I hate how LLMs have contaminated creative writing communities.
Uncategorized
20
Posts
9
Posters
0
Views
-
I hate how LLMs have contaminated creative writing communities. I start reading something that has an interesting premise, and then my heart slowly sinks as I start noticing the LLM telltales sprinkled throughout. It quickly becomes evident that the story is using a lot of words to say nothing, and the writing is terrible.
I don't even mind terrible writing if it's earnest! I've read more cringe fanfic than I care to admit. I'm willing to push through clumsy writing when it's a fellow mind trying their best to show me this neat thing they made.
But LLM terrible writing is completely bland and soulless, just a kilo of store brand words from the words warehouse. You can spot it a mile away, and it devalues anything it touches.
All creative writing communities should have a tool that lets me mark another user as a sloperator, and never see another word they extruded from the words machine. Get that fucking shit out of our idea exchanges.
-
I hate how LLMs have contaminated creative writing communities. I start reading something that has an interesting premise, and then my heart slowly sinks as I start noticing the LLM telltales sprinkled throughout. It quickly becomes evident that the story is using a lot of words to say nothing, and the writing is terrible.
I don't even mind terrible writing if it's earnest! I've read more cringe fanfic than I care to admit. I'm willing to push through clumsy writing when it's a fellow mind trying their best to show me this neat thing they made.
But LLM terrible writing is completely bland and soulless, just a kilo of store brand words from the words warehouse. You can spot it a mile away, and it devalues anything it touches.
All creative writing communities should have a tool that lets me mark another user as a sloperator, and never see another word they extruded from the words machine. Get that fucking shit out of our idea exchanges.
@danderson yes!
i'm a big fan of coding agents (i know you aren't), for boring pragmatic reasons, but WOW do i ever want to never read another line of LLM-written prose again.
not fiction, not technical writing, just...nothing. when so many of my interactions with others are via text, i want _their_ words.
-
I hate how LLMs have contaminated creative writing communities. I start reading something that has an interesting premise, and then my heart slowly sinks as I start noticing the LLM telltales sprinkled throughout. It quickly becomes evident that the story is using a lot of words to say nothing, and the writing is terrible.
I don't even mind terrible writing if it's earnest! I've read more cringe fanfic than I care to admit. I'm willing to push through clumsy writing when it's a fellow mind trying their best to show me this neat thing they made.
But LLM terrible writing is completely bland and soulless, just a kilo of store brand words from the words warehouse. You can spot it a mile away, and it devalues anything it touches.
All creative writing communities should have a tool that lets me mark another user as a sloperator, and never see another word they extruded from the words machine. Get that fucking shit out of our idea exchanges.
And on that note, kudos to SCP, which has a straightforward rule for their creative writing community: any use of generative AI that results in words you didn't write appearing on the site gets you a permanent ban.
And yes, I take very special pleasure in reading new submissions and flagging those with LLM telltales, and seeing the "permanent ban implemented" update in the moderation log a few days later.
-
And on that note, kudos to SCP, which has a straightforward rule for their creative writing community: any use of generative AI that results in words you didn't write appearing on the site gets you a permanent ban.
And yes, I take very special pleasure in reading new submissions and flagging those with LLM telltales, and seeing the "permanent ban implemented" update in the moderation log a few days later.
@danderson SCP deeply understands memetic hazards, after all
-
@danderson SCP deeply understands memetic hazards, after all
@aburka I'm kinda surprised that there hasn't yet been an SCP that's just a direct potshot at LLMs. There probably is and I missed it :)
-
@danderson yes!
i'm a big fan of coding agents (i know you aren't), for boring pragmatic reasons, but WOW do i ever want to never read another line of LLM-written prose again.
not fiction, not technical writing, just...nothing. when so many of my interactions with others are via text, i want _their_ words.
@commaok @danderson it's funny how a lot of people are fine with slop, just not in the media they enjoy most
-
And on that note, kudos to SCP, which has a straightforward rule for their creative writing community: any use of generative AI that results in words you didn't write appearing on the site gets you a permanent ban.
And yes, I take very special pleasure in reading new submissions and flagging those with LLM telltales, and seeing the "permanent ban implemented" update in the moderation log a few days later.
@danderson @aburka I’m delighted they’ve taken such a strong stance. (It’s such a terrific and impressive project in general. I just finished the novel version of There is No Antimemetics Division, and it’s great.)
-
@aburka I'm kinda surprised that there hasn't yet been an SCP that's just a direct potshot at LLMs. There probably is and I missed it :)
@danderson @aburka Are we sure Sam Altman doesn’t need Containment?
-
And on that note, kudos to SCP, which has a straightforward rule for their creative writing community: any use of generative AI that results in words you didn't write appearing on the site gets you a permanent ban.
And yes, I take very special pleasure in reading new submissions and flagging those with LLM telltales, and seeing the "permanent ban implemented" update in the moderation log a few days later.
Oh, and the way to appeal an AI ban is pretty straightforward: show evidence that you did work, for example a draft with versions that show the writing appearing progressively over time, the way it does when a human is writing. As a bonus, if you use the sandbox wiki to draft your things, you get that for free!
I don't think I've ever seen an AI ban appeal succeed, because guess what the slop article appears fully formed in v1, or appears in large multi-paragraph chunks over a few revisions with no rewriting or tinkering (other than removing the more obvious LLM telltales that even sloperators seem to know about, and even that's pretty rare).
-
@danderson @aburka I’m delighted they’ve taken such a strong stance. (It’s such a terrific and impressive project in general. I just finished the novel version of There is No Antimemetics Division, and it’s great.)
@michaelgemar Yeah, the antimemetics stories are a really great little arc!
I can't think of another thematic hub that works as well for me, but plenty of standalone articles are a treat.
-
@michaelgemar Yeah, the antimemetics stories are a really great little arc!
I can't think of another thematic hub that works as well for me, but plenty of standalone articles are a treat.
@danderson It’s unusual in that it accommodates so well bits that are just cool ideas without a lot of attached narrative. I love just dipping into random SCPs.
-
Oh, and the way to appeal an AI ban is pretty straightforward: show evidence that you did work, for example a draft with versions that show the writing appearing progressively over time, the way it does when a human is writing. As a bonus, if you use the sandbox wiki to draft your things, you get that for free!
I don't think I've ever seen an AI ban appeal succeed, because guess what the slop article appears fully formed in v1, or appears in large multi-paragraph chunks over a few revisions with no rewriting or tinkering (other than removing the more obvious LLM telltales that even sloperators seem to know about, and even that's pretty rare).
@danderson I've been reviewing code samples from job applicants. Some of them seem to manually re-type things into the tool. Looks like non-LLM behaviour, but not a clear cut.
-
@danderson @aburka I’m delighted they’ve taken such a strong stance. (It’s such a terrific and impressive project in general. I just finished the novel version of There is No Antimemetics Division, and it’s great.)
@michaelgemar @danderson yeah it's in my TBR heap (I was going to say pile but it's nowhere near that organized)
-
@aburka I'm kinda surprised that there hasn't yet been an SCP that's just a direct potshot at LLMs. There probably is and I missed it :)
@danderson this one requires monitoring of LLM capabilities https://scp-wiki.wikidot.com/scp-8196
-
@danderson this one requires monitoring of LLM capabilities https://scp-wiki.wikidot.com/scp-8196
@danderson and then there's... this https://scp-wiki.wikidot.com/y-ai/offset/1
-
@danderson @aburka Are we sure Sam Altman doesn’t need Containment?
@michaelgemar @danderson pretty sure that would be a net bonus to society, yeah
-
And on that note, kudos to SCP, which has a straightforward rule for their creative writing community: any use of generative AI that results in words you didn't write appearing on the site gets you a permanent ban.
And yes, I take very special pleasure in reading new submissions and flagging those with LLM telltales, and seeing the "permanent ban implemented" update in the moderation log a few days later.
@danderson SCP knows all about containing humanity draining, energy devouring monstrosities
-
@commaok @danderson it's funny how a lot of people are fine with slop, just not in the media they enjoy most
@aburka @commaok @danderson More than hallucinations or producing disinformation, the biggest danger from AI is it causes people to say "This is brilliant!" about mediocre slop they don't have the expertise to identify.
-
@aburka @commaok @danderson More than hallucinations or producing disinformation, the biggest danger from AI is it causes people to say "This is brilliant!" about mediocre slop they don't have the expertise to identify.
@AlSweigart @aburka @commaok @danderson If only people read and believed human experts with the same trust they put into LLMs - is it because if the answer comes from my computer it must be right?
-
And on that note, kudos to SCP, which has a straightforward rule for their creative writing community: any use of generative AI that results in words you didn't write appearing on the site gets you a permanent ban.
And yes, I take very special pleasure in reading new submissions and flagging those with LLM telltales, and seeing the "permanent ban implemented" update in the moderation log a few days later.
@danderson I wasn’t aware of SCP, so thanks! What a rabbit hole!
-
undefined oblomov@sociale.network shared this topic on
Gli ultimi otto messaggi ricevuti dalla Federazione
-
In 2022 I wrote about designing for where the internet isn’t. Now there’s an open-source space station OS built on ROS 2 doing exactly that - life support that can’t wait for a round-trip to Earth.
https://www.sealjay.com/blog/the-future-of-edge-demassification/
-
-
Juslin Bodo Bodo Bodo
-
Tired: The AI industry is bad because copyright infringement is bad
Wired: The AI industry is bad AND THEREFORE copyright infringement is awesome
https://www.garbageday.email/p/the-only-taboo-left-is-copyright-infringement
-
@repubblica @cronache-italiane-la-repubblica-repubblica porcaccia eva... ma in Italia a sabotare le ferrovie di Salvini allora sono sessanta milioni!
-
@brozu vigiliamo che non broglino...e vadano a votare anche coloro che si asterrebbero ma già salvarono Ilaria Salis!
-
Perché dovreste aver paura a dirlo?
Per chiunque si senta schifato del fatto che “le #Santanché” restino impunite:
sappiate che con il SÌ quella impunità diventa strutturale.
E stranamente proprio la destra estrema combatte questo referendum con tanta aggressività.
Quel sogno di casta le appartiene per natura.
-
@b0rk my least favorite man pages were back when all the GNU tools man pages were just stubs telling you to use info. User hostility taken to a level only GNU could manage.
