Forum
>
General / Nonfiction
>
Thread
Log in |
Register
Some people like real women, some people have kinks about unrealistic representations of women. Some like both. Most prefer free content.
I think the first group is by far larger and I don't think this AI stuff appeals to them. There's too much free content shared by models for paid content to be preferred, especially if you're trying to just recreate a similar type of content without leaning into some kink. Sora was losing millions of dollars per day as it shut down, and even with that, it couldn't hook enough users to monetize it in a mainstream way.
AI will have its niche in kink content that real models can't do, just like fmg comics have their niche, of course. I think people care less about the medium then they care that it's appealing to their kink. I also think people will prefer free content.
I guess if total fantasy is your thing, it will do the trick. I don't mind taking a look but it doesn't do anything for me because I know it's not real. My biggest complaint is when pics of already amazing REAL women are being doctored to make them appear bigger, stronger, bustier....whatever. It's a slap in the face to the hard work these women put in FOR REAL to suggest they're still not quite good enough. That's a red line for me.
This technology is only going to become more powerful, especially with quantum computing on the horizon. AI will become exponentially faster and more efficient. Of course, it's still in its infancy, but the implications are revolutionary. I'm very excited for the future of AI generated entertainment.
This technology is only going to become more powerful, especially with quantum computing on the horizon. AI will become exponentially faster and more efficient. Of course, it's still in its infancy, but the implications are revolutionary. I'm very excited for the future of AI generated entertainment.
Quantum computing will have zero impact on AI. Quantum computing isn't a better form of classical computing, it's simply different with much more niche applications to things like cryptography, sensing, etc. Quantum computing is actually much, much worse than classical computing when it comes to handling massive amounts of data. Think of quantum computing as simple data, big math. AI is simple math, big data. They're incompatible.
Quantum computing will have zero impact on AI. Quantum computing isn't a better form of classical computing, it's simply different with much more niche applications to things like cryptography, sensing, etc. Quantum computing is actually much, much worse than classical computing when it comes to handling massive amounts of data. Think of quantum computing as simple data, big math. AI is simple math, big data. They're incompatible.
Well, quantum computing has the potential to address the current challenges in AI. And I never claimed that quantum computing is a better form of classical computing. Yes, it's true that quantum systems are worse than classical systems at processing large amounts of data, currently. Yet, there isn't a law of physics that forbids quantum systems from becoming better or as good as classical systems at handling big data. Researchers are working to solve this bottleneck in a variety of ways ranging from QRAM to hybrid integration. However, I don't know how you figured that Quantum and AI are incompatible when they can mutually advance each other. Major tech companies and research institutions have already begun investing heavily in Quantum AI. Anyway, this is off-topic.
It’s gonna happen a lot earlier than 10 years — closer to 10 months than 10 years.
AI can already produce breathtakingly muscular depictions of girls, but right now it’s only working from a very small bank of natty/large base data — and doing so with what is (arguably) an equally limited amount of processing power. Once that data gets fully ingested and the analytics refine further — i.e., once the AI figures out:
1) where the most aesthetically pleasing insertion points are, and
2) what (inter) muscle proportions appeal most to the wider population —
and then applies that information to generating GWM…
WOW.
Want an idea of what Taylor Swift would have looked like if she went the Tina Lockwood route? You’d have it. You could even test it for lifelikeness: take a couple of mainstream celebs and create muscular versions of them, plus take a couple of Vladas/Speegles and create demuscularised versions of them. Show all the video pairs to a Kalahari Bushman who’s never seen any of those women, and ask him which version of each looks more like a real person. Run the same test on a dozen bushmen.
Their hit rate won't beat random chance…
It won’t.
Well, quantum computing has the potential to address the current challenges in AI. And I never claimed that quantum computing is a better form of classical computing. Yes, it's true that quantum systems are worse than classical systems at processing large amounts of data, currently. Yet, there isn't a law of physics that forbids quantum systems from becoming better or as good as classical systems at handling big data. Researchers are working to solve this bottleneck in a variety of ways ranging from QRAM to hybrid integration. However, I don't know how you figured that Quantum and AI are incompatible when they can mutually advance each other. Major tech companies and research institutions have already begun investing heavily in Quantum AI. Anyway, this is off-topic.
It absolutely does not have the potential to address current challenges in AI nor will it be useful within that modality in the future. I work in the field, you clearly have no idea what you're talking about.
It absolutely does not have the potential to address current challenges in AI nor will it be useful within that modality in the future. I work in the field, you clearly have no idea what you're talking about.
So, you know more than the big tech companies investing in the research then, too? You've not even substantiated your claim.
> mittengoddess: > > You're correct that it's not economical, but subsiding and operating at a loss for years in pursuit of profitability isn't new. > > Uber was founded in 2009 and wasn't profitable until 2023. Investors burned through several billions for 14 years. > > NIMBY resistance to data centers is more likely to present an obstacle; but that often loses to brute force.
In general you're not wrong that companies will eat losses to gain market share, but that doesn't apply here. There are severe constraints at basically every step of the AI supply chain. Every hyperscaler is capacity constrained, there's a shortage of memory chips, there's not enough power to keep the data centers running, there's a shortage of helium, etc. OpenAI shuttered Sora after just six months because it is not cost effective. They desperately need that compute for things with a higher ROI. The party with subsidies was always going to end, but the timetable has accelerated dramatically due to the reasons I listed. Uber is a really bad example because 1) they were not similarly constrained and 2) their capex needs were significantly smaller. For example, Uber's 2025 capex was $336 million. OpenAI's 2025 capex was $8 billion. Entirely different situations. This has nothing to do with strategy and everything to do with scarcity.
Intriguing that you've run everything through an AI, yet know nothing of what you've pasted. And been allowed to get away with the deception, as nobody has yet thought to run your AI genned output through another.
We live in strange times. A fool makes a fool of so many.
No. It is not economical. The only reason ordinary people can afford LLMs now is because they're heavily subsidized. That will quickly end as generative video is a huge waste of compute.
Do you imagine that computer technology isn't going to progress, making computing power cheaper and cheaper... you know, the way it has literally done since the dawn of the age of computing?
Do you imagine that computer technology isn't going to progress, making computing power cheaper and cheaper... you know, the way it has literally done since the dawn of the age of computing?
Moore's law hasn't really applied since 2010. And we could easily go backwards in compute capacity for 10-15 years if those invasion barges China have built aren't for show and something exceptionally stupid happens in the south china sea.
The SOTA do-everything models are a non-commercially viable dick measuring contest. The future is likely hundreds of small highly specialized models built on more carefully curated training data. But the obvious use case of using AI to control a commercial VFX package to sculpt / texture / rig a character then animate a scene just doesn't work yet.
Everyone here is worried about the implications this will have on porn consumption, yet no one is concerned about the impact it will have on the workforce.
And that's in keeping with the topic. It's nice to see a thread able to stay on topic here.
Personally speaking, I'm not unconcerned or disinterested regarding the workforce ramifications. Just sticking with the OP topic.
No, but I am worried it might gouge the female bodybuilding business all together.
If people come across muscular females via shitloads of AI postings before they get to know the real people, they will consider it as artificial. Even more so than the majority of people is doing now.
there are other social dynamics at play.. even before AI started becoming mainstream as it is.. the web and specially social networks were already dealing addictive fixes to both genders..for us it became the conduit for visual stimulation.. for girls, the stimulation has been delivered via attention and likes (just to test the theory, ask any high volume IG/tiktok/even OF girl to consider stopping their activities and watch them get triggered) ..
I see AI supplying this fix to both sexes especially those who will prefer it as a shortcut or as a substitute to real social interaction..
Eventually AI for imagery is going to hit a wall and not be allowed to get too "good" (I personally think it's already gone too far. It's the new "that's photoshopped"). Once people with actual power start getting fucked over by it they aren't going to mess with digital watermarks or any of that. They're just going to say to the nerds "just because you can do something doesn't mean you should, we're shutting this down." Obviously organizations like the CIA will retain these capabilities but for everyone else it'll be like having your own oil company; incredibly inaccessible.
For other uses AI will hit a wall of nature. Despite Roko's Basilisk bullshit and Singularity dreams of autonomous computers, at the end of the day you can't create "life" (including consciousness) out of non-biological components and you can't create anything independent of the creators of it. Somebody's gotta program it.
There will probably be a short ai dark age before the tech becomes cheaper. some people in 1999 tried to make youtube like sites .that collapsed because of the costs once the investors gave up. but a few years later it became more feasible so its probably going be like that