Earlier today, Dr. Gumby brought to my attention that Johnny Moronic had covered a seldom-seen alternate cut of Alien Visitor (1996) that included quite a bit of extra nudity from Ullie Birve.
Here is an example of a brief but excellent scene. (Apparently only available in DVD quality.) In the cut seen more often, her lower body appears only momentarily in this scene, too briefly to see without the pause button. But you get a lingering look in this version. While the real clip is not the topic at hand here, we can’t fail to note that Ullie had a great figure. You have to love the way her large natural breasts bounce.
Anyway … I thought to ask AI for something way beyond my video editing capability – “Imagine this scene in daylight, with natural skin tones.” Here is the result. I have to stress here that this scene never existed. It is a complete fake, a product of pure AI invention, loosely based on the real scene above.
But it’s just about perfect. The only real flaw in it is that the guy, Syd Brisbane, speaks with an American accent. He’s really an Aussie, as you might guess from the unlikely name of Sydney Brisbane. The accent could have been fixed if I had given the proper instructions, but I had no idea the fuckin’ AI thing was going to invent dialogue.
I was skeptical of you prophets of AI doom, but you Cassandras are absolutely right. Your paranoia is not paranoia at all. It is merely a recognition of reality. In a year or two, we will have no idea whether an actress got naked. We won’t even know whether she was really in the movie with the other actors, or was just strolling around her back yard for images to be altered later and inserted into the movie’s context. Shit, I might put myself in Casablanca in place of Bogie!
Scary shit.
Say, I wonder if Sydney Brisbane’s middle name is Melbourne.
Impressive fidelity. What AI is this?
vireels.ai
Amazing! Scary amazing, but still amazing.
What ai product did you use ?
Vireels.ai
Dumb question, I guess, but is this Vireels.ai something you have on your own computer, or is it something that requires a server farm, and your own computer just communicates with it? I suppose that now, server farm is a far more likely proposition, but I don’t know how fast things have grown,
The simple answer is that I do nothing on my computer other than upload a picture or video, plus instructions.
The more complicated answer is that Vireels appears to be a middle-man that resells the technologies of others. To be fair, they also have their own home-grown AI engine. When you submit your videos and instructions, they offer you a choice of AI engines. Since I don’t know jack about those engines, I choose one at random, more or less. Nothing is done on my computer. The processing is done on their computers or, in the cases where they are re-selling as a third party, at the data centers of the originators. Somebody mentioned in this thread that the engine used to create the video in this post is Seedance 2.0, and the processing is therefore done in China.
You should understand that I do nothing and use no skills or talent or computing power. It’s not a case of my laboring away on a photo, tinkering with transformations in Photoshop until I find one that looks right. It’s simply a matter of me, or anyone else, uploading a picture of monkeys, then telling the computer to make the monkeys dance on a two-dollar bill.
Thanks! That’s exactly what I wanted to know – whether this was stuff people could have on their desk already, or if it still took major computing muscle. I think the fact that is does is a bit comforting.
Scoop, just curious about this. I’m trying something similar and it keeps saying “content sensitive” failed. wondering how you got it to produce nudity ? Thx!
Never mind. Got it! Thx.
AI itself is neither good, nor evil.
In this case, AI is being used for good.
And much like other extremely powerful things, it should be regulated so terrible people don’t send videos to women of them being tortured
Gonna have to take it up with China then. Because I can guarantee you this clip was not made on US soil.
I think Vireels is an Aussie company, although that doesn’t exclude Japanese or Chinese ownership.
I saw they have Seedance and Kling as options, both Chinese AI companies. And that’s why you are able to make videos of celebrities without getting moderated – these AI’s are basically trained on movies and film without worrying about copyright lawsuits the way the US LLMs are. Try this on Google or Midjourney and you’ll get shown the door.
There you go! I think I used the Seedance option!
I just made a post, and I just realized that what I said in the wordy paragraphs, you said in one sentence. That is vastly more effective, and I thank you for doing it.
Yes, an inanimate object is neither good nor bad in itself. As I also learned in Philosophy 101, it has no “ding an sich”, i.e., inherent morality. That is not the question. The question for a society is, even if it a thing is used for good ends, what are the side effects? Are those bad, and can they be controlled? And what are the consequences of it being used for bad ends, and can that be controlled?
This point occurs to us most readily with weapons, which is why people are not allowed to own tanks, artillery pieces, hand grenades, or nuclear weapons. But it also applies to things which pollute, like burning coal, or are detrimental in other ways, like the use of radium on watch dials, lead in paint, or arsenic in green wallpaper. Heck, the use of white phosphorus for making matches was banned as far back as the 1600’s because it was killing workers.
We need to have the same discussion with regard to AI. Or at least we need to debate if we need to have that discussion. Certainly the billionaires who are betting the US economy on it are not doing that; they are only dreaming of how rich it might make them. And they have amply demonstrated how little they care about the rest of us, IMO.
We’re heading towards yet another social contract restructuring, just like it’s always been in history. Musk is about to become a trillionaire in net worth, and Altman probably in the nine figures once the IPOs for SpaceX and OpenAI hit the market. The way risk-reward, merit, ownership, and hierarchy are structured in society is getting to a tipping point.
Every chart in the world will show you how income inequity has gone despite gains in GDP since 1980, and any social scientist will tell you how much income inequity is tied to the hip with recidivism, crime, drug use, etc. Hell, you can look up the mouse endurance swim experiment to see how mammals react naturally when they lose hope.
These people are not that smart, they’ve just become adept opportunists in our casino culture and rely upon capitalistic administrative mechanisms to keep accruing more in a snowball effect. If it wasn’t them, it would be someone else, just because of the flaws in the system. Take away whatever the IPO worth Musk and Altman will get going forward from the employee base, tax credits, and collective knowledge accrued through taxpayer funding of federally funded universities and labs and you’ll see where the ‘worth’ is here. We’ve only defined what they’re allotted through our flawed systems, not any skill, ability, knowledge, or effort they’ve provided to do anything more than be a glorified project manager or overseer of their companies.
There’s some irony in the fact we’re always saying China is passing us, even with a repressive totalitarian state, because they’ve reached far more in collectivism than us when we’ve started from such a greater position. People can’t get out of their people worship, with the top 0.1% and their addiction to greed, or the individuals all along the chain below them who consider themselves disenfranchised future millionaires only a good idea or a lottery ticket away from vapid luxury. It’s pretty damn sad – we’ll either learn or not exist in the future. Simple as that.
Well, I guess: a) yes, those are the two options. And b) we won’t learn. That’s the bottom line from our track record. All I can really add is, I think whatever happens will happen in slow-enough motion that it won’t pan out one way or the other while I’m still around.
When Generation Alpha and later generations grow through the decades, they’re going to look at the 20th century and this time with similar disdain as those today’s look at plantation owners in the 1800s.
So many of the right-wing lunatics retiring the past couple of decades took advantage of New Deal Keynesian economics to make a life for themselves with ready-to-go training, cheap colleges, cheap housing, and big government programs that gave them a living wage and a home to build a family post FDR and then flipped it with Reaganomics to profit heavily via investments during the tech boom with heavy returns off others labor and lax intellectual property laws during the 80s.
We’re being torched by a similar repeat with the tech billionaires via web boom of the 90s and consolidation of corporations through mergers and acquisitions and stock buybacks. All with credulous claims of goods or services that provide any sort of real value to society, so the pyramid schemes of the stock market can keep rising so more people on top can pull the rug out from those under them. All because their ego believes they’re ‘gods’ rather than giving ample credit and resources to the labor and consumer market as well as the inventions and progress funded by the taxpayer over generations intended to give back to society – not to their wallets.
Meta is an appropriate name for one of the companies causing the downfall with their huge datacenters. Because that’s exactly what this economy is – its playing the ‘meta’ – the administrative laws of hierarchy, ownership, intellectual property, and so on that have been carefully manipulated by those on top to continually accrue resources without providing anything of real merit in return.
We literally have the same technology from years ago going UP in price while the biggest value proposition we’re getting in today’s boom is the ability to cut jobs, hope to achieve vendor lock-in through reliance on AI, and then raise licensing pricing to achieve gains.
At some point down the line, generations will look up and see what the French did a couple hundred years ago: a modern tech aristocracy without merit, ability, skill, or effort who simply played the game like a Vegas casino setting the rules for Blackjack.
As George Carlin once said, they call it the American Dream, because you have to be asleep to believe it.
I comments on this in another post on this site earlier, Scoop. It’s crazy that we really won’t know if a nude scene is real or not in a few years lol, maybe that’s already happening and we won’t even know. This moment we’re living in is totally insane
A few years ago I told you this was being done and you said I was wrong my cousin is producing films and the nude scenes you thought were nude scenes are all Ai generated and why do you think actresses don’t care anymore if done right they are better than they would have looked. bad example is Jean Smart done horribly.
If that is actually AI, it is scary good. But if Scoop can do that on his home computer with a couple of simple instructions it makes you wonder why movie animation and CGI is not better and still creates very noticeable ‘uncanny valley’ effects. Something feels off in that respect.
And also, if it is AI, it still screwed up based on the instructions given. Don’t trust it folks.
That is absolutely not what I expected to see when it was finished, but I forgot the cardinal rule – always say “minimize creativity, err on the side of realism.”
So I basically let it run wild. But damn, did it do it well. I was dazzled. “Credits” cost a penny, and that job was 800 credits. It made that video for eight bucks, and did so in a few minutes.
Eight bucks.
It looks as good as a network TV show. Sure, it’s only five seconds long, but that still works out to only a hundred bucks a minute – $12,000 for a two-hour show.
You can see why it will take everybody’s job.
And, as I told you last week, it gets better almost by the day because of how fast it can learn. Remember, it can learn from us and from its own mistakes, can think millions of times faster than we can, and knows – well, basically everything there is to know. Think of everything mankind has learned from the beginning of the stone age. Super-computers can learn all of that in days, limited only by how slowly we can feed it into them. 20,000 years in a few days. And in a few more days, they may be 20,000 years ahead of us.
Why does this little amateur clip look better than some professional shows you have seen? Because they were done last week, or last month, or last year. And two years ago the whole thing was a joke.
Even scarier than the fact that I can do it with some simple commands is the fact that you can do the same. My editing skills and years of Photoshopping, give me ZERO advantage. You are now as good as I am.
I will point out one thing regarding AI in this situation: It’s not creating anything. It’s simply taking something that already exists and editing it a bit.
Trying to use this stuff for original work is a whole different ball game. The camera doesn’t go where you want. The people look the wrong way. The guy in the background does something weird. The doors open the wrong way. Dozens of little things like that crop up in every single shot. And yeah, every AI still has that 6 to 8 second limit. Sometimes you can stitch multiple shots together, but there is almost always a perceptible shift at the cut.
Not to say this stuff isn’t amazing, and the porno implications are wild. But I still think the idea of this replacing Hollywood is far-fetched. It’s another tool in their arsenal, just like CGI was.
You are underestimating how fast it learns. Those mistakes it made last year are not made this year. This year’s errors will be gone next year.
They aren’t using websites like these. There are ways to do this offline without limitations. Lots of processing power though. And much higher learning curve.
And this is why every acre of farm land, every computer hard drive and gpu, every drop of water, and every kwh of electricity in this country will soon be devoted to data centers. So that random nobodies online can train AI to make fake porn. Yay. We didn’t need to eat or drink anyway.
That is not the necessary future, but it is a possible one.
I still have the feeling that something must be ‘off’. Maybe I am wrong – probably I am – but my belief until it is specifically proven otherwise to me is that the video you posted was not ‘whole-cloth’ created from AI. Sure, it was the product of an AI instruction, but I just can’t believe that it was put together as an originally-created video made pixel by pixel by machine. Who knows the manner in which AI complied with the instruction, especially since it did not actually comply all that closely with the details of what you asked? It may have just scoured what is out there in the digital world, found a deleted outtake from that movie (we know now that they exist), maybe made some small alterations, and presented you with that. That is mostly what AI does: it searches for, recompiles, and feeds you in altered form what is already out there, rather than creating something truly new.
Again, given how even professionally-produced films, with millions of dollars in budgets and access to high-end computer programs, try and fail, so far, to make original images of digital humans look entirely real, I have a hard time believing that a home computer could do it. But probably I am wrong. If that is the case, then among other implications, acting careers – and many other careers – in the film business will very soon be finished. Because a talented director using now commonly-available tools will simply be able to create all the necessary images by giving instructions to AI. The movie stars of the future will simply be digital creations.
repeating here, but I mentioned this to Scoop years back, this very thing taking place. You’re correct, nobody will be able to tell what is real and what isn’t. Here now as this proves. If an actor doesn’t publicly state it’s real, nobody will ever know. We’ll assume it’s fake.
Back then I mentioned the need for a new type of rating or classification, which indicates ai is not being used. Or is being used, one or the other.
Authenticity will remain but how do we authenticate? What will the market be? AI content is a cheap commodity. It will likely be valued as such. Doesn’t mean no value, just less value.
I have to believe a classification system of some kind is being looked at by the industry. I haven’t kept up, so I wouldn’t know.
Other issue with clips like this, if she saw this, she may not be happy. It’s a slippery slope at some point, with regards to the actors and the manor in which their likeness is being recreated.
True. I used it without her permission, and she might be upset, although I had no idea that AI was going to create a whole new scene. I was just expecting a slick edit of a scene she had actually performed..
But as we go forward, we won’t have to worry about the actors being upset because we won’t need them at all. . The computer will not need a direct human model as in this instance. If I want to, or if you want to, you can probably take that scene above and tell AI to make it between a guy who looks a lot like Kevin Costner and a woman who resembles Sydney Sweeney, but not exactly. But don’t use those voices. Make the Costner guy sound like James Earl Jones with an Irish accent, and Sweeney sound like Kathleen Turner.
Or whatever else you would like.
Oh, brave new world that has no people in it.
Let’s face it, you can probably tell AI to create a slicker, better, funnier version of Other Crap, and it would probably be better than my page. It would not be me, with all my quirks, but that doesn’t mean it can’t be better than me. And in the extreme case, you can probably tell it to mimic my style, and you’d never know I was gone.
Is that your plan of succession?
I actually died in 2023, and AI has been writing the page.
or you faked your death and are hanging out with Andy Kaufman, while N. Robert “Scoop” Parking continues the site pretending to be “you”
The clip is impressive. Wow.
Yeah, remember when Fred Astaire dancing with a vacuum cleaner was as bad as it got?
Also, Ullie Birve’s middle name is Kalgoorlie.
Now our friend Dawid’s concern will be whether the penis scenes are AI or real, and whether the actresses’ pubic hair is AI 😂😂. And he would be right in that case, ironically lol, it’s a question we’ll all have in the coming years.
Nude scenes haven been being digitally altered, manipulated and faked for years, particularly within American entertainment. It’s nothing that new.
Yep, I have no doubt that many nude scenes that we thought were real in recent years and decades are actually CGI, prosthetics, etc. And this isn’t just in Hollywood; today we know that it has even happened in European films countries.
But you know what? It’s not a new concern in celebrity and nudity forums, lol. With the advancement of CGI and realistic prosthetics in recent decades, many people already worried about whether nude scenes were real or not, whether much of it was prosthetics and CGI. Now the focus will just shift to AI. But this discussion has always existed.
It’s going to put out-of-work everyone that creates media.
Artists
Writers
Actors
Musicians
News broadcasters (broadcasters of any kind)
Plus all the support jobs for all the above such as camera and audio people, etc
(I’m sure there are thousands of jobs I haven’t listed.)
It’s also going to change elections. How hard would it be to show a politician accepting a suitcase of money on a grainy film? or getting with a 9 year old? or just speaking about betraying the country? What is real??
However, I look forward to the days when people can create their own full length movies or TV shows. An x-rated episode of Friends? why not. A Pixar quality movie in the Marvel universe? sure.
I’ve already seen scenes of sitcoms that have been nudified. It’s happening now and more is being created every day. Most of it looks AI, but some are approaching can’t-tell-the-difference quality.
I can tell you from personal experience that it’s being used extensively in the broadcast and advertising fields already, but it’s mainly replacing stock footage. Instead of digging through Adobe or Istock for that perfect video, we can just generate it ourselves. Putting photographers out of business for sure, but the day where we can create a full ad in a single prompt is far away. We’re still doing all the editing, graphic overlays, animation etc.
Two things are going to hold this up. One is cost, it’s not going to get cheaper. All of these AI companies are losing billions at the moment to get us hooked. They ARE going to jack up the prices once we’re reliant on it. And this isn’t speculation, check the Adobe AI EULA when you sign up. It specifically says all the current prices are temporary.
The second is copyright. AI generated assets cannot be copyrighted. This might not be a big deal for social media influencers, but it is a deal breaker for big companies. They will still need humans in the pipeline even if it’s just for show.
The current mantra in the industry is “AI will not replace you, but someone who uses AI will”. It’s ironically the people hating on AI and protesting against it who are going to be victimized the most by it.
good points.
Also I left out programmers, such as video game creators.
DeepSeek (open source) is dirt cheap because it doesn’t need the expensive AI data centers to run. The Chinese have developed algorithms to greatly reduce the costs of AI training. US companies are behind on this.
Imagine that sinking feeling you would have if you are just about to graduate college and you have studied 4 years in a field that is about to be replaced by AI.
I would think that colleges must be in chaos right now with students (the ones paying attention to AI’s impact) request to change majors.
Hence, see the boos the commencement speakers on AI got.
From a tech point of view, I think it’s best to think of it as Aggregate Information. It’s up to the end user to verify context in prompting and also apply discernment with the information it gives out. I’ve posed technical questions to the various internet models (Claude/Anthropic, Google Gemini, ChatGPT and CoPilot based on Open AI) and it’s funny that at times you’ll get disagreements on answers to questions when given options. Then you read the explanations and without context they ALL sound reasonable with a reasonable explanation. It can be a good starting point to wrap your head around something, but still has to be verified and isn’t a source of truth.
These LLMs aren’t thinking, they’re trained and aggregating and spitting out in the patterns they’re trained on – seeded by Reddit, Google searches, etc. I have a pretty eclectic background and have used the same areas these models have trained on to research myself, and you still get variance in answers from the usual suspects.
In the end, its just another tool, no matter how the tech companies wish to market it. And just like any licensed product or service, cloud infrastructure, etc – they want organizations to get locked into it and will eventually come calling with big price increases. The ironic thing is, we’re just layering this thing called AI that may provide some knowledge to the user when utilized correctly when if most industries would come together and drop all the bullshit competitive advantage/trade secret/IP laws etc we could accomplish that by information sharing with standardized methodologies and frameworks to provide interoperability and efficiency.
Of course, that would be unheard of in a society that treats businesses like a casino for unlimited return on ‘investment’, but it just goes to show the way AI is utilizing is simply exposing fundamental flaws in societal order.
The AI is now beyond the LLM stage.
I don’t think I’m as terrified by AI as I ought to be. Essentially, it’s still an imitation of intelligence. It is not – apparently – self conscious or self motivating to the same extent that humans are. It just simulates self-consciousness.
In my real life, there are a couple of minor and unimportant historical subjects – specifically focused on WW2 and Nazi Germany – on which I am the undeniable world authority, and have published several books etc. I fed queries about these – separately – into various AIs and the responses were pathetic. A number of hallucinated academic papers and authors apparently ‘balancing’ my work, but which don’t, in reality, exist. At the same time, AIs were producing magnificent reviews of my books. From what I can see, they just aggregate information and then apply some filters to it; they make shit up to simulate debate; and essentially they just fluff us.
It’s great that AI will generate celebrity porn: I’ve always wanted to see Princess Diana in a threesome with Genghis Khan and Joseph Stalin, but I’m not going to believe it’s real. My Insta feed seems to be about 95% AI porn slop which is really tedious. Porn only really works if the people are real.
I’ve had a similar experience. I suppose I am the most knowledgeable person about the Black Sox. When I ask an AI bot questions that I know the answer to, it tends to spit out the common misconceptions instead of well researched facts and suppositions. But I can’t stress too strongly how well a good model learns. The answers it gives you next week will be better, and they’ll be still better the week after that. Eventually they will eclipse us. That is inevitable.
The problem is that the general population does not know that AI injects made up crapola into its answers. I know it. You know it. But I bet 98% of people will believe what they see. I have relatives that try to show me AI videos that are clearly AI that they believe are real.
I’m not scared of AI. In fact, I think it’s a fun thing to screw around with, but I do think it will replace a ton of jobs in the media creation fields. Does it bother me? Not really. I’m close to retirement and my kids are in medical. They’ll be fine. However, if I were a romantasy fiction writer, I’d be worried. AI can churn those things out in a few minutes. All it takes is a good editor to “co-author” it.
Is anyone here very familiar with Kurt Vonnegut’s first novel, “Player Piano”? Do you think it is relevant to the future of AI that is being discussed here?
I am not much a of a Kurt Vonnegut reader. I respect him, but I don’t enjoy reading his stuff. I read a plot summary of “Player Piano”, but it kind of reminded me of Fritz Lang’s silent movie “Metropolis”, with the son of a powerful man being the protagonist.
Similar questions are now being raised. In that story, the rebellion against the machines failed when the mass of humanity demanded that the destroyed machines be rebuilt. The human dependence on the machines proved to be greater than the human desire to fill life with purpose and originality.
Vonnegut was into a deeper truth than that confined within the details of that story. Throughout human history, technology that might have made us better has just made us lazier and more pleasure-seeking. The greed of those who own the technology abets our sloth, but has only been a small part of that problem. The REAL problem is our inevitable willingness to let the new technologies appeal to our basest and laziest instincts. We knew that television could turn us into a nation of couch potatoes, yet we submitted. We knew the risk of allowing the internet and our phones to have full control of our lives, yet almost all of us submit to it because those things make our lives so much easier and more entertaining. We know the risks of ceding ever more of our autonomy to AI, but we again submit because, as always, we love to watch the monkeys dance.
Thanks again. I think what you say explains a lot of the public reaction (or, rather, lack of reaction) to Trump’s dismantling of the lawful government of the United States.
I think I would find reading the book depressing right now. But I am impressed by your knowledge of it, and ability to write a summary of the themes it deals with (from your point of view) in short order!
I was a huge Vonnegut nerd. Before Bill Shatner and George Costanza as inspirations in my life, there was Kilgore Trout. I loved Vonnegut like Colbert loves Lord of the Rings. Unfortunately his work has been turned into movies I don’t like, except for Slaughterhouse Five, which I did not like in 1972, but now love because I have come unstuck in time.
At least they didn’t make a movie of Cat’s Cradle. Things to see if you haven’t yet – Venus on the Halfshell ‘by Kilgore Trout’ (actually Philip Jose Farmer). A lot of fun but it pissed Kurt off, so wasn’t reprinted, worth a look on ebay.
-The early 80s Rolling Stone interview where he grades all of his earlier books.
-God Bless You, Dr. Kevorkian. He could have just cranked out another recycling of his good stuff and made another million, but I got one page into this and like goddammit he’s BACK. Then he died.
And so it goes…
Although Vonnegut himself compared it to “Brave New World,” “Metropolis” is the better comparison. The author of “Metropolis,” Thea von Harbou is a bit of an enigma. Until her dying day, there were three framed photographs in her apartment which she dusted daily: one was of Ayi Tendulkar, another of Mahatma Gandhi and the third of Adolf Hitler.
I still have my doubts about what exactly that video is.
Anyway, as a funny antidote to all of the above – to the idea that AI in its current form can ‘take over’ – here is a video showing ChatGPT trying to count to 65:
The public version of ChatGPT is a joke. The difference between the paid and free version has become enormous. Ask Claude or Gemini.
You know, I don’t know enough about AI to even know if Figaro’s last line is a joke.
To put it another way, judging ChatGPT from the free version is like judging a film from a teaser. It’s true in some cases that a particular movie trailer contains every good foot of film from the whole movie. That’s not true in general. Usually there’s still some meat left on the bone. Likewise, while AI is still pretty much slop, & will be for some time, that fact is going to get harder & harder to detect, & faster than we thought not so long ago. We were comforted by that thought, but it was wishful thinking. Harder to detect the fakery is going to cost a lot of us little guys a lot of our bargaining power.
Like, just the idea that vibe coding is going to radically change a coder’s workflow, even though it won’t replace a single worker, is still a harbinger of a possible RIF in many workplaces. It won’t completely get rid of SaaS, but it could enable many small businesses to wing it, instead of having to fork over big bucks for a service contract.
One thing I have noticed using the free tools, & I don’t have any experience with paid plans, is that implicit time limits can lead to some kinds of hallucinations. I believe I’ve seen accounts of being able to give them more time in prompts. They would’ve been on paid plans, & I haven’t tried the same in free.
Anyway, back to what I noticed. The engine is multitasking its information gathering with its presentation. It has a running outline which as it decides what it’s going to say, shapes how it will say it. When it decides the time has come to output its work product as is, both the outline & its conclusions are incomplete. You can then see some of its placeholder material, which may be made up: hallucination.
I’ve seen output where the engine was in the middle of considering, say, network drive paths as a possible last resort. In the followup, most of the options it had outlined saying it would give details if I want, were taken suddenly off the table, once it had found out that what I had asked for (a host OS file-backed emulated USB stick in a virtual machine guest OS) could be done easily with a loopback device, it suggested that & only that. All the other ideas were effectively deprecated.
And, of course, that seems like what we do, too. So it’s a bit unclear if this kind of hallucinating is a bug or a feature.
Getting back to the nekkid ladies…
After a few high-profile cases of AI-faked nudity (Jean Smart was just the beginning), I predict a pendulum swing back to actual bodies. Just as there are purists who will only shoot on film, I can see some actress in an interview stating “That’s all me in the nude scene. I worked hard for that body and I’m proud of it. Oh, and it was necessary for the character.”
I hope that you are right.