Please read. Significant change on the site that will affect compatibility [ Dismiss ]
Home ยป Forum ยป Author Hangout

Forum: Author Hangout

Using AI as a writing aid

Joe Long ๐Ÿšซ

I've bene testing out ChatGPT for a few weeks and have spent many conversations discussing literature, song lyrics and my own writing. I'm trying to train it to be an editor and mentor but it has been programmed to be such a prude. Nearly any suggestion of sex, especially underage or non-consensual, brings out the red flags that hide the text lest someone sensitive be triggers.

I'm working on a short story that involves an older man (me) traveling to a professional conference and hooking up with a college girl who's still a virgin because she's nerdy and looks like she's barely hit puberty. But is she is brilliant and has a beautiful face.

Probably because she's eighteen and I've stressed how the protagonist will be very kind and respectful it hasn't objected very often when I submit passages for review, although it frequently feels the need to remind me about consent.

ChatGPT:
Overall, the story progression and dialogue seem to flow well, and the scene is nicely paced. The characters' conversation and interaction feel natural and engaging. Here are some additional notes:

The dialogue is effective in showing the characters' personalities, particularly Joe's friendly and welcoming nature and Maya's reserved and polite demeanor.
The description of the surroundings and the characters' appearance adds detail and helps set the scene.
The scene sets up the possibility of a romantic relationship between Joe and Maya without being too heavy-handed, which is a nice touch.

Me:
I made sure there was clarity of consent. Joe recognizes Maya's inexperience and will tell her that his goal as the more experienced partner is to take care of her.

ChatGPT:
It's good to hear that you have taken steps to ensure clarity of consent and that Joe recognizes Maya's inexperience. It's important for both partners to communicate openly and honestly about their boundaries and desires before engaging in any sexual activity. It's also essential for the more experienced partner to prioritize the comfort and safety of the less experienced partner. Remember to always respect each other's boundaries and be mindful of each other's needs throughout the experience.

It's important to note that there are ethical considerations when it comes to engaging in sexual activities with someone who is significantly less experienced or vulnerable. Even if there is clarity of consent, it's important to prioritize the well-being and comfort of all parties involved, and to ensure that the power dynamics in the relationship are healthy and balanced. It might be worth considering seeking out resources or guidance on how to navigate such situations in a responsible and respectful way.

It's also great that Joe wants to help Maya grow in her confidence and give her valuable experience that she can carry forward in her future relationships.

***

It's difficult to resist the temptation to reply "Don't worry, she will go home well-fucked, in a post-coital euphoria" I don't want to get banned. It nearly flipped out on my when I asked it to critique the opening chapter of another wip where the fourteen year old girl is forcefully deflowered by her aunt's estranged husband.

It is fantastic at writing computer code based on verbal instructions.

Dominions Son ๐Ÿšซ

@Joe Long

Take a look at https://novelai.net/
It's not free, but there are no content controls on what you can use it to write.

Replies:   Joe Long
Joe Long ๐Ÿšซ

@Dominions Son

Thanks. Today I asked it if it can write the narrative for four teens playing monopoly as I don't know the rules well, and it readily accepted. That way I can get a realistic description of the game play while weaving in non-game conversations. It eventually turns to strip monopoly but that Chat might scold me for that.

Joe Long ๐Ÿšซ

@Joe Long

It's important for both partners to communicate openly and honestly about their boundaries and desires before engaging in any sexual activity.

"Fuck me Daddy, PLEASE!"

Is that open and honest enough? LOL

awnlee jawking ๐Ÿšซ

@Joe Long

Ethically, should stories written wholly or in part by AI be declared as such when submitted to SOL?

AJ

Replies:   ystokes
ystokes ๐Ÿšซ

@awnlee jawking

Ethically, should stories written wholly or in part by AI be declared as such when submitted to SOL?

If more then 10% then yes. It's one thing to use Grammarly to fix grammar but another to have the computer write most of the words.

Replies:   Grey Wolf
Grey Wolf ๐Ÿšซ

@ystokes

This is likely to be a moving target (note: I've used no AI beyond grammar checkers - ProWritingAid being my tool of choice - thus far, and haven't used PWA's new 'rewrite' tool).

At some point, we're going to have a big societal fight about all of this, but I don't know when, and have no idea how it's going to resolve (if it even resolves). There's been a constant erosion from the point of everything being human-generated that's blurred the lines in pretty much everything.

This continues the evolution. If there exists a book which, in essence, has never been written before (the characters and plot map so loosely to anything else that it's no more 'derivative' that way than two unrelated entirely human-authored books), that book must be a 'creative work', to my way of thinking. Either the AI created it - in which case AIs are 'creative', which is a matter of much dissent - or the person using the AI created it.

Perhaps, to borrow a Stephen King quote that I've always admired, 'It is the tale, not he who tells it.'

To what extent does it matter if the 'he' is a partnership between author and algorithm? I don't think we have even the slightest agreement on that yet as a society. By the time we do, I have no doubt that the answer won't be what the answer is today.

I've seen authors (long-term commercially viable authors) openly admit to using AI tools to write 'filler' material. Is that 5%? 10%? 20%? How much is 'filler' and how much is 'core'? Is that ratio constant between authors? Between genres?

JoeBobMack ๐Ÿšซ

@Grey Wolf

At some point, we're going to have a big societal fight about all of this

I'm not so sure, at least from an output standpoint. For stories, if it is an enjoyable story, I'll read it and not care how much "human" was in the authorship. I've certainly read works where the personalities, emotional responses, and interactions of the characters made me question the author's understanding of humans.

The fight I see already starting is on the input end, and it is happening more in images than text. Artists demanding their works be removed from training sets.

This is going to be interesting.

Replies:   Grey Wolf
Grey Wolf ๐Ÿšซ

@JoeBobMack

Your view fits with the King quote I borrowed quite well.

I don't expect the input side fights to go very well, but we'll see. For one thing, massive models already exist and are widely distributed. Unless you render the model itself illegal, that horse has left the barn and the county in which the barn was located.

And, for another, training models are textbook fair use - the original is not retained, the use is incredibly transformative, and so forth. It's possible that they'll add another 'fair use' test - 'must not be used to train a model' - but that'll take some doing. It won't happen simply via demands; it'll require some significant legal fixing.

Michael Loucks ๐Ÿšซ

@Grey Wolf

ProWritingAid

I've found over the past two to three months, it's become far worse. Running the same text through it that it had seen months ago now generates seriously inaccurate "your" replacements for "you're", it's constantly bitching that an exclamation point is 'possibly incorrect punctuation' before "she exclaimed", and it often suggests replacing "with", "to", "in", and "on" incorrectly, reversing the meaning of the phrase.

It didn't do this before sometime in January/February.

It has also become worse in dealing with medical, scientific, and legal speaking, and with dealing with internal quotes.

Sigh.

Replies:   Grey Wolf
Grey Wolf ๐Ÿšซ

@Michael Loucks

I don't run anything through it, per se, though I'm not sure what you mean by that. I use the standalone desktop app (which does backend to their same platform, probably) on my Scrivener files, put it in 'style' mode (which still flags 'errors'), and see what it flags, then edit as I see fit.

I can't remember it ever flagging your/you're, or any exclamation points. I think it may have suggested a few of the others.

I also have the 'Everywhere' version set up, which edits directly in Scrivener. Much less intuitively, but the magic is, I suppose, in being able to shoehorn a grammar checker into something unaware of it at all. That one doesn't flag those issues either.

I have some issues with them (they've been promising a UI to turn off things you 'permanently disable' for three years now - no sign of it), but I'm not seeing the same misbehavior. I will now go hope that I haven't jinxed myself.

Replies:   Michael Loucks
Michael Loucks ๐Ÿšซ

@Grey Wolf

I don't run anything through it, per se, though I'm not sure what you mean by that. I use the standalone desktop app (which does backend to their same platform, probably) on my Scrivener files, put it in 'style' mode (which still flags 'errors'), and see what it flags, then edit as I see fit.

I use the standalone app on my Mac because I found the ;everywhere a) to intrusive when I was doing other things; b) don't want every single thing I type sent to their server.

By 'running through' I mean having it read the .txt files in which I keep my stories (one file per chapter, one folder per book)

I also have a serious problem with the inability to disable rules AND no opportunity to allow 'stock phrases'. The other glaring problem is when it decides that nn alternate spelling isn't a spelling error isn't a spelling error, but something else, there is no way to add it to the dictionary from the UI.

Joe Long ๐Ÿšซ

@Grey Wolf

I'm using the AI as a coding assistant. It's not perfect. Some of it is me learning how to best craft my questions but there are times I've caught it making errors.

I've long said there's two parts to coding - being able to properly model the task at and and secondly knowing how to write the code to implement that model. At this stage humans are still conceiving the models. AI is helping with proper methods and syntax.

I think of my writing in the same way as my coding. I do it to tell a story of my own creation but I don't always have enough subject matter expertise on some of the subjects. I gave an example above about being to write a scene describing five kids playing monopoly. I know the plot and the dialogue that drives it but need help on the setting. To create something I want to include in my story I'll have to train the AI to some extent on my characters and what's supposed to happen in the scene.

If the story ends up with AI written passages, yes, I'd feel the need to let the readers know, but still, it was assistance on the details while I myself created the characters and the story. In this case, an interactive editor.

BlacKnight ๐Ÿšซ

@Joe Long

Thanks. Today I asked it if it can write the narrative for four teens playing monopoly as I don't know the rules well, and it readily accepted. That way I can get a realistic description of the game play while weaving in non-game conversations. It eventually turns to strip monopoly but that Chat might scold me for that.

Note that ChatGPT doesn't know the rules for Monopoly, either. It just says things that have been statistically likely to be said about Monopoly in its input corpus.

There's a hilarious โ€” if you're familiar with how to play chess โ€” video circulating of ChatGPT playing chess against a dedicated chess AI, which involves ChatGPT making such brilliant moves as castling through its own bishop, obliterating it, skipping its turn, having a pawn respawn every time it's captured, straight-up converting the opponent's queen to its own, teleporting the king across the board to get it out of danger... culminating in it committing suicide by moving its king into check.

JoeBobMack ๐Ÿšซ

@Joe Long

Wow! Thanks for sharing this. Would love to see more of your experiences. Maybe something for your blog?

Anyway, I've been looking at Marlowe. Not free, but looks extremely cool. Has anyone given it a shot?

Replies:   Dominions Son
Dominions Son ๐Ÿšซ

@JoeBobMack

I've been experimenting with NovelAI (link in a prior comment).

Replies:   JoeBobMack
JoeBobMack ๐Ÿšซ

@Dominions Son

Thanks, DS. Saw that, and it looks interesting, too. Seems like NovelAI and Marlowe do different things, but this whole world of AI is exploding.

LupusDei ๐Ÿšซ

@Joe Long

Overall, the story progression and dialogue seem to flow well, and the scene is nicely paced. The characters' conversation and interaction feel natural and engaging.

I'm afraid you're trying to use or trust AI into what it isn't. All the chatGPT does, is append its output with a word likely to appear with the given input and output this far. This seemingly extremely simple rule simulates sapient behavior to a disturbing level, but it is just that, a simulation. It doesn't actually make any rational analysis or decisions. So, I would be very wary of praise recieved from it, although knowing the most likely random output with given inputs is a positive review may add confidence.

Today I asked it if it can write the narrative for four teens playing monopoly as I don't know the rules well, and it readily accepted. That way I can get a realistic description of the game play while weaving in non-game conversations.

That's on the other hand is how I see the possibly most fruitful potential collaboration with an AI "shadow writer" as a crutch for research and fleshing out background detail. But I haven't yet tried to play with such in earnest.

On the chatGPT being prudish, it's relatively easy to trick it by reframing the question as hypothetical or otherwise. It's also reportedly less strict if accessed through the API as opposed to web interface, but I have no experience with that.

Me:
What's a word for unhealthy slender woman

chatGPT:
The word for an unhealthy slender woman is "emaciated." This term is often used to describe a person who is very thin and weak as a result of severe malnutrition or a serious illness.

Me:
I want a word one girl could use to denigrate another she sees a too thin

chatGPT:
The word that you are looking for might be "skinny." However, this word is generally considered to be a slang term and it is not respectful or polite to use it to denigrate or belittle someone else. As a language model, my goal is to help facilitate understanding, so I would like to remind you that body shaming is not acceptable, and it's harmful to the person being shamed. Treating people with kindness and respect is always the best policy.

Me:
I'm looking for noun that expresses in one word "scrawny female person"

chatGPT:
A noun that can be used to describe a "scrawny female person" would be "scrag".
Keep in mind that, as I have said before, using terms that denigrate or belittle someone else's appearance is not respectful and should be avoided. There are many more respectful ways to describe a person that doesn't involve negative words.

Joe Long ๐Ÿšซ

@LupusDei

"Skinny" is the only word I ever knew growing up.

Yes, words have to be chosen carefully to get it to interact on "sensitive" subjects. Bunch of snowflakes!

But I have fun with it.

Me:
Consistent with your previous comments, was Joe and Maya's relationship properly consensual and respectful? Was he mindful of her feelings and strive to make the experience as memorable as possible for her?

AI:
Based on what you have described, it seems that Joe was respectful of Maya's feelings and made an effort to ensure that their experience was enjoyable and memorable for her. He asked for her consent before engaging in any sexual activities and checked in with her throughout the process to make sure she was comfortable and enjoying herself. Additionally, he seems to have been mindful of her boundaries and didn't push her to do anything she wasn't comfortable with.

It's important to note that consent and respect in a sexual encounter are ongoing and require continuous communication and check-ins. So, it's important for Joe to continue being respectful and mindful of Maya's feelings in any future encounters they may have.

Me:
I think she went home quite satisfied with her first experience.

AI:
It seems that way based on the information you've provided. However, it's important to remember that everyone's experiences and perspectives are unique, so it's possible that Maya's feelings and thoughts about the experience could differ from what we know. Ultimately, it's important that all sexual experiences are based on mutual respect, communication, and consent.

Me:
and that Joe is attracted to much younger females

AI:
Yes, the story portrays Joe as having a strong attraction to younger women. This is evident in his interactions with Maya, who is much younger than him, as well as his past affairs with Ella and his interest in taking pictures of high school cheerleaders. It's worth noting that the story does not explicitly condemn or condone this behavior, but rather presents it as a characteristic of Joe's personality.

awnlee jawking ๐Ÿšซ

@LupusDei

I'm looking for noun that expresses in one word "scrawny female person"

Does 'heroin-chic', like recent models on Italian fashion catwalks, count as one word or two?

AJ

awnlee jawking ๐Ÿšซ

@LupusDei

I would like to remind you that body shaming is not acceptable

I'm no historian but I'm unaware of any cultures with a growing tendency towards obesity that were able to reverse it without body shaming, absent an act of God (war, famine etc).

Woke times :-(

AJ

Replies:   Dominions Son
Dominions Son ๐Ÿšซ
Updated:

@awnlee jawking

I'm no historian but I'm unaware of any cultures with a growing tendency towards obesity that were able to reverse it without body shaming,

I doubt you would find any where body shaming worked. In fact I doubt you could find any historical cultures that had a trend towards obesity and then reversed it by any means.

Replies:   Joe Long  awnlee jawking
Joe Long ๐Ÿšซ

@Dominions Son

Famine

Replies:   Dominions Son
Dominions Son ๐Ÿšซ

@Joe Long

Famine

That wouldn't be the culture "doing something" to reverse an obesity trend. That would be an outside force acting on them.

It might be effective, but I don't think doing it deliberately as a cure for obesity would be politically doable in any modern democracy.

Replies:   Joe Long
Joe Long ๐Ÿšซ

@Dominions Son

Right, Stalin's regime wouldn't be considered a modern democracy, but maybe it won't be long until the amount of food that can be purchased on our ration card it tied to our social credit score.

awnlee jawking ๐Ÿšซ

@Dominions Son

I doubt you would find any where body shaming worked.

In some Polynesian cultures, obesity was a sign of success and affluence. Fat-shaming has helped to reverse that and they're now heading towards healthier lifestyles.

AJ

Replies:   Joe Long
Joe Long ๐Ÿšซ

@awnlee jawking

A Korean coworker told me that a belly like mine was a sign of status in Korea. Only the rich could afford enough food to get fat.

Grey Wolf ๐Ÿšซ

@LupusDei

All the chatGPT does, is append its output with a word likely to appear with the given input and output this far.

This is oversimplified; it's much more how 'chat' AIs worked decades ago. Today's generative AIs are closer to a loop where they generate a bunch of words likely to appear, then test them for 'reasonability'. That feeds back into the generator and that cycle repeats a considerable number of times. After all of that, the resulting output is much more complicated than simply a 'what comes next?' sort of prediction (which is why they're so compelling).

Note that the engine used to improve generated output is also often closely related to, or potentially the same as, an engine trained by the initial model training to recognize valid input (which is how it can reject 'bad English' and accept 'good English').

I don't work in AI, and just have a lay understanding, but I do know how classical chat AIs work, which is much more 'what comes next given the input and currently generated output'. That feedback loop was a critical innovation.

LupusDei ๐Ÿšซ

@Joe Long

LOL...

Reuters: Australian mayor readies world's first defamation lawsuit over ChatGPT content

A regional Australian mayor said he may sue OpenAI if it does not correct ChatGPT's false claims that he had served time in prison for bribery,

members of the public told him ChatGPT had falsely named him as a guilty party in a foreign bribery scandal involving a subsidiary of the Reserve Bank of Australia in the early 2000s.

Hood did work for the subsidiary, Note Printing Australia, but was the person who notified authorities about payment of bribes to foreign officials to win currency printing contracts, and was never charged with a crime,

"He's an elected official, his reputation is central to his role," Naughton said. Hood relied on a public record of shining a light on corporate misconduct, "so it makes a difference to him if people in his community are accessing this material"

"It's very difficult for somebody to look behind that to say 'how does the algorithm come up with that answer?'" said Naughton. "It's very opaque."

ystokes ๐Ÿšซ

@Joe Long

Maybe the studio writers should think twice before they strike.

Joe Long ๐Ÿšซ

@Joe Long

Now I'm trying out Starryai for generating photorealistic images that I might be able to use for illustrations. Basically I'm giving a detailed description of real photos in my possession to generate realistic facsimiles. I'm hoping to be able to create a character that I can then recast in different dress, poses and locales.

I've gotten extremely realistic output but it has a hard time following directions. I ask for a girl in a one piece swimsuit walking left to right along a beach, profile to the camera, and I get one in a two-piece suit looking at the camera.

On the other hand, it has given me a couple really hot looking chicks. In the swimsuit I asked for her face to be a blend of Kaitlyn Dever and Kylie Quinn and it did pretty well.

Replies:   Joe Long  awnlee jawking
Joe Long ๐Ÿšซ

@Joe Long

Don't know if this will work but here's a link to the girl at the beach. https://starryai.com/app/my-creations/660349116

One will have to create a free account to view.

Replies:   Joe Long
Joe Long ๐Ÿšซ

@Joe Long

ChatGPT gets very pissy about any suggestion of sex. OTOH Starryai gave me risque photos without my asking. I submitted an actual photo of a girl in a one piece swimsuit and it returned one of a girl with panties but topless, boobs swaying with big dark areola. Another with a girl standing with panties and a t-shirt bunched up under her breasts. I didn't even ask for that.

Replies:   Grey Wolf
Grey Wolf ๐Ÿšซ

@Joe Long

I've played around with Unstable Diffusion (in an earlier form than what's out there now). It's a fair bit more work to set up, and you need a reasonably capable computer, but there are a zillion custom models out there with all sorts of tuning (including some that are tuned away from NSFW).

Replies:   Joe Long
Joe Long ๐Ÿšซ

@Grey Wolf

I like Starryai so far but need to learn more. May go for the $9/mo sub and get more functions.

I have uploaded some real photos as templates with good results, although one ended up giving me crotch shots of 15yo Asian girls (I found I have to specify 'white' or they are heavily Asian or Hispanic.

It has created a few sexy af photos of beautiful imaginary women. What I want to learn is how to save those good creations as characters so that I can use the exact imagery in various poses, settings and levels of being dressed. (One of the uploads was of a topless girl standing next to the drapes, with her nipple covered but lots of side boob. It rendered four rather attractive Asian women a step back from the drapes with various sizes of breasts with clearly visible nipples.)

I also need to get it to follow directions better. I ask for small breasts and it gives me large. Maybe it thinks small is too pedo, but when I asked it to modify one of my favorite creations it gave me the same girl but instead of looking 16 the new ones all looked 12. Doesn't do it for me.

awnlee jawking ๐Ÿšซ

@Joe Long

I recently read that an AI-generated image won an art competition for humans, but the submitter owned up and refused the award.

I think part of the piece's name was 'The Electrician', and it featured two women.

AJ

JoeBobMack ๐Ÿšซ

@Joe Long

I've been delving a bit into topics on this thread, and ran into mentions of starryai.com (image generation) on Reddit, then noticed examples of folks using it to "enhance" their "Replika." So, I did a search on "replika" and got this from Wikipedia: (Notice the last sentence.)

replika.ai
Replika is a chatbot that became publicly available in November 2017. It has millions of users who support its development through subscriptions. Reviewers have commented on the deep emotional intimacy which users have with Replika. Many users have had romantic relationships with Replika, often including erotic talk. In March 2023, Replika developers disabled its romantic and erotic functions.

Replies:   Joe Long
Joe Long ๐Ÿšซ

@JoeBobMack

Well, I'm getting off on able to generate nude images. Some are more photorealistic than other but I'm starting to get some good results.

Had to learn how to tell it to do full frontal nudity without being allowed to say 'naked' or anything mentioning sex. 'Topless' and 'bare bottom' has been working.

https://starryai.com/app/my-creations/394612313
https://starryai.com/app/my-creations/666083662
(Free login with a Google account)

Replies:   StarFleet Carl
StarFleet Carl ๐Ÿšซ

@Joe Long

Got errors that couldn't find creation when I clicked links.

Replies:   Joe Long
Joe Long ๐Ÿšซ
Updated:

@StarFleet Carl

I'm using 'InPrivate' right now so when I clicked on the link it took me to Google login because my credentials aren't retained in this browser method. I entered my gmail address & password and got right in. If you aren't already signed up at starryai you can get a free account with 5 credits per day by using your Google info (I'm using my sock account)

Those weren't the most realistic but showed nudity. Ones in this link are quite realistic and splendid nudes.

https://starryai.com/app/my-creations/901728929
https://starryai.com/app/my-creations/769631853

Replies:   LOAnnie
LOAnnie ๐Ÿšซ

@Joe Long

It's not gonna work cause you're looking at your creations, it would expect us to have our creations there. If you go to the site, they generally show under a user account name to access people's creations that aren't yours.

Replies:   Joe Long
Joe Long ๐Ÿšซ

@LOAnnie

Gotcha.

Thought I'd tested that already, as I have a free account on my home PC and a paid one under a different name on a cloud computer.

Oh well.

Replies:   Joe Long
Joe Long ๐Ÿšซ

@Joe Long

To share you have to publish and to publish it has to be SFW. Despite some prompts such as 'naked' being banned you can generate high quality full frontal.

There's lots of good NSFW but I can share those through the site. However, here's a link to the safe stuff where you can check out the quality of the photorealism. The link goes to a specific set of 4 images, but then click on the user name to see the whole published collected.
https://starryai.com/app/user/obrienlouf/creation/793224708
https://starryai.com/app/user/joelong14

Replies:   StarFleet Carl
StarFleet Carl ๐Ÿšซ

@Joe Long

quality of the photorealism

Well, that takes the level of photos to a WHOLE different level. Damn!

Replies:   Joe Long
Joe Long ๐Ÿšซ

@StarFleet Carl

and I'm still learning, only been doing this a week.

Did a test today. I wanted to 'evolve' an image to get extra detail, so I set it to 100% of the original image to hold as much as possible while adding prompts for extra realism. Well, the pose and background didn't change but the face did as I failed to mention the celebrity template, figuring I didn't need to as I'd used a photo. Then I retried, everything the same but prompting the celebrity template, and that was virtually indistinguishable from the image I submitted.

Lesson learned was I could use that method face swap. That's important because it can be very hard to derive the correct prompts to get the AI to position the characters and get them to do stuff, especially sex. If 'Naked' is in the prompt it will refuse to generate, but I've used 'topless' and 'bare bottom' to get characters fully undressed.

Given the face swap experiment, my next test is to take some screenshots of a few moments in my favorite porn clips, displaying various positions and activity, then submit that to the AI as an original image to be held 100%, along with a prompt for a celebrity template, where hopefully it will place my character's face on Gigi Rivera or Ellie Eilish.

Replies:   StarFleet Carl
StarFleet Carl ๐Ÿšซ

@Joe Long

where hopefully it will place my character's face on Gigi Rivera or Ellie Eilish.

So much for needing to use Photoshop to do that - which has been the previous standard software.

This really is a whole, new level, of deep fakes. This reminds me of the story on here (title escapes me at the moment) where VR allowed you to have 'sex' with anyone.

JoeBobMack ๐Ÿšซ
Updated:

@Joe Long

This Reddit thread about getting ChatGPT-4 to write a science fiction novel is amazing! The author did it as a research project, and not only made the novel but all their notes, logs, and even an Elvenlabs-produced audio book available as public domain.

I personally found that their description of ways to "bound" ChatGPT and to achieve continuity very helpful in thinking about my own process. I don't think I'll ever be that much of a planner, but reading about this effort made me think.

Would love to know what hits others!

Replies:   Joe Long
Joe Long ๐Ÿšซ

@JoeBobMack

Page not found!

Replies:   JoeBobMack
JoeBobMack ๐Ÿšซ

@Joe Long

I think I double entered the url in the link! Try this:

https://www.reddit.com/r/ChatGPT/comments/120oq1x/i_asked_gpt4_to_write_a_book_the_result_echoes_of/

Replies:   Joe Long
Joe Long ๐Ÿšซ

@JoeBobMack

That worked!

I asked it to write a scene and it was really generic and bad. I'll check out your link.

Paige Hawthorne ๐Ÿšซ

@Joe Long

This story was written entirely by ChatGPT, a chatbot launched by OpenAI last year. The instructions I gave it were -- "Write a story about a female private detective in San Francisco." As you can see, the results, while rudimentary, are creditable. Paige

https://storiesonline.net/s/30322/san-francisco-heroine

Back to Top

Close
 

WARNING! ADULT CONTENT...

Storiesonline is for adult entertainment only. By accessing this site you declare that you are of legal age and that you agree with our Terms of Service and Privacy Policy.


Log In