"AI" Rejection

[PandorasBachs]

There's nothing to fix. The hurdle is convincing her it's not AI, not fixing a problem.

The problem is there's literally no information being given out as to what causes her/any software she's using to claim something is written with AI. The reality is absolutely none of the supposed AI detection software is even remotely accurate. I work in ed tech/instructional design at a tier 1 research university and I am literally on a team piloting a few new solutions to identifying the actual writing capabilities of applicants to our graduate programs. There is NOTHING out there that has been able to show above 25% accuracy for AI detection in any moderate to large scale scenario.

If she's using software, then it's her fault that things have broken so badly even for established writers. If she's using people to do this, then they don't know what they're looking for.

 

[OverconfidentSarcasm]

Now let's pretend that someone takes all your stuff and posts it on some random fuckstories site. It's been up for months and you have no idea because it's some little low-traffic site. Then you take your stuff down here for whatever reason and start posting somewhere else. Some fan points out that your stuff has been stolen on this random fuckstories site. You report it.

... I didn't think of that scenario. Good point!

 

[ElectricBlue]

I kinda feel like it says a lot about Lit itself if authors have to maintain accounts here, to make sure Lit isn't making profit off their stolen stories...

That makes no sense. You have an account here so you CAN publish stories, and you're the owner of those stories, not Lit. It's not Lit's job to police what happens to your content. That's down to you, if you can be bothered doing it.

 

[Bramblethorn]

Now let's pretend that someone takes all your stuff and posts it on some random fuckstories site. It's been up for months and you have no idea because it's some little low-traffic site. Then you take your stuff down here for whatever reason and start posting somewhere else. Some fan points out that your stuff has been stolen on this random fuckstories site. You report it.

One little problem. Their posting date is older than yours, because you've removed the stories here with all the original posting dates.

This is a valid argument for keeping stories up, and it's probably useful when dealing with sites that are operating in good faith, but it's not foolproof when dealing with unscrupulous operators. I've seen one site which stole stories from Literotica and reposted them with fabricated dates preceding the original posting. Equally, somebody who'd stolen stories and posted them on another site could allege that the Literotica posting dates are fabricated, and while we know that's unlikely to be true, it might be hard to convince an outsider of that.

Even on a good-faith site, it may be possible for an bad-faith user of that site to manipulate publication dates. Some blogging services etc. allow backdating of posts (there are legitimate reasons for doing so) and it's also possible to create a placeholder post to lock in a publication date, then go back to that years later and replace the content with material stolen from a recent story.

For authors who want better documentation of publication date without paying for registration, one option is going to the Wayback Machine (https://web.archive.org/) and using the "save page now" option to preserve a snapshot of the story on the day it was posted. Those archives have been accepted in some court proceedings as date evidence in IP proceedings, though this may depend on your jurisdiction.

 

[RejectReality]

Frankly, there's nothing we can do about what Laurel decides is AI. Telling us how she determines it would allow anyone who wanted to pass AI off as human-written a pathway to do so. If pitching a bitch was going to accomplish anything, this would have been resolved months ago.

Absent any control over that, useful information about something you can control — such as cultivating another venue for rejected work — at least provides an attainable goal.

 

[ragal2]

Not to worry bawdybloke. These kind folk didn't mean to disparage your great work.

What they meant to say is Lit has a 'heightened alert' screening for AI works. The inclusion of those might get Lit into a legal quagmire over copyright ownership, so it forbids the text generation aspects - though, as noted, you can still use your assisting tools e.g. word processor spell checker and such - Just not something that writes the story as you feed it an idea and it pounds out your plot, actions, etc and you paste and claim it is truly your work with those glitchy things everyone else has tagged above.

As electricblue66 noted, "If it genuinely isn't AI, resubmit it with a note to the Editor and confirm that it's not." Laurel will follow through and get it back on track for our readership.

Lately, I noticed too many stories I sent got rejected or got 'postponed' for too long (some of these were rejected too). I do not understand the issue. I was raised in a country where people didn't speak English. When I sent my first 80-90 stories, many of the comments were about my English (Even though about 60% were still rated HOT). They suggested I use a word-proofer. I started using it, but the ideas, the writing - Everything was mine. Still, rejections or delays continued. These days I use the software VERY sparingly, and my stories keep piling in the pending column( Now I have 5...).
What exactly is the issue, and why do all these new changes affect so significantly the publication of my stories?
BTW, I resubmitted the story and emphasized there is no AI involvement. I still got the standard rejection without any clear explanation. Frustrating, because I won't know what to avoid doing next time!

 

[ElectricBlue]

What exactly is the issue, and why do all these new changes affect so significantly the publication of my stories
BTW, I resubmitted the story and emphasized there is no AI involvement. I still got the standard rejection without any clear explanation. Frustrating, because I won't know what to avoid doing next time!

My guess is the site is being swamped by AI junk, and genuine content is getting sucked into the quagmire, and how does the site tell the difference?

The problem is that the ChatGPT toys (I won't credit them as tools at this point) are new, and I suspect too many people are playing with them. It's the same as the AI drawing tools - they're obviously easy to use (if you're incapable of seeing the deformities and can't count fingers), so every wannabe is having a go.

I'm sure Laurel wants people to write their stories using their own words, but if there are clowns out there generating AI content, they're fucking it up for the rest of us. Ire should be directed towards those people, not the site, and sympathy for those poor bastards trying to get their story published.

 

[ElectricBlue]

So, creating complete stories for LIT etc. isn't happening anyway. So, we basically only talk about translation tools, grammar checkers and integrated thesaurus and dictionaries. And then you have stories written by someone who might use the tools to auto-rewrite parts of it. And that might make them sound boring, but they are still not written by AI. And for boring, we have ratings. And in the end, if Laurel wants to be an additional filter, why doesn't she just say "Rejected for being badly written, look at the help pages for new authors or get help from one of the editors". And not just use "Sounds like AI" as a cheap excuse to not admit that she is censoring content she just doesn't like.

I take your point, but why is now different to way back when? By the logic of that argument, she'd have been knocking back bad writing right from the start, so why start now, all of a sudden?

I wish I knew, both the facts of the matter, and the solution.

 

[StillStunned]

That is exactly what happens when someone uses a translator. And for some reason, Laurel seems to believe that this is what an LLM, or things like deepl or grammerly would do.

But that is what DeepL actually does. If you use DeepL as a translator, that's the result you'll get. If you use a human translator (a good one), you'll get English that reads like English.

And I speak as someone who's edited more DeepL translations than I care to recall.

 

[CalebZhass]

Hey, total newbie when it comes to writing, though I have probably read hundreds of stories on this site alone over the years. Ironically enough, for me it was the AI that gave the push to create stories and it was this thread that makes me question that desire.

I started looking at forum topics on the subject after reading through the AI policy when I was preparing to post a story series I recently created.

See, I am not in this for glory or fame. I simply have stories to tell that I believe others might find interesting / fun and in general I like sharing, especially when it costs me very little over what I would already do. I am however not certain if I am ready to take on the stress of trying to fight a system that even veterans of the site have trouble with.

I am an IT professional and understand LLMs probably way more than most. I definitely can see the moral, legal (and technical) hurdles that come with them and would definitely understand why someone would want to limit AI use that mimics the style of famous artists. Banning AI content just for being written by an AI however is beyond me.

For starters, I have seen humans write worse prose than some AI can and while yes, LLMs have technological limitations that lead to certain writing flaws, those flaws, like repetition or inconsistency could also be present in badly written human content as well.

I believe that AI based content generation is a Pandora's box that we have to learn to live with and embrace it. I say that with the heaviest heart, as I know all too well what sort of problems we are facing down the road as more and more people will start using AI to churn out generic spam content.
BUT: I also know that as with any tool, when used well, AI can greatly improve productivity even in a creative process.

So blame the person who uses the tool for bad, do not condemn the tool and even more importantly, do not condemn the people just for using the tool.

I for example am using the AI to make writing stories more thrilling for me as well. I never bothered writing stories before, as I always knew where they went and since I am not in it for the fame or glory, I never felt it worth putting in the effort to write those thoughts out.

With AI however, my stories have a life of their own. I give the setting, the characters and the direction and watch the story take shape. I course correct when it is derailed, but in general I can enjoy writing the story, as I can never know where exactly I will end up.

You might argue, that at the end it is not me who writes the story then, but the AI. Maybe so, but I beg to differ. The AI might put the words on the screen, but it is ultimately me who decides where the story goes and it is me who writes entire pages worth of lore, content if the AI is stuck and needs a nudge, or if I feel I have something very specific in mind I want to be included.

If it would not be for my fantasy or idea, my characters, my world, my direction and my decisions, the story would not exist. So no. It is not as simple as pressing a button and presto, you have a story. Just the lore alone for the series I am writing consists of more than 3000 words of original writing no one will ever directly see, detailing characters, people, places, etc.. all so that that AI can draw on those descriptions and build around them.

So yea.. I am a bit torn after reading all the desperate / sad posts in this topic and I can only hope that the site finds its way around the subject in a healthy manner.

I will probably still try my luck, just to see how bad things really are, but I do not hold my breath My 150Kb of raw text is only about 50% "original writing", with the other half coming from the AI. Even though I painstakingly post-process my text to eliminate inconsistencies / grammar errors and to flash out bits and pieces I am not satisfied with, it would still leave 30-40% at least that I did not write by hand. Will see I guess.

Anyway, so much for my cheeky first post on the forum. I honestly, truly wish I could have found a better first post to contribute with, and thank you long time authors for all the stories you shared with me over the years.

------ addendum reflecting on some replies I got, which makes me think I have to clarify certain points I might not have been clear about -----
- I respect the site and its operators, but more importantly, I respect the authors who have shared their work here. I will not do anything to disrespect them or the effort they've put into making this site work.
- I will never think less of anyone for how they think about AI generated content, as it is a rather new technology that many are still struggling to grasp and place in their lives. I will however maintain my view and belief, that we will eventually all have to come to terms living with it, as it is not going to just go away.
- I would never try to 'mislead' anyone by misrepresenting my work as I simply have nothing to lie for. If my stories are thrown back because of a technicality and not because of quality, is that really my loss? No. If I share something, I do it because I feel it meets my standards and expectations. More importantly though, I do it because I feel it might meet the standards and expectations of others and be of interest to them.

Yes, I have a series of stories I am in the process of rewriting, that I originally created with an AI. Why am I rewriting them? So they meet the standards I expect from something that I would consider publishing. When I said that "I will probably still try my luck" posting these stories, I did not mean to try and mislead the moderators into thinking its not written using assistance from an AI, but rather that I find the stories to be worth sharing and will try to see if they go through, even though they were admittedly made with the assistance of an AI. I am sorry if that part was not clear enough.

 

[djrip]

Hey, total newbie when it comes to writing, though I have probably read hundreds of stories on this site alone over the years. Ironically enough, for me it was the AI that gave the push to create stories and it was this thread that makes me question that desire.

I started looking at forum topics on the subject after reading through the AI policy when I was preparing to post a story series I recently created.

See, I am not in this for glory or fame. I simply have stories to tell that I believe others might find interesting / fun and in general I like sharing, especially when it costs me very little over what I would already do. I am however not certain if I am ready to take on the stress of trying to fight a system that even veterans of the site have trouble with.

I am an IT professional and understand LLMs probably way more than most. I definitely can see the moral, legal (and technical) hurdles that come with them and would definitely understand why someone would want to limit AI use that mimics the style of famous artists. Banning AI content just for being written by an AI however is beyond me.

For starters, I have seen humans write worse prose than some AI can and while yes, LLMs have technological limitations that lead to certain writing flaws, those flaws, like repetition or inconsistency could also be present in badly written human content as well.

I believe that AI based content generation is a Pandora's box that we have to learn to live with and embrace it. I say that with the heaviest heart, as I know all too well what sort of problems we are facing down the road as more and more people will start using AI to churn out generic spam content.
BUT: I also know that as with any tool, when used well, AI can greatly improve productivity even in a creative process.

So blame the person who uses the tool for bad, do not condemn the tool and even more importantly, do not condemn the people just for using the tool.

I for example am using the AI to make writing stories more thrilling for me as well. I never bothered writing stories before, as I always knew where they went and since I am not in it for the fame or glory, I never felt it worth putting in the effort to write those thoughts out.

With AI however, my stories have a life of their own. I give the setting, the characters and the direction and watch the story take shape. I course correct when it is derailed, but in general I can enjoy writing the story, as I can never know where exactly I will end up.

You might argue, that at the end it is not me who writes the story then, but the AI. Maybe so, but I beg to differ. The AI might put the words on the screen, but it is ultimately me who decides where the story goes and it is me who writes entire pages worth of lore, content if the AI is stuck and needs a nudge, or if I feel I have something very specific in mind I want to be included.

If it would not be for my fantasy or idea, my characters, my world, my direction and my decisions, the story would not exist. So no. It is not as simple as pressing a button and presto, you have a story. Just the lore alone for the series I am writing consists of more than 3000 words of original writing no one will ever directly see, detailing characters, people, places, etc.. all so that that AI can draw on those descriptions and build around them.

So yea.. I am a bit torn after reading all the desperate / sad posts in this topic and I can only hope that the site finds its way around the subject in a healthy manner.

I will probably still try my luck, just to see how bad things really are, but I do not hold my breath My 150Kb of raw text is only about 50% "original writing", with the other half coming from the AI. Even though I painstakingly post-process my text to eliminate inconsistencies / grammar errors and to flash out bits and pieces I am not satisfied with, it would still leave 30-40% at least that I did not write by hand. Will see I guess.

Anyway, so much for my cheeky first post on the forum. I honestly, truly wish I could have found a better first post to contribute with, and thank you long time authors for all the stories you shared with me over the years.

You can have fun writing with AI, but it's not for this site. The major problem we're having here is people being falsely accused, not convincing the site to let AI content through.

 

[CalebZhass]

You can have fun writing with AI, but it's not for this site. The major problem we're having here is people being falsely accused, not convincing the site to let AI content through.

I see your point, but I do believe that the difference you are pointing at is very small. When someone says that AI assisted work has no place somewhere, it implies that the work is not original, which essentially diminishes the author, falsely accusing them of trying to pass someone else's work off as their own.

I do not expect anyone change on my account. I have no real stake in this as an author as... well... let's face it, I am not one at the moment and I might never be one after all. I am perfectly fine doing stuff just to amuse myself, not seeking the validation or appreciation of others.

As a reader however, I would hate to lose out on great stories, just because people are slow to adopt to the emergence of new technologies that shape our lives. If an otherwise great story is killed in its infancy because the writer used an AI to augment its work, that is a tragedy for literature, may those be erotic or more mundane works of fiction.

At least it is for me, the reader losing out on potentially great content.

 

[StillStunned]

As a reader however, I would hate to lose out on great stories, just because people are slow to adopt to the emergence of new technologies that shape our lives. If an otherwise great story is killed in its infancy because the writer used an AI to augment its work, that is a tragedy for literature, may those be erotic or more mundane works of fiction.

At least it is for me, the reader losing out on potentially great content.

I doubt any reader here sees every story written, or even close to every story. Until they start complaining to Laurel about the lack of new stories, I doubt she'll change her policy for "what might have been".

Also, if I'm reading, I prefer a poor story that's well written over an amazing story that's poorly written. I'm sure I'm not the only one.

 

[CalebZhass]

I doubt any reader here sees every story written, or even close to every story. Until they start complaining to Laurel about the lack of new stories, I doubt she'll change her policy for "what might have been".

Also, if I'm reading, I prefer a poor story that's well written over an amazing story that's poorly written. I'm sure I'm not the only one.

It was more of a philosophical statement. You are right of course in terms of what the average reader might or might not consider a problem. Ignorance is bliss. You will never worry about the problem you do not know exists. Just as you will never cry about unpublished stories you never knew existed in the first place. Unfortunately for me, I peeked behind the curtains and will have to come to terms with the reality of how things are.

 

[Bramblethorn]

I will probably still try my luck, just to see how bad things really are, but I do not hold my breath My 150Kb of raw text is only about 50% "original writing", with the other half coming from the AI. Even though I painstakingly post-process my text to eliminate inconsistencies / grammar errors and to flash out bits and pieces I am not satisfied with, it would still leave 30-40% at least that I did not write by hand. Will see I guess.

Literotica's site rules prohibit this kind of content. If Laurel notices or guesses that your stories have this much AI-generated content, they will be rejected. If she becomes aware that you're deliberately trying to publish stories that break the site rules, don't be too surprised if you catch a ban.

Everybody gets to have their own opinion about what the rules for stories here should be, and about whether the site's current methods for enforcing its rules are appropriate. If we don't like the rules we can try to persuade Laurel and Manu to change them, or we can go find some other site whose rules are more to our liking.

But when you're in somebody else's house voluntarily, it's common courtesy to respect their house rules, even if you think those rules are mistaken. If you think the rules are bullshit and you can't live with them, the answer is to leave, not to break the rules behind the owners' backs.

Deliberately trying to get an AI-based story published on a site with a "no AI" policy is extremely poor form, and it's very unlikely to shift the site's policies towards greater tolerance of AI.

 

[Bramblethorn]

I asked an English teacher for feedback to my story. He told me that the dialogue sounds mechanical. Why? Because I write like I learned in school. I also used too many dialogue tags, even at places where they don't add any value because the mood etc. was already known. I also used a lot of complicated, long sentences with a lot of additional information. The dialogue was not natural, because a natural person would omit words and use contractions. Besides some inconsistencies in the way I wrote, for example, times. And that's exactly what someone would do, who learned English as second or third language. That is exactly what happens when someone uses a translator. And for some reason, Laurel seems to believe that this is what an LLM, or things like deepl or grammerly would do.

Do we know that this is in fact what Laurel is using to flag AI-written stories? I've seen plenty of speculation here to that effect - and it's not an unreasonable guess - but AFAICT there has been nothing to confirm it.

 

[Bramblethorn]

https://forum.literotica.com/threads/ai-allegations-thread.1599778/page-5

https://forum.literotica.com/threads/ai-rejection.1597700/page-8#post-97910848

"Bland and toneless" can be interpreted many different ways. It's not obvious to me that this is necessarily about the same specific characteristics you mentioned in your post.

 

[RejectReality]

I think those reported replies are definitive. She's using a program. My reading of them also suggests she's not using the base evaluation, but rather the reports they generate of passages the program believes were generated by AI, and making a determination from there.

They also indicate she's giving it an even closer look when you resubmit without changes stating it's not AI.

If it gets pinged as AI the second time, there's probably not much sense in trying again. Time to cultivate another venue. If you're genuinely not using AI, that sucks, but it's the reality we're currently living in. You can probably blame a mountain of AI-generated crap-flap that's piling up in the queue.

Time will eventually sort it out. The robo-garbage generators will move on or give up when they can't get anything through. As the garbage piles get smaller, things will move closer to the old status quo.

Cold comfort for anyone currently getting the short end of the stick, but sometimes life sucks donkey balls, and there's nothing you can do about it.

 

[CalebZhass]

Literotica's site rules prohibit this kind of content.

Deliberately trying to get an AI-based story published on a site with a "no AI" policy is extremely poor form, and it's very unlikely to shift the site's policies towards greater tolerance of AI.

I guess its on me for not being clear enough in that paragraph, so I edited my original post to clarify.

The portion relevant to you is this:
Yes, I have a series of stories I am in the process of rewriting, that I originally created with an AI. Why am I rewriting them? So they meet the standards I expect from something that I would consider publishing. When I said that "I will probably still try my luck" posting these stories, I did not mean to try and mislead the moderators into thinking its not written using assistance from an AI, but rather that I find the stories to be worth sharing and will try to see if they go through, even though they were admittedly made with the assistance of an AI. I am sorry if that part was not clear enough.

As for cheating/lying/disrespecting the rules being poor form, I do agree wholeheartedly, that is why I would never lie about something as trivial as this, nor did I ever do so. I save my lies for the important stuff

Anyway, if/when (if ever) my work gets rejected, I will just live with it. It will likely not diminish my love for this site and the many wonderfully twisted fantasies people shared here. Life is a though cookie and sometimes we just can't bite hard enough.

 

[Bramblethorn]

I guess its on me for not being clear enough in that paragraph, so I edited my original post to clarify.

Apologies for misunderstanding you, then!

 

[ElectricBlue]

Anyway, so much for my cheeky first post on the forum. I honestly, truly wish I could have found a better first post to contribute with, and thank you long time authors for all the stories you shared with me over the years.

Would it be safe to say your cheeky first post has had a considerable AI assist?

 

[CalebZhass]

Would it be safe to say your cheeky first post has had a considerable AI assist?

I am not quite sure if this is a pun or if you really feel like what I wrote there smells of AI. Curious.
No, I do not use AI to write my forum posts. I tend to use AI to remove tedium from my life or to spice up things with a little bit of randomness and unpredictability. I am perfectly capable of producing mediocre English text in excessive quantities on my own, thank you very much.

Apologies for misunderstanding you, then!

Nothing to apologize for, I was the one not being clear enough, and given the views I voiced on the subject it was a reasonable jump to conclude that I would want to "test the system" in an unethical manner.

 

[ElectricBlue]

I am not quite sure if this is a pun or if you really feel like what I wrote there smells of AI. Curious.
No, I do not use AI to write my forum posts. I tend to use AI to remove tedium from my life or to spice up things with a little bit of randomness and unpredictability. I am perfectly capable of producing mediocre English text in excessive quantities on my own, thank you very much.

To be frank, it reeked of AI, which perfectly illustrates the dilemma (since you say it is not).

As a forum post, it was very, very long (which suggests you've not looked around the forum much, to gauge the TLDR factor - which is something a computer wouldn't do), it's a fairly bland and dry style, meanders to the point where you have one, but the point gets lost because of the verbosity. You need a rigorous editor!

 

[CalebZhass]

To be frank, it reeked of AI, which perfectly illustrates the dilemma (since you say it is not).

As a forum post, it was very, very long (which suggests you've not looked around the forum much, to gauge the TLDR factor - which is something a computer wouldn't do), it's a fairly bland and dry style, meanders to the point where you have one, but the point gets lost because of the verbosity. You need a rigorous editor!

Well, thank you for the feedback. Your analysis is actually spot on with the only new element being that you consider the style dry/bland. I mean, it is a forum post, I did not intend it to be colorful, but I did not expect it to be considered/called bland or dry either.

I would not impose on you, but if you feel like it, I would love to hear your insight on that portion, as that is genuinely new feedback I have not heard before and I am curious as to what I could do to not elicit those feelings.

The rest: length, meandering, overly verbose, those are flaws of mine I am aware of and try to counter as much as I can. Sometimes I succeed, other times.. well.. I could say you saw that, but this is far from me at my worst.

As for it "reeking like AI". The sad part for me is the paranoia this induces in people. Where people second guess and question everything and everyone, looking for boogeymen even when there are none to find.

It is just a phase and it is bound to happen with similar life altering technologies. It is a technological shock, that happens where something is so far out of the norm, that people have a hard time adjusting to it. It will pass. Still, it creates quite a lot of conflicts while it lasts. The best we can do is to try and keep an open mind.

Thanks again for your candid feedback.

 

[XerXesXu]

Well, thank you for the feedback. Your analysis is actually spot on with the only new element being that you consider the style dry/bland. I mean, it is a forum post, I did not intend it to be colorful, but I did not expect it to be considered/called bland or dry either.

I would not impose on you, but if you feel like it, I would love to hear your insight on that portion, as that is genuinely new feedback I have not heard before and I am curious as to what I could do to not elicit those feelings.

The rest: length, meandering, overly verbose, those are flaws of mine I am aware of and try to counter as much as I can. Sometimes I succeed, other times.. well.. I could say you saw that, but this is far from me at my worst.

As for it "reeking like AI". The sad part for me is the paranoia this induces in people. Where people second guess and question everything and everyone, looking for boogeymen even when there are none to find.

It is just a phase and it is bound to happen with similar life altering technologies. It is a technological shock, that happens where something is so far out of the norm, that people have a hard time adjusting to it. It will pass. Still, it creates quite a lot of conflicts while it lasts. The best we can do is to try and keep an open mind.

Thanks again for your candid feedback.

The good news is that over half a dozen AI detectors your OP scored between 0% and 6% AI. You'll have no trouble posting stories on Lit on that account, if you post your own writing.

There's a belief that LLMs are developing a way of writing that's different to humans - an AI style - but that's not the case. LLMs are trained to infer rules, just as the human brain infers rules, we don't know what those rules it infers, but it's trained on a lot of text available on the internet, written by people who strive to follow the, so-called, 'Rules for Writing', a frequent topic of conversation on AH.

The reason I checked you text was because I immediately spotted it as human, you're too free with 'The Rules'.

Some people think that being too free with their rules makes for bad writing and call out writing which strays beyond their permitted range of eccentricity as AI, using the allegation as a term of abuse.

The unfortunates are those who adhere closely to the 'The Rules' as inferred by LLMs and are called as false positives by AI detectors. Those detectors don't say anything about the merits of a particular piece of writing, only that it's too invested in LLMs' rules.