand answers. A book that is written to challenge all readers, no matter your worldview, 2084 shows how the Christian worldview, properly understood, can provide evidence–based, credible answers that will bring you real hope for the future of humanity."
Does 'the Christian worldview' invalidate any conclusions the authors may present?
Describing the "Christian worldview" as "evidence-based" suggests that a healthy dose of skepticism is warranted when evaluating the authors' conclusions, yes.
'The Christian worldview' is actually evidence-based, it's not based on childish stories of "ghosts" and "angels". Please, read the book first before judging by the cover.
As another reference, see https://news.ycombinator.com/item?id=42877709 (https://www.vatican.va/roman_curia/congregations/cfaith/docu...)