[HN Gopher] LLMs can get "brain rot"
       ___________________________________________________________________
        
       LLMs can get "brain rot"
        
       Author : tamnd
       Score  : 253 points
       Date   : 2025-10-21 14:24 UTC (8 hours ago)
        
 (HTM) web link (llm-brain-rot.github.io)
 (TXT) w3m dump (llm-brain-rot.github.io)
        
       | AznHisoka wrote:
       | Can someone explain this in laymen terms?
        
         | PaulHoule wrote:
         | They benchmark two different feeds of dangerous tweets:
         | (1) a feed of the most popular tweets based on likes, retweets,
         | and such       (2) an algorithmic feed that looks for clickbait
         | in the text
         | 
         | and blend these in different proportions to a feed of random
         | tweets that are not popular nor clickbait and find that feed
         | (1) has more of damaging effect on the performance of chatbots.
         | That is, they feed that blend of tweets into the model and then
         | they ask the models to do things and get worse outcomes.
        
           | ForHackernews wrote:
           | Blended in how? To the training set?
        
             | PaulHoule wrote:
             | Very early training.
        
         | sailingparrot wrote:
         | train on bad data, get a bad model
        
           | xpe wrote:
           | > train on bad data, get a bad model
           | 
           | Right: in the context of supervised learning, this statement
           | is a good starting point. After all, how can one build a good
           | supervised model if you can't train it on good examples?
           | 
           | But even in that context, it isn't an incisive framing of the
           | problem. Lots of supervised models are resilient to some
           | kinds of error. A better question, I think, is: _what kinds_
           | of errors at _what prevalence_ tend to degrade performance
           | and _why_?
           | 
           | Speaking of LLMs and their ingestion processing, there is a
           | lot more going on than purely supervised learning, so it
           | seems reasonable to me that researchers would want to try to
           | tease the problem apart.
        
         | rriley wrote:
         | The study introduces the "LLM Brain Rot Hypothesis," asserting
         | that large language models (LLMs) experience cognitive decline
         | when continuously exposed to low-quality, engaging content,
         | such as sensationalized social media posts. This decline,
         | evident in diminished reasoning, long-context understanding,
         | and ethical norms, highlights the critical need for careful
         | data curation and quality control in LLM training. The findings
         | suggest that standard mitigation strategies are insufficient,
         | urging stakeholders to implement routine cognitive health
         | assessments to maintain LLM effectiveness over time.
         | 
         | TL;DR from
         | https://unrav.io/#view/8f20da5f8205c54b5802c2b623702569
        
       | pixelmelt wrote:
       | Isn't this just garbage in garbage out with an attention grabbing
       | title?
        
         | philipallstar wrote:
         | Attention is all you need.
        
           | echelon wrote:
           | In today's hyper saturated world, attention is everything:
           | 
           | - consumer marketing
           | 
           | - politics
           | 
           | - venture fundraising
           | 
           | When any system has a few power law winners, it makes sense
           | to grab attention.
           | 
           | Look at Trump and Musk and now Altman. They figured it out.
           | 
           | MrBeast...
           | 
           | Attention, even if negative, wedges you into the system and
           | everyone's awareness. Your mousey quiet competitors aren't
           | even seen or acknowledged. The attention grabbers suck all
           | the oxygen out of the room and win.
           | 
           | If you go back and look at any victory, was it really better
           | solutions, or was it the fact that better solutions led to
           | more attention?
           | 
           | "Look here" -> build consensus and ignore naysayers -> keep
           | building -> feedback loop -> win
           | 
           | It might not just be a societal algorithm. It might be one of
           | the universe's fundamental greedy optimization algorithms. It
           | might underpin lots of systems, including how we ourselves as
           | individuals think and learn.
           | 
           | Our pain receptors. Our own intellectual interests and
           | hobbies. Children learning on the playground. Ant colonies.
           | Bee swarms. The world is full of signals, and there are
           | mechanisms which focus us on the right stimuli.
        
             | peterlk wrote:
             | You're absolutely right!
        
             | ghurtado wrote:
             | Something flew approximately 10 miles above your head that
             | would be a good idea for you to learn.
        
               | scubbo wrote:
               | There were plenty of kinder ways to let someone know that
               | they had missed a reference - https://xkcd.com/1053/
        
               | echelon wrote:
               | What makes you think I didn't know the reference? That
               | paper is seminal and essential reading in this space.
               | 
               | The intent was for you to read my comment at face value.
               | I have a point tangential to the discussion at hand that
               | is additive.
        
             | lawlessone wrote:
             | Is this copypasted from LinkedIn?
        
               | echelon wrote:
               | If you traverse back the fourteen years of my comment
               | history (on this account - my other account is older),
               | you'll find that I've always written prose in this form.
               | 
               | LLMs trained on me (and the Hacker News corpus), not the
               | other way around.
        
             | alganet wrote:
             | You're not accounting for substrate saturation.
             | 
             | If you could just spam annoy until you win, we'd be all
             | dancing to remixed versions of Macarena.
        
           | dormento wrote:
           | In case anyone missed the reference:
           | https://arxiv.org/abs/1706.03762
           | 
           | > (...) We propose a new simple network architecture, the
           | Transformer, based solely on attention mechanisms, dispensing
           | with recurrence and convolutions entirely.
        
         | wat10000 wrote:
         | Considering that the current state of the art for LLM training
         | is to feed it massive amounts of garbage (with some good stuff
         | alongside), it seems important to point this out even if it
         | might seem obvious.
        
           | CaptainOfCoit wrote:
           | I don't think anyone is throwing raw datasets into LLMs and
           | hoping for high quality weights anymore. Nowadays most of the
           | datasets are filtered one way or another, and some of them
           | highly curated even.
        
             | BoredPositron wrote:
             | I doubt they are highly created you would need experts in
             | every field to do so. Which gives me more performance
             | anxiety for LLMs because one of the most curated fields
             | should be code...
        
               | nradov wrote:
               | OpenAI has been literally hiring human experts in certain
               | targeted subject areas to write custom proprietary
               | training content.
        
               | BoredPositron wrote:
               | I bet the dataset is mostly comprised of certain
               | areas(tm).
        
               | groby_b wrote:
               | The major labs are hiring experts. They carefully build &
               | curate synthetic data. The market for labelled non-
               | synthetic data is currently ~$3B/year.
               | 
               | The idea that LLMs are just trained on a pile of raw
               | Internet is severely outdated. (Not sure it was ever
               | fully true, but it's far away from that by now).
               | 
               | Coding's one of the easier datasets to curate, because we
               | have a number of ways to actually (somewhat) assess code
               | quality. (Does it work? Does it come with a set of tests
               | and pass it? Does it have stylistic integrity? How many
               | issues get flagged by various analysis tools? Etc, etc)
        
               | satellite2 wrote:
               | Is that right? Isn't the current way of doing thing to
               | throw "everything" at it then fine tune?
        
         | Barrin92 wrote:
         | Yes, I am concerned about the Computer Science profession
         | 
         | >""Brain Rot" for LLMs isn't just a catchy metaphor--it
         | reframes data curation as cognitive hygiene for AI"
         | 
         | A metaphor is exactly what it is because not only do LLMs not
         | possess human cognition, there's certainly no established
         | science of thinking they're literally valid subjects for
         | clinical psychological assessment.
         | 
         | How does this stuff get published, this is basically a blog
         | post. One of the worse aspects of the whole AI craze is that is
         | has turned a non-trivial amount of academia into a complete
         | cargo cult joke.
        
           | bpt3 wrote:
           | It is a blog post, it was published as a Github page and on
           | arXiv.
           | 
           | I think it's intended as a catchy warning to people who are
           | dumping every piece of the internet (and synthetic data based
           | on it!) that there are repercussions.
        
             | pluc wrote:
             | I think it's an interesting line of thought. So we all
             | adopt LLMs and use it everywhere we can. What happens to
             | the next generation of humans, born with AI and with
             | diminished cognitive capacity to even wonder about
             | anything? What about the next generation? What happens to
             | the next generation of AI models that can't train on
             | original human-created datasets free of AI?
        
               | iwontberude wrote:
               | They will accept that their orders come from a terminal
               | and they will follow them.
        
               | fragmede wrote:
               | Manna. https://marshallbrain.com/manna1
        
             | gowld wrote:
             | arXiv is intended to host research papers, not a blog for
             | researchers.
             | 
             | Letting researchers pollute it with blog-gunk is an abuse
             | of the referral/vetting system for submitters.
        
         | otterley wrote:
         | And with extra steps!
        
           | Insanity wrote:
           | Garbage in -> Magic -> Hallucinated Garbage out
        
         | icyfox wrote:
         | Yes - garbage in / garbage out still holds true for most things
         | when it comes to LLM training.
         | 
         | The two bits about this paper that I think are worth calling
         | out specifically:
         | 
         | - A reasonable amount of post-training can't save you when your
         | pretraining comes from a bad pipeline; ie. even if the
         | syntactics of the input pretrained data are legitimate it has
         | learned some bad implicit behavior (thought skipping)
         | 
         | - Trying to classify "bad data" is itself a nontrivial problem.
         | Here the heuristic approach of engagement actually proved more
         | reliable than an LLM classification of the content
        
           | satellite2 wrote:
           | Yes but the other interesting bit which is not clearly
           | addressed is that increasing the garbage in to 100% does not
           | result in absolute garbage out. So visibly there is still
           | something to learn there.
        
         | ashleyn wrote:
         | Yes, but the idea of chatgpt slowly devolving into Skibidi
         | Toilet and "6 7" references conjures a rather amusing image.
        
           | 1121redblackgo wrote:
           | 6-7 9(**)_
        
             | stavros wrote:
             | Can someone explain this? I watched a South park episode
             | that was all about this, but I'm not in the US so I have no
             | idea what the reference is.
        
               | Sparkle-san wrote:
               | It's a meme without a lot of real meaning behind it.
               | While it has its origins, I wouldn't say it's a
               | "reference" to anything specific.
               | 
               | https://en.wikipedia.org/wiki/6-7_(meme)
        
               | stavros wrote:
               | Ahh, thanks, so it's just a thing kids say.
        
               | 1121redblackgo wrote:
               | Yep
        
               | lexandstuff wrote:
               | It's a line from a banger Skrilla song, nothing more than
               | that.
        
       | CaptainOfCoit wrote:
       | > continual exposure to junk web text induces lasting cognitive
       | decline in large language models (LLMs).
       | 
       | TLDR: If your data set is junk, your trained model/weights will
       | probably be junk too.
        
       | b0gb wrote:
       | AIs need supervision, just like regular people... /s
        
       | thelastgallon wrote:
       | If most of the content produced by younger generations is about
       | skibidi toilet[1] and 67[2], isn't that what LLMs are going to be
       | trained on?
       | 
       | [1] https://en.wikipedia.org/wiki/Skibidi_Toilet
       | 
       | [2] https://en.wikipedia.org/wiki/6-7_(meme)
        
         | micromacrofoot wrote:
         | only if the trends last long enough (which they rarely do!),
         | skibidi is already old news according to some kids I know
        
           | ciaranmca wrote:
           | Agreed, " Popularity as a better indicator". Hypothetically
           | you could look at popularity over time to filter out viral
           | rot content and work out if people feel the content is
           | useful.
        
       | Isamu wrote:
       | Another analogy to help us understand that LLMs are a useful part
       | of what people do but are wildly misconstrued as the whole story
        
       | bbstats wrote:
       | making a model worse is very easy.
        
       | moffkalast wrote:
       | Ah yes, something the local LLM fine tuning community figured out
       | how to do in creative ways as soon as llama 1 released. I'm glad
       | it has a name.
        
       | killshotroxs wrote:
       | If only I got money every time my LLM kept looping answers and
       | telling stuff I didn't even need. Just recently, I was stuck with
       | LLM answers, all while it wouldn't even detect simple syntax
       | errors...
        
       | rriley wrote:
       | This paper makes me wonder the long lasting effects of the
       | current media consumption patterns by the alpha-gen kids.
        
         | AznHisoka wrote:
         | why just kids?
        
           | rriley wrote:
           | I am mostly concerned with the irreversibility part. More
           | developed brains probably would not be affected as much.
        
             | jama211 wrote:
             | Have you opened facebook recently? Seems the older folk are
             | plenty affected to me.
        
               | FactolSarin wrote:
               | But don't worry, us middle aged people are definitely
               | immune.
        
               | rriley wrote:
               | Good point :-)
        
             | vanderZwan wrote:
             | I recently saw an article about the history of Sesame
             | Street that claimed that in the late 1960s American
             | preschool kids watched around _twenty-seven hours of
             | television per week on average_ [0]. And most of that was
             | not age-appropriate (education TV had yet to be invented).
             | So maybe we should check in on the boomers too if we're
             | sincere about these worries.
             | 
             | [0] https://books.google.se/books?id=KOUCAAAAMBAJ&pg=PA48&v
             | q=ses...
        
               | ordu wrote:
               | It is an interesting hypothesis. Seriously. There is a
               | trend in Homo Sapience cultural evolution to treat
               | children in more and more special ways from generation to
               | generation. The (often implicit) idea it helps children
               | to develop faster and to leverage their sensitive and
               | critical periods of development, blah-blah-blah... But
               | while I can point to some research on importance of
               | sensitive and critical periods of development, I can't
               | remember any research on the question if a deficit of
               | age-inappropriate stimuli can be detrimental for
               | development.
               | 
               | There were psychologists who talked about zone of
               | proximal development[0], about importance of exposing a
               | learner to tasks that they cannot do without a support.
               | But I can't remember nothing about going further and
               | exposing a learner to tasks far above their heads when
               | they cannot understand a word.
               | 
               | There is a legend about Sofya Kovalevskaya[1], who became
               | a noteworthy mathematician after she were exposed to
               | lecture notes by Ostrogradsky when she was 11 yo. The
               | walls of her room were papered with those notes and she
               | was curious what are all that symbols. It doesn't mean
               | that there is a causal link between these two events, but
               | what if there is one?
               | 
               | What about watching deep analytical TV show at 9 yo? How
               | it affect the brain development? I think no one tried to
               | research that. My gut feeling that it can be
               | motivational, I didn't understand computers when I met
               | them first, but I was really intrigued by them. I learned
               | BASIC and it was like magic incantations. It had build a
               | strong motivation to study CS deeper. But the question is
               | are there any other effects beyond motivation? I remember
               | looking at the C-program in some book and wondering what
               | does it all mean. I could understand nothing, but still I
               | had spent some time trying to decipher the program.
               | Probably I had other experiences like that, which I do
               | not remember now. Can we say with certainty that it had
               | no influence on my development and hadn't make things
               | easier for me later?
               | 
               |  _> So maybe we should check in on the boomers too if we
               | 're sincere about these worries._
               | 
               | Probably we should be sincere.
               | 
               | [0] https://en.wikipedia.org/wiki/Zone_of_proximal_develo
               | pment
               | 
               | [1] https://en.wikipedia.org/wiki/Sofya_Kovalevskaya
        
       | avazhi wrote:
       | "Studying "Brain Rot" for LLMs isn't just a catchy metaphor--it
       | reframes data curation as cognitive hygiene for AI, guiding how
       | we source, filter, and maintain training corpora so deployed
       | systems stay sharp, reliable, and aligned over time."
       | 
       | An LLM-written line if I've ever seen one. Looks like the authors
       | have their own brainrot to contend with.
        
         | askafriend wrote:
         | If it conveys the intended information then what's wrong with
         | that? You're fighting a tsunami here. People are going to use
         | LLMs to help their writing now and forever.
        
           | binary132 wrote:
           | The brainrot apologists have arrived
        
             | askafriend wrote:
             | Why shouldn't the author use LLMs to assist their writing?
             | 
             | The issue is how tools are used, not that they are used at
             | all.
        
               | grey-area wrote:
               | Because they produce text like this.
        
               | xanderlewis wrote:
               | Is it really so painful to just think for yourself? For
               | one sentence?
               | 
               | The answer to your question is that it rids the writer of
               | their unique voice and replaces it with disingenuous
               | slop.
               | 
               | Also, it's not a 'tool' if it does the entire job. A
               | spellchecker is a tool; a pencil is a tool. A machine
               | that _writes for you_ (which is what happened here) is
               | not a tool. It 's a substitute.
               | 
               | There seem to be many falling for the fallacy of 'it's
               | here to stay so you can't be unhappy about its use'.
        
               | SkyBelow wrote:
               | Assist without replacing.
               | 
               | If you were to pass your writing it and have it provide a
               | criticism for you, pointing out places you should
               | consider changes, and even providing some examples of
               | those changes that you can selectively choose to include
               | when they keep the intended tone and implications, then I
               | don't see the issue.
               | 
               | When you have it rewrite the entire writing and you past
               | that for someone else to use, then it becomes an issue.
               | Potentially, as I think the context matter. The more a
               | writing is meant to be from you, the more of an issue I
               | see. Having an AI write or rewrite a birthday greeting or
               | get well wishes seems worse than having it write up your
               | weekly TPS report. As a simple metric, I judge based on
               | how bad I would feel if what I'm writing was being
               | summarized by another AI or automatically fed into a
               | similar system.
               | 
               | In a text post like this, where I expect others are
               | reading my own words, I wouldn't use an AI to rewrite
               | what I'm posting.
               | 
               | As you say, it is in how the tool is used. Is it used to
               | assist your thoughts and improve your thinking, or to
               | replace them? That isn't really a binary classification,
               | but more a continuum, and the more it gets to the
               | negative half, the more you will see others taking issue
               | with it.
        
               | dwaltrip wrote:
               | The paragraph in question is a very poor use of the tool.
        
           | avazhi wrote:
           | If you can't understand the irony inherent in getting an LLM
           | to write about LLM brainrot, itself an analog for human
           | brainrot that arises by the habitual non use of the human
           | brain, then I'm not sure what to tell you.
           | 
           | Whether it's a tsunami and whether most people will do it has
           | no relevance to my expectation that researchers of LLMs and
           | brainrot shouldn't outsource their own thinking and
           | creativity to an LLM in a paper that itself implies that
           | using LLMs causes brainrot.
        
             | nemonemo wrote:
             | What you are obsessing with is about the writer's style,
             | not its substance. How sure are you if they outsourced the
             | thinking to LLMs? Do you assume LLMs produce junk-level
             | contents, which contributes human brain rot? What if their
             | contents are of higher quality like the game of Go?
             | Wouldn't you rather study their writing?
        
               | avazhi wrote:
               | Writing is thinking, so they necessarily outsourced their
               | thinking to an LLM. As far as the quality of the writing
               | goes, that's a separate question, but we are nowhere
               | close to LLMs being better, more creative, and more
               | interesting writers than even just decent human writers.
               | But if we were, it wouldn't change the perversion
               | inherent in using an LLM here.
        
           | moritzwarhier wrote:
           | What information is conveyed by this sentence?
           | 
           | Seems like none to me.
        
           | uludag wrote:
           | Nothing wrong with using LLMs--until every paragraph sounds
           | like it's A/B tested for LinkedIn virality. That's the rot
           | setting in.
           | 
           | The problem isn't using AI--it's sounding like AI trying to
           | impress a marketing department. That's when you know the
           | loop's closed.
        
             | drusepth wrote:
             | Brilliantly phrased -- sharp, concise, and perfectly
             | captures that uncanny "AI-polished" cadence everyone
             | recognizes but can't quite name. The tone strikes just the
             | right balance between wit and warning.
        
               | solarkraft wrote:
               | You are absolutely right!
        
               | ewoodrich wrote:
               | Lately the Claude-ism that drives me even more insane is
               | "Perfect!".
               | 
               | Particularly when it's in response to pointing out a big
               | screw up that needs correcting and CC utterly unfazed
               | just merrily continues on like I praised it.
               | 
               | "You have fundamentally misunderstood the problems with
               | the layout, before attempting another fix, think deeply
               | and re-read the example text in the PLAN.md line by line
               | and compare with each line in the generated output to
               | identify the out of order items in the list."
               | 
               | "Perfect!...."
        
               | glenstein wrote:
               | One thing I don't understand, there was (appropriately) a
               | news cycle about sycophancy in responses. Which was real,
               | and happening to an excessive degree. It was claimed to
               | be nerfed, but it seems strong as ever in GPT5, and it
               | ignores my custom instructions to pare it back.
        
               | anjel wrote:
               | "Any Compliments about my queries cause me anguish and
               | other potent negative emotions."
        
           | stavros wrote:
           | The problem is that writing isn't only judged on whether it
           | conveys the intended information or not. It's also judged on
           | whether it does that well, plus other aesthetic criteria.
           | There is such a thing as "good writing", distinct from "it
           | mentioned all the things it needed to mention".
        
           | grey-area wrote:
           | It's a text generator regurgitating plausible phrases without
           | understanding and producing stale and meaningless pablum. It
           | doesn't even know what the intended information is, and
           | judging from the above neither did the human involved.
           | 
           | It doesn't help writing it stultifies and gives everything
           | the same boring cheery yet slightly confused tone of voice.
        
             | zer00eyz wrote:
             | > It's a text generator regurgitating plausible phrases
             | without understanding and producing stale and meaningless
             | pablum.
             | 
             | Are you describing LLM's or social media users?
             | 
             | Dont conflate how the content was created with its quality.
             | The "You must be at least this smart (tall) to publish
             | (ride)" sign got torn down years ago. Speakers corner is
             | now an (inter)national stage and it written so it must be
             | true...
        
               | grey-area wrote:
               | I really could only be talking about LLMs but social
               | media is also low quality.
               | 
               | The quality (or lack of it) if such texts is self
               | evident. If you are unable to discern that I can't help
               | you.
        
           | AlecSchueler wrote:
           | Style is important in writing. It always has been.
        
           | sailingparrot wrote:
           | > If it conveys the intended information then what's wrong
           | with that?
           | 
           | Well, the issue is precisely that it doesn't convey any
           | information.
           | 
           | What is conveyed by that sentence, exactly ? What does
           | reframing data curation as cognitive hygiene for AI entails
           | and what information is in there?
           | 
           | There are precisely 0 bit of information in that paragraph.
           | We all know training on bad data lead to a bad model,
           | thinking about it as "coginitive hygiene for AI" does not
           | lead to any insight.
           | 
           | LLMs aren't going to discover interesting new information for
           | you, they are just going to write empty plausible sounding
           | words. Maybe it will be different in a few years. They can be
           | useful to help you polish what you want to say or otherwise
           | format interesting information (provided you ask it to not be
           | ultra verbose), but its just not going to create information
           | out of thin air if you don't provide it to it.
           | 
           | At least, if you do it yourself, you are forced to realize
           | that you in fact have no new information to share, and do not
           | waste your and your audience time by publishing a paper like
           | this.
        
           | Angostura wrote:
           | it's not really clear whether it conveys an "intended
           | meaning" because it's not clear whether the meaning -
           | whatever it is - is really something the authors intended.
        
           | dwaltrip wrote:
           | Because it sounds like shit? Taste matters, especially in the
           | age of generative AI.
           | 
           | And it doesn't convey information that well, to be honest.
        
         | standardly wrote:
         | That is indeed an LLM-written sentence -- not only does it
         | employ an em dash, but also lists objects in a series -- twice
         | within the same sentence -- typical LLM behavior that renders
         | its output conspicuous, obvious, and readily apparent to HN
         | readers.
        
           | kcatskcolbdi wrote:
           | thanks, I hate it.
        
           | Jackson__ wrote:
           | LLM slop is not just bad--it's degrading our natural
           | language.
        
           | itsnowandnever wrote:
           | why do they always say "not only" or "it isn't just x but
           | also y and z"? I hated that disingenuous verbosity BEFORE
           | these LLMs out and now it'll all over the place. I saw a post
           | on linked in that was literally just like 10+ statements of
           | "X isn't just Y, it's etc..." and thought I was having a
           | stroke
        
             | Starlevel004 wrote:
             | GPT loves lists and that's a variant of a list
        
               | wizzwizz4 wrote:
               | Lists have a simpler grammatical structure than most
               | parts of a sentence. Semantic similarity makes them easy
               | to generate, even if you pad the grammar with filler.
               | And, thanks to Western rhetoric, they nearly _always_
               | come in threes: this makes them easy to predict!
        
             | moritzwarhier wrote:
             | It's not just a shift of writing style. It symbolizes the
             | dangerous entrapment of a feedback loop that feeds the
             | worst parts of human culture back into itself.
             | 
             | scnr
        
             | heavyset_go wrote:
             | They're turns of phrase I see a lot in opinion articles and
             | the like. The purpose is to take a popular framing and
             | reframe it along the lines of the author's own ideas.
             | 
             | LLMs fundamentally don't get the human reasons behind its
             | use, see it a lot because it's effective writing, and
             | regurgitate it robotically.
        
           | turtletontine wrote:
           | I think this article has already made the rounds here, but I
           | still think about it. I love using em dashes! It really makes
           | me sad that I need to avoid them now to sound human
           | 
           | https://bassi.li/articles/i-miss-using-em-dashes
        
             | janderson215 wrote:
             | The em dash usage conundrum is likely temporary. If I were
             | you, I'd continue using them however you previously used
             | them and someday soon, you'll be ignored the same way
             | everybody else is once AI mimics innumerable punctuation
             | and grammatical patterns.
        
               | astrange wrote:
               | They didn't always em-dash. I expect it's intentional as
               | a watermark.
               | 
               | Other buzzwords you can spot are "wild" and "vibes".
        
             | jader201 wrote:
             | Same here. I recently learned it was an LLM thing, and I've
             | been using them forever.
             | 
             | Also relevant:
             | https://news.ycombinator.com/item?id=45226150
        
           | veber-alex wrote:
           | hehe, I see what you did there.
        
           | AlecSchueler wrote:
           | Don't forget the "it's not just X, it's Y" formulation and
           | the rule of 3.
        
           | hunter-gatherer wrote:
           | Lol. This is brilliant. I'm not sure if anyone else has this
           | happen to them, but I noticed in college my writing style and
           | "voice" woukd shift quite noticeably depending on whatever I
           | was reading heavily. I wonder if I'll start writing more like
           | an LLM naturally as I unavoidably read more LLM-generated
           | content.
        
             | actionfromafar wrote:
             | Yes. It's already shifting spoken language.
        
             | MarcelOlsz wrote:
             | I've always read AI messages in this voice/style [0]
             | 
             | [0] https://www.youtube.com/watch?v=KiqkclCJsZs.
        
           | b33j0r wrote:
           | I talked like that before this happened, and now I just feel
           | like my diction has been maligned :p
           | 
           | I think it's because I was a pretty sheltered kid who got A's
           | in AP english. The style we're calling "obviously AI" is most
           | like William Faulkner and other turn-of-the-20th-century
           | writing, that bloggers and texters stopped using.
        
             | dingnuts wrote:
             | IDK all the breathless "it's not just X, it's Y --" reminds
             | me of press releases
        
               | b33j0r wrote:
               | Yeah it was trained on bullshit more than Faulkner for
               | sure. +1 you.
        
         | mtillman wrote:
         | I recently saw someone on HN comment about LLMs using
         | "training" in quotes but no quotes for thinking or reasoning.
         | 
         | Making my (totally rad fwiw) Fiero look like a Ferrari does not
         | make it a Ferrari.
        
           | snickerbockers wrote:
           | I like to call it tuning, it's more accurate to the way they
           | "learn" by adjusting coefficients and also there's no proven
           | similarity between any existing AI and human cognition.
           | 
           | Sometimes I wonder if any second order control system would
           | qualify as "AI" under the extremely vague definition of the
           | term.
        
         | mvdtnz wrote:
         | What is actually up with the "it's not just X, it's Y" cliche
         | from LLMs? Supposedly these things are trained on all of the
         | text on the internet yet this is not a phrasing I read pretty
         | much anywhere, ever, outside of LLM content. Where are they
         | getting this from?
        
       | gaogao wrote:
       | Brain rot texts seems reasonably harmful, but brain rot videos
       | are often surreal and semantically dense in a way that probably
       | improves performance (such as discussed on this German brain rot
       | analysis https://www.youtube.com/watch?v=-mJENuEN_rs&t=37s). For
       | example, Svankmajer is basically proto-brainrot, but is also the
       | sort of thing you'd watch in a museum and think about.
       | 
       | Basically, I think the brain rot aspect might be a bit of
       | terminology distraction here, when it seems what they're
       | measuring is whether it's a puff piece or dense.
        
         | f_devd wrote:
         | I do not think this is the case, there has been some research
         | into brainrot videos for children[0], and it doesn't seem to
         | trend positively. I would argue anything 'constructed' enough
         | will not classify as far on the brainrot spectrum.
         | 
         | [0]: https://www.forbes.com/sites/traversmark/2024/05/17/why-
         | kids...
        
           | gaogao wrote:
           | Yeah, I don't think surrealism or constructed is good in the
           | early data mix, but as part of mid or post-training seems
           | generally reasonable. But also, this is one of those cases
           | where anthropomorphizing the model probably doesn't work,
           | since a major negative effect of Cocomelon is kids only
           | wanting to watch Cocomelon, while for large model training,
           | it doesn't have much choice in the training data
           | distribution.
        
         | moritzwarhier wrote:
         | For this reason, I believe thar the current surge we see in AI
         | use for people manipulation (art is also a form of
         | manipulation, even if unintended) is much more important than
         | their hyped usage as a technical information processors.
         | 
         | Brainrot created by LLMs is important to worry about, their
         | design as "people pleasers".
         | 
         | Their anthropomorphization can be scary too, no doubt.
        
       | conception wrote:
       | This is a potential moat for the big early players in a pre-
       | atomic steal sort of way as any future players won't have a non-
       | AI-slop/dead internet to train new models on.
        
       | andai wrote:
       | I encourage everyone with even a slight interest in the subject
       | to download a random sample of Common Crawl (the chunks are
       | ~100MB) and see for yourself what is being used for training
       | data.
       | 
       | https://data.commoncrawl.org/crawl-data/CC-MAIN-2025-38/segm...
       | 
       | I spotted here a large number of things that it would be unwise
       | to repeat here. But I assume the data cleaning process removes
       | such content before pretraining? ;)
       | 
       | Although I have to wonder. I played with some of the base/text
       | Llama models, and got very disturbing output from them. So
       | there's not _that_ much cleaning going on.
        
         | throwaway314155 wrote:
         | > But I assume the data cleaning process removes such content
         | before pretraining? ;)
         | 
         | I didn't check what you're referring to but yes, the major
         | providers likely have state of the art classifiers for
         | censoring and filtering such content.
         | 
         | And when that doesn't work, they can RLHF the behavior from
         | occurring.
         | 
         | You're trying to make some claim about garbage in/garbage out,
         | but if there's even a tiny moat - it's in the filtering of
         | these datasets and the purchasing of licenses to use other
         | larger sources of data that (unlike Common Crawl) _aren't_
         | freely available for competition and open source movements to
         | use.
        
         | dist-epoch wrote:
         | Karpathy made a point recently that the random Common Crawl
         | sample is complete junk, and that something like an WSJ article
         | is extremely rare in it, and it's a miracle the models can
         | learn anything at all.
        
           | andai wrote:
           | >Turns out that LLMs learn a lot better and faster from
           | educational content as well. This is partly because the
           | average Common Crawl article (internet pages) is not of very
           | high value and distracts the training, packing in too much
           | irrelevant information.
           | 
           | >The average webpage on the internet is so random and
           | terrible it's not even clear how prior LLMs learn anything at
           | all. You'd think it's random articles but it's not, it's
           | weird data dumps, ad spam and SEO, terabytes of stock ticker
           | updates, etc. And then there are diamonds mixed in there, the
           | challenge is pick them out.
           | 
           | https://x.com/karpathy/status/1797313173449764933
           | 
           | Context: FineWeb-Edu, which used Llama 70B to [train a
           | classifier to] filter FineWeb for quality, rejecting >90% of
           | pages.
           | 
           | https://huggingface.co/spaces/HuggingFaceFW/blogpost-
           | fineweb...
        
       | commandlinefan wrote:
       | My son just sent me an instagram reel that explained how cats
       | work internally, but it was a joke, showing the "purr center" and
       | "knocking things off tables" organ. It was presented completely
       | seriously in a way that any human would realize was just supposed
       | to be funny. My first thought was that some LLM is training on
       | this video right now.
        
         | Night_Thastus wrote:
         | I'm reminded of this 'repair' video:
         | https://www.youtube.com/watch?v=3e6motL4QMc
        
       | chuckreynolds wrote:
       | is that why chatGPT always tells me "6 7 lol"? ;)
        
       | jdkee wrote:
       | " Studying "Brain Rot" for LLMs isn't just a catchy metaphor--it
       | reframes data curation as cognitive hygiene for AI, guiding how
       | we source, filter, and maintain training corpora so deployed
       | systems stay sharp, reliable, and aligned over time."
       | 
       | Is this slop?
        
         | Profan wrote:
         | ... it sure reads like slop
         | 
         | and you know what they say, if it walks like slop, quacks like
         | slop and talks like slop, it's probably slop
        
       | earth2mars wrote:
       | duh! isn't that obvious. is this some students wanted a project
       | with pretty graphs on writing experience?! I am not trying to be
       | cynical or anything. just questioning the obvious thing here.
        
       | antegamisou wrote:
       | My Goodness, looks like Computer 'Science' is a complete
       | euphemism now.
        
         | guelo wrote:
         | It's turning into a social science.
        
       | nakamoto_damacy wrote:
       | Our metaphorical / analogical muscle is too well developed. Maybe
       | there is a drug we can take to reduce how much we lean into it.
       | 
       | If you look at two random patterns of characters and both contain
       | 6s you could say they are similar (because you're ignoring that
       | the similarity is less than 0.01%). That's how comparing LLMs to
       | brains feels like. Like roller skates to a cruise ship. They both
       | let you get around.
        
       | buyucu wrote:
       | I don't understand why people have a hard time understanding
       | 'garbage in, garbage out'. If you train your model on junk, then
       | you will have a junk model.
        
       | nomel wrote:
       | "Brain rot" is just the new term for "slang that old people don't
       | understand".
       | 
       | "Cool" and "for real" are no different than "rizz" and "no cap".
       | You spoke "brain rot" once, and "cringed" when your parents
       | didn't understand. The cycle repeats.
        
         | kcatskcolbdi wrote:
         | This both has nothing to do with the linked article (beyond the
         | use of brain rot in the title, but I'm certain you must have
         | read the thing you're commenting on, surely) and is simply
         | incorrect.
         | 
         | Brain rot in this context is not a reference to slang.
        
       | buellerbueller wrote:
       | By all means, let's make sure the LLMs have healthier media diets
       | than the humans. We wouldn't want the humans to realize they are
       | being dumbed down into cattle. /s
        
       | themafia wrote:
       | > as cognitive hygiene
       | 
       | LLMs are not cognizant. It's a terrible metaphor. It hides the
       | source of the issue. The providers cheaped out on sourcing their
       | data and now their LLMs are filled with false garbage and
       | copyrighted material.
        
         | donaldihunter wrote:
         | Likewise, cognitive decline isn't what's happening here since
         | that would require cognition. At best it is a simulation of
         | cognitive decline.
        
       | numpad0 wrote:
       | Not surprising that trending tweets as data is junk, not only
       | from brainrots-be-brainrots perspective: trending tweets are
       | contextual. They don't make sense without the rest of the
       | timeline.
       | 
       | And now I know why bots on Twitter don't even work, even with
       | humans in it - they're shooting blind.
        
       ___________________________________________________________________
       (page generated 2025-10-21 23:00 UTC)