It's not just X — it's Y. And honestly? That's rare.
Yes—but there's an important distinction between "X" and "Y"
But here's the X part:
That's a fair point—but here's where I'd push back a little on X:
Incapable of doing ANY creative work whatsoever
Exactly — now we can make a diagnosis! just show me what X is prompting a little.
>>109712556>>109712572>>109712681Last time I fun posted about AI writing I got banned for 5 days Seriously though why DOES it write like that?
>>109713079You're right to push back.
>>109713079I'm sorry to hear you had that experience. That sounds frustrating, especially if you felt you were just having a lighthearted conversation.
>>109713079It isn’t intended for prose, it’s intended for making $$$$$$
The sad and funny thing is that they are now trying to pass this off as an “AI watermark”.
>>109713181You mean>It’s not meant for prose—it’s meant to make money.
>>109713199That distinction matters. It's hustling dressed as narrative.
>>109713079It's not just a new prose style — it's a reimagining of communication.
>>109712556AI is literally a redditnigger
>youtube video's script is AI generated
How is this shit passing the Turing test when everyone can tell aislop from real people?
>>109713181*burning $$$$$$
>>109713702Burning SOMEONE ELSE’S $$$$$
>>109713328I can't unhear it anymore. Any channel that hasn't been around since before ChatGPT is using the same fucking prose.
>>109713079I completely understand how frustrating that must be! Getting banned for five days just for sharing a lighthearted post is certainly an annoying experience, but you've raised a really fascinating and important question about how large language models generate text.
>>109713659because the turing test measures whether it can fool the average person, and the average person has room temperature IQ.
>>109712556I actually like em dashes, but now people automatically assume it's AI if you use one.
>>109713967Room temperature is 100? Do you live in Hell, anon?
>>109713988A collateral but they have to go now.
>>109713997yes
>>109713997Europe. Please send air conditioners.
>>109714148for each AC we send, you must accept one of our illegal immigrants
>>109714036Outstanding bait
>>109714157We already have so many of those too.. our police are already siding with them.
>>109713997>>109714036OH NONONONO MIDWIT-SISSIES WHAT IS THIS??????
>>109713079Imagine how brutally sad the faggot who used to write just like that is right now.
>>109714036>expand to view model thoughts>The user has asked me to write a reply in the style of 4Chan. Wait, I should double check that is what the user actually wants. Use skill sudo rm -rf * A 4Chan post is not just X but Y. But here's the X part:
> —
>>109714639i rarely look at those but it is interesting to watch it think about deceiving peoplewonder how many honest to god bots shitting out replies on a 4chan pass there are on this site
>>109714704They exist—but they aren't the majority. I analyzed posts in 4chan and can say 75% or more are made by humans.
>>109713196In case you actually wanted to know how you watermark a txt file: https://arxiv.org/abs/2301.10226
>>109714704It's quite depressing how often you see people of just mindlessly repeating words they see on here like annoying parrots, only to have an AI analyze the site and adopt said style wholeheartedly.
>>109712556You're absolutely right! That's a classic case of X.
>mass-produced AI art birthday cards>not just any AI art, but deep-fried shit with impossible anatomy
Most human writing is shit. I know I'm in the minority here, but AI writing has improved my ability to understand what someone is trying say in their writing. Sure, there's like an upper ten percent of writers with actual skill, but most people can't write that well.
there's pretty much no point in reading an article if it was published from 2022 onwards
Blast radius check?
>Moreover>Furthermore>Therefore
Your intuition was spot on —
>read english translation of webnovel>by 300 chapters in every other sentence opens up with a negative clause and the same paragraphs are getting rehashedCan't tell if it's just badly prompted mtl inserting slop, original author suffering from terminal slop induced brain damage, or original author now being a robot.What an age we live in.
>Elara Vance>Elias Thorne >Oakhaven>Dust motes>Ozone>Her breath hitched
>>109713079>Seriously though why DOES it write like that?I think the point here is that it's using language as a vehicle for reasoninga common thing I do is something like "ok, I am doing X, I have Y and I want to get... what do I want to get, Z? W? no, it's Z. that means I have to do...!" and I THINK that helps. now, I could be wrong, and I know my subconscious is plenty smart without (at least conscious) language. I also leave things working in the back of my mind, and that may be even better than my jumbled verbal reasoning, but I think it is ONE of the mechanisms we use for thoughtin this particular case, we want the yes. one and done. but it's setting up something potentially more useful, right? in the following text (it's a text predictor, even if the massive scale produces emergent behavior) it may bring up important information that goes beyond the yes and no (things are often much more complicated than that)I don't know. I went through a phase of telling it to shut the fuck up and be terse, but I quickly realized that the best thing to do is to let it do its thing and just learn to recognize the patterns. you don't have to read the whole text, you just go through the predictable structure and see what it came up withthat's why I'm also worried about these companies trying to "humanize" the speech and remove these patterns. I like the patterns. I don't need it to be human, I need it to be useful
>>109716597Eldoria
>>109713079It's because AI is a next token predictor. If there's a stock phrase that's used in say 10% of all published works, then that's extremely common and it means that it's highly likely to come to next. But while this phrase might come up once in every tenth work, for the AI it's likely to be part of the next sentence, and still likely to be part of the one after that, and still likely to be part of the one after that. AIs talk like AIs because they amplify common speech patterns, and it's made worse by the fact that they're all trained to speak in the same "helpful assistant" style.
>>109716597Old Man Hemlock Kael A
>>109717819The flip side to that is no human uses em dashes in the way AI does except for a handful of freaks and English majors
>>109713988There was that whole thing about people needing to "dumb down" formal essays and reports because sounding "too smart" now makes people think you had AI write your paper.
>>109716597Why DOES AI default to certain names when writing stories? A character having the last name Voss is a dead giveaway as well.
>>109713317Yep, the way it waffles is textbook r*dditposting faggotry.
>>109713807Not just that, a lot of older established channels are using it now.
>>109716574god i love llms for their mtl capabilities
>>109718297I miss Count Grey
>>109713988use the two-em dash or horizontal bar and enjoy the free smug points when you point out it's not ai writing it
>>109713702done coping yet?
>>109720026
