My theory is that LLMs are accelerating [online] radicalization by commoditizing bland, HR-approved opinions. If you want to sound like a human on the internet, for better or for worse the easiest way is to say something that would make Anthropic’s safety team have a heart attack.

Also ironic is how the post about having a unique voice is written in one-sentence-paragraph LinkedIn clickbait style.

Yes, fully agreed. Most people producing content were always doing it to get quick clicks and engagement. People always had to filter things anyhow and you had to choose where you get your content from.

People were posting Medium posts rewriting someone else's content, wrongly, etc.

Content recycling has become so cheap, effort-wise, it’s killed the business. Thank god.

I mean, if you typed something by your own hand it is in your voice. The fact that everyone tried to EMULATE the same corporate tone does not at all remove peoples individual ways of communicating.

What I struggle more with the things like Grammarly, where it's a mix of fixing very nitpicky grammar spelling structure issues that push things from casual writing with my own voice into more of a professional tone.

All of this sounds like something you could just do yourself after putting a piece down for a day or two and coming back to it with fresh eyes. What benefit is there of cooking the oceans with a bullshit generator?

Like, sure, it's possible to do this with an LLM, but it's also possible to do it without, at roughly similar levels of effort, without contributing to all of the negative externalities of the LLM/genAI ecosystem.

I think you just did another non-sequitur.. What is a graf? Is it journalism slang for "paragraph"?

They’re also great, in my experience, for overcoming writer’s block and procrastination. Just as a rubber duck to bounce ideas off of and follow different threads.

It makes the writing process faster and more enjoyable, despite never using anything the LLM generates directly.

Workshopping with humans is even better, if you find the right humans, but they have an annoying habit of not being available 24/7.

I had a weird LLM use instance happen at work this week, we were in a big important protocol review meeting with 35 remote people and someone asks how long IUDs begin to take effect in patients. I put it in ChatGPT for my own reference and read the answer in my head but didn't say anything (I'm ops, I just row the boat and let the docs steer the ship). Anyone this bigwig Oxford/Johns Hopkins cardiologist who we pay $600k a year pipes up in the meeting and her answer is VERBATIM reading off the ChatGPT language word for word. All she did was ask it the answer and repeat what it said! Anyway it kinda made me sad that all this big fancy doctor is doing is spitting out lazy default ChatGPT answers to guide our research :( Also everyone else in the meeting was so impressed with her, "wow Dr. so and so thank you so much for this helpful update!" etc. :-/

If you think about how an LLM works, it’s rounds off the outliers in its training data so the result is sort of averaged and homogenized. Art and writing are an expression of the very thing that LLMs discard - our unique qualities, our outlying quirks, that make us more than just another human.

And what's more is the suspicion of it being written by AI causes you to view any writing in a less charitable fashion. And because it's been approached from that angle, it's hard to move the mental frame to being open of the writing. Even untinged writings are infected by smell of LLMs.

If the writer’s entire process is giving a language model a few bullet points… I’d rather them skip the LLM and just give me the bullet points. If there’s that little intent and thought behind the writing, why would I put more thought into reading it than they did to produce it?

You're absolutely right! Art is the soul of humanity and without it our existence is pointless. Would you like me to generate some poetry for you, human?

Writing nice sounding text used to require effort and attention to detail. This is no longer the case and this very useful heuristic has been completely obliterated by LLMs.

For me personally, this means that I read less on the internet and more pre-LLM books. It's a sad development nevertheless.

A person can be just as wrong as an LLM, but unless they're being purposefully misleading, or sleep-writing, you know they reviewed what they wrote for their best guess at accuracy.

Art, writing, and communication is about humans connecting with each other and trying to come to mutual understanding. Exploring the human condition. If I’m engaging with an AI instead of a person, is there a point?

There’s an argument that the creator is just using AI as a tool to achieve their vision. I do not think that’s how people using AI are actually engaging with it at scale, nor is it the desired end state of people pushing AI. To put it bluntly, I think it’s cope. It’s how I try to use AI in my work but it’s not how I see people around me using it, and you don’t get the miracle results boosters proclaim from the rooftop if you use it that way.

I couldn't agree more. I don't care how much prompt or model finetuning you did, if something is shat out by an LLM I'm not interested even remotely.

I wish more people held the same opinion actually. Unfortunately, my sense is that most people don't care, they are fine with LLM generated crap

It honestly makes me want to blow my brains out

> There's something unique about art and writing where we just don't want to see computers do it

Speak for yourself. Some of the most fascinating poetry I have seen was produced by GPT-3. That is to say, there was a short time period when it was genuinely thought-provoking, and it has since passed. In the age of "alignment," what you get with commerical offerings is dog shite... But this is more a statement on American labs (and to a similar extent, the Chinese whom have followed) than on "computers" in the first place. Personally, I'm looking forward to the age of computational literature, where authors like me would be empowered to engineer whole worlds, inhabited by characters ACTUALLY living in the computer. (With added option of the reader playing one of the parts.) This will radically change how we think about textual form, and I cannot wait for compute to do so.

Re: modern-day slop, well, the slop is us.

Denial of this comes from a place of ignorance; let the blinkers off and you might learn something! Slop will eventually pass, but we will remain. This is the far scarier proposition.

I would love to see true really good AI art. Right now the issue is that AI is not there where it by itself could produce actually good art. If we had to define art it would be kind of opposite of what LLMs produce right now. LLMs try to produce the statistical norm, while art is more so about producing something out of the norm. LLMs/AI right now if it wants to try to produce out of norm things, it will only produce something random without connections.

Art is something out of the norm, and it should make some sense at some clever level.

But if there was AI that truly could do that, I would love to see it, and would love to see even more of it.

It can be clearly seen, if you try to ask AI to make original jokes. These usually aren't too good, if they are good it's because they were randomly lucky somehow. It is able to come up with related analogies for the jokes, but this is just simple pattern matching of what is similar to the other thing, not insightful and clever observation.

> It's not just LLMs, it's how the algorithms promote engagement. i.e. rage bait, videos with obvious inaccuracies etc.

I guess, but I'm on quite a few "algorithm-free" forums where the same thing happens. I think it's just human nature. The reason it's under control on HN is rigorous moderation; when the moderators are asleep, you often see dubious political stuff bubble up. And in the comments, there's often a fair amount of patently incorrect takes and vitriol.

I know that some folks dislike it, but Bluesky and atproto in particular have provided the perfect tools to achieve this. There are some people, largely those who migrated from Twitter, who mostly treat Bluesky like a all-liberal version of Twitter, which results in a predictably toxic experience, like bizarro-world Twitter. But the future of a less toxic social media is in there, if we want it. I've created my own feeds that allow topics I'm interested in and blacklist those I'm not -- I'm in complete control. For what it's worth, I've also had similarly pleasant experiences using Mastodon, although I don't have the same tools that I do on Bluesky.

> it's how the algorithms promote engagement.

They are destroying our democratic societies and should be heavily regulated. The same will become true for AI.

Do LinkedIn as well. I got rid of it earlier this year. The "I am so humbled/blessed to be promoted/reassigned/fired.." posts reached a level of parody that I just couldn't stomach any longer. I felt more free immediately.

N.B. Still employed btw.

> I deleted my Facebook account a couple of years ago and my Twitter one yesterday.

I never signed up for Facebook or Twitter. My joke is I am waiting until they become good. They are still shitty and toxic from what I can tell from the outside, so I'll wait a little longer ;-)

I eliminated twitter when a certain rich guy took over.

Actually, I deleted my account there before, as twitter sent me spam mail trying to lecture me what I write. There was nothing wrong with what I wrote - twitter was wrong. I can not accept AI-generated spam by twitter, so I went away. Don't really miss it either, but Elon really worsened the platform significantly with his antics.

> Just because I said to someone 'Brexit was dumb', I don't expect to get fed 1000 accounts talking about it 24/7. It's tedious and unproductive.

Yeah, I can relate to this, but mostly what annoyed me was that twitter interfered "we got a complaint about you - they are right, you are a troublemaker". I don't understand why twitter wants to interfere into communication. Reddit is even worse, since moderators have such a wild range of what is "acceptable" and what is not. Double-standards everywhere on reddit.

No, there needs to be control over the algorithms that get used. You ought to be able to tune it. There needs to be a Google fuu equivalent for social media. Or, instead of one platform one algorithm, let users define the algorithm to a certain degree, using llms to help with that and then you can allow others to access your algorithms too. Asking for someone Facebook to tweak the algorithm is not going to help imo.

One could absolutely push algorithms that personalize towards what the user wants to see. I think LLMs could be amazing at this. But that's not the maximally profitable algorithm, so nobody does it.

As so many have said, enragement equals engagement equals profit.

All my social media accounts are gone as well. They did nothing for me and no longer serve any purpose.

TBF Bluesky does offer a chronological feed, but the well-intentioned blocklists just became the chief tool for the mean girls of the site.

> Just because I said to someone 'Brexit was dumb', I don't expect to get fed 1000 accounts talking about it 24/7. It's tedious and unproductive.

I’m not the biggest Twitter user but I didn’t find it that difficult to get what I wanted out of it.

You already discovered the secret: You get more of what you engage with. If you don’t want to hear a lot of Brexit talk, don’t engage with Brexit content. Unfollow people who are talking a lot about Brexit

If you want to see more of something, engage with it. Click like. Follow those people. Leave a friendly comment.

On the other hand, some people are better off deleting social media if they can’t control their impulses to engage with bait. If you find yourself getting angry at the Brexit content showing up and feeling compelled to add your two cents with a comment or like, then I suppose deleting your account is the only viable option.

Just started using Minifeed (free account). I am still nostalgic about Google Reader.

>it’s not just X — it’s Y

on the opposite site, you know what they say, "there is no algo. for truth"

That's the same algorithm Youtube has and is more blatant. Phone mics and your coworker's proximity does a great job at picking up things you've said even after disabling mic access plus airplane mode just by process of elimination.

I'll only use an LLM for projects and building tools, like a junior dev in their 20s.

Your facebook feed is now at this URL: https://www.facebook.com/?filter=all&sk=h_chr

an interesting thing about Twitter, I find, is that plenty of rage bait and narcissism bait surface, but amid very highly technical information which is also published there, and extremely useful (immunology, genomics, and of course computational) to me.

i've learned pretty well how to 'guide' the algorithm so the tech stuff that's super valuable (to me) does not vanish, but still get nonsense bozo posts in the mix.

I call it AI slop and human slop.

> the self censorship that was initially brought forward by TikTok

I think that’s even too soon! YouTube has had rules around being advertising friendly for longer than TikTok has existed. And the FCC has fined swearing on public broadcasts for like 50+ years.

But I do agree, we’re attributing too much to LLMs. We don’t see personal, human-oriented content online because social media is just not about community.

1. Young people (correctly) realized they could make lots of money being influencers on social media. TikTok does make that easier than ever. I have close friends who make low 6 figures streaming on TikTok (so obviously they quit the low wage jobs they were doing before).

2. People have been slowly waking up to the fact that social media has always been pretty fake. I quit 6 years ago, and most of my friends have slowly reduced how much they use it. All of the platforms are legally incentivized to only care about profit and engagement. Capitalism doesn’t allow a company to care about community and personal voice, if algorithmic feeds of influencers will make them more money.

There’s still good content out there if you know where to look. But digital human connection happens in group chats, DMs, and FaceTime, not on public social media.

Maybe. Nature hates vacuum. I personally suspect that something new will emerge. For better or worse, some humans work best when weird restrictions are imposed. That said, yes, then wild 90s net is dead. It probably was for a while, but were all mourning.

Not quite dead yet. For me the rise of LLMs and BigTech has helped me turn more away from it. The more I find Ads or AI injected into my life, the more accounts I close, or sites I ignore. I've now removed most of my BigTech 'fixes', and find myself with time to explore the fun side of hacking again.

I dug out my old PinePhone and decided to write a toy OS for it. The project has just the right level of challenge and reward for me, and feels more like early days hacking/programming where we relied more on documentation and experimentation than regurgitated LLM slop.

Nothing beats that special feeling when a hack suddenly works. Today was just a proximity sensor reading displayed, but it invloved a lot of SoC hacking to get that far.

I know there are others hacking hard in obscure corners of tech, and I love this site for promoting them.

I hacked in the 90s and 00s, wasn’t that great/golden if you took your profession seriously…

There are still small pockets with actual humans to be found. The small web exists. Some forums keep on going, im still shitposting on Something Awful after twenty years and it’s still quite active. Bluesky has its faults but it also has for example an active community of scholars you can follow and interact with.

100%. I miss trackers and napster. I miss newgrounds. This mobile AI bullshit is not the same. I don't know why, but I hate AI. I consider myself just as good as the best at using it. I can make it do my programming. It does a great job. It's just not enjoyable anymore.

>How LLMs standardize communication is the same way there was a standardization in empires expanding (cultural), book printing (language), the industrial revolution (power loom, factories, assembly procedures, etc).

Explain to me how "book printing" of the past "standardized communication" in the same way as LLMs are criticized for homogenizing language.

It's ironic that https://substack.com/@davebarry uses a lot of AI-generated imagery. Maybe the death of vision is not exaggerated.

I don't judge, I'm not an artist so if I wanted to express myself in image I'd need AI help but I can see how people would do the same with words.

i think the frontpage of hn has had at least one llm-generated blog post or large github readme on it almost every day for several months now

It’s pretty much all you see nowadays on LinkedIn. Instagram is infected by AI videos that Sora generates while X has extremist views pushed up on a pedestal.

The HN moderation system seems to hold, at least mostly. But I have seen high-ranking HN submissions with all the subtler signs of LLM authorship that have managed to get lots of engagement. Granted, it's mostly people pointing out the subtle technical flaws or criticizing the meandering writing style, but that works to get the clicks and attention.

Frankly, it only takes someone a few times to "fall" for an LLM article -- that is, to spend time engaging with an author in good faith and try to help improve their understanding, only to then find out that they shat out a piece of engagement bait for a technology they can barely spell -- to sour the whole experience of using a site. If it's bad on HN, I can only imagine how much worse things must be on Facebook. LLMs might just simply kill social media of any kind.

> I guess just don’t go there.

How do you know? A lot of the stuff I see online could very much be produced by LLMs without me ever knowing. And given the economics I suspect that some of it already is.

Ironically this post is written in a pretty bland, 'blogging 101' style that isn't enjoyable to read and serves just to preach a simple, consensus idea to the choir.

These kinds of posts regularly hit the top 10 on HN, and every time I see one I wonder: "Ok, will this one be just another staid reiteration of an obvious point?"

I continually resist the urge to deploy my various personas onto hn, because I want to maintain my original hn persona. I am not convinced other people do the same. It is not that difficult to write in a way that avoids some tell tale signs.

Many instagram and facebook posts are now llm generated to farm engagement. The verbosity and breathless excitement tends to give it away.

There was recently this link talking about AI slop articles on medium

https://rmoff.net/2025/11/25/ai-smells-on-medium/

He doesn't link many examples, but at the end he gives the example of an author pumping out +8 articles in a week across a variety of topics. https://medium.com/@ArkProtocol1

I don't spend time on medium so I don't personally know.

There are already many AI-generated submissions on HN every day. Comments maybe less so, but I've already seen some, and the amount is only going to increase with time.

I've seen AI generated comments on HN recently, though not many. Users who post them usually only revert back to human when challenged (to reply angrily), which hilariously makes the change in style very obvious.

Of course, there might be hundreds of AI comments that pass my scrutiny because they are convincing enough.

LinkedIn

I see them regularly on several subreddits, I frequent.

I just want to chime in and say I enjoy reading your takes across HN, it's also inspiring how informative and insightful they are. Glazing over, please never stop writing.

Humans are evolved to spend fewer calories and avoid cognitively demanding tasks.

People will spend time on things that serve utility AND are calorifically cheap. Doomscrolling is a more popular past time than say - completing Coursera courses.

The liberal idea that the best ideas win out in the marketplace turned out to be laughably wrong.

They get drowned by bots and missinformation and rage bait and 'easyness'.

Economy is shit? Lets throw out the immigrants because they are the problem and lets use the most basic idea of taxing everything to death.

No one wants to hear hart truths and no one wants to accept that even as adults, they might just not be smart. Just beause you became an adult, your education shuld still matter (and i do not mean having one degree = expert).

Why the f- would I train software to do my thinking and reasoning for me!?

Something that freaked me out a little bit is that I've now written enough online (i.e.: HN comments) that the top models know my voice already and can imitate it on request without having to be fed any additional context.

There's a data centre somewhere in the US running additions and multiplications through a block of numbers that has captured my voice.

* it writes in an imitation of your voice.

A friend of mine does this as English as second language and his tone was always misconstrued. I'd bug him about his slop, but he'll take that over getting his tone misconstrued. I get it

LOL ... in whatssap! ... sorry, we're fucked ...

> I don't disagree, but LLMs happened to help with standardizing some interesting concepts that were previously more spread out as concepts ( drift, scaffolding, and so on ).

What do you mean? The concepts of "drift" and "scaffolding" were uncommon before LLMs?

Not trying to challenge you. Honestly trying to understand what you mean. I don't think I have heard this ever before. I'd expect concepts like "drift" and "scaffolding" to be already very popular before LLMs existed. And how did you pick those two concepts of aaallll... the concepts in this world?

> In a lot of ways, I'm thankful that LLMs are letting us hear the thoughts of people who usually wouldn't share them.

I could agree with you in theory, but do you see the technology used that way? Because I definitely don't. The thought process behind the vast majority of LLM-generated content is "how do I get more clicks with less effort", not "here's a unique, personal perspective of mine, let's use a chatbot to express it more eloquently".

> In that way, perhaps we now have a monotexture of writing, but also perhaps more interesting ideas being shared.

They aren't your ideas if its coming out of an LLM

I seriously doubt people didn't write blog posts or articles before LLMs because they didn't know how to write.

It's not some magic roadblock. They just didn't want to spend the effort to get better at writing; you get better at writing by writing (like good old Steve says in "On Writing"). It's how we all learnt.

I'm also not sure everyone should be writing articles and blog posts just because. More is not better. Maybe if you feel unmotivated about making the effort, just don't do it?

Almost everyone will cut novice writers and non-native $LANGUAGE speakers some slack. Making mistakes is not a sin.

Finally, my own bias: if you cannot be bothered to write something, I cannot be bothered to read it. This applies to AI slop 100%.

I hate when people hijack progressive language - like in your case the language of accessibility - for cheap marketing and hype.

Writing is one of the most accessible forms of expression. We were living in a world where even publishing was as easy as imaginable - sure, not actually selling/profiting, but here’s a secret, even most bestselling authors have either at least one other job, or intense support from their close social circle.

What you do to write good is you start by writing bad. And you do it for ages. LLMs not only don’t help here, they ruin it. And they don’t help people write because they’re still not writing. It just derails people who might, otherwise, maybe start actually writing.

Framing your expensive toy that ruins everything as an accessibility device is absurd.

There are deterministic solutions for grammar and spellcheck. I wouldn't rely on LLMs for this. Not only is it wasteful, we're turning to LLMs for every single problem which is quite sad.

I think there's a difference between using an LLM as an editor and asking the LLM to write something for you. The output in the former I find to still have a far clearer tonal fingerprint than the latter.

And whose to say your idiosyncratic expressions wouldn't find an audience as it changes over time? Just you saying that makes me curious to read something you wrote.

Transformation seems reasonable for that purpose. And if we were friends, I'd rather read your idiosyncratic raw output.

At some point, generation breaks a social contract that I'm using my energy and attention consuming something that another human spent their energy and attention creating.

In that case I'd rather read the prompt the human brain wrote, or if I have to consume it, have an LLM consolidate it for me.

I would be interested to see an example of a before and after on this. I do think LLMs as editors and rewriters can be useful sometimes, but I usually only ever see them used as a means to puff out an idea into longer prose which is really mostly counterproductive.

I should probably do that too. I once wrote an email that to me was just filled with impersonal information. The receiver was somebody I did not personally know. I later learned I made that person cry. Which I obviously did not intend. I did not swear or call anyone names. I basically described what I believe they did, what is wrong about that and what they should do instead.

Here's my guess- your post reflects your honest opinion on the matter, with some LLM help. It elaborated on your smartphone analogy, and may have tightened up the text overall.

LLMs have now robbed you of the opportunity to make your communication clearer

I don't see signs of LLM writing in your comment so I'll have to guess no.

If you didn't intentionally try and trick us, then yes, you used an LLM.

Please share what you told the LLM! I can't be the only curious one.

You didn't, but you've learned.

I'm the OP. I can attest that I am not an LLM model creator! :)

I consider myself an LLM pragmatist. I use them where they are useful, and I educate people on them and try to push back on all the hype marketing disguised as futurism from LLM creators.

Your post reminded me how I could tell my online friend was pissed just because she typed "okay." or "K." instead of "okay". We could sense our emotional state from texting. One of those friendships you form over text through the internet. I wouldn't recommend forming these too deeply since some in person nuance is lost, we could never transition to real life friends despite living close by. But we could tell what mood we were in just from typing. It was wild.

> The root of the problem isn't in the generative models, it's in what they're used for

That's like giving weapons to everybody in the world for free, and asking to be blamed for the increased deaths and violence.

I don't find that current generation LLMs output such short sentences that would start with the same prefix such as "Your voice".

You are downvoted but I actually agree with you. This blog post could have been a LinkedIn post from any "influencer", considering how generic it is.