This weekly roundup thread is intended for all culture war posts. 'Culture war' is vaguely defined, but it basically means controversial issues that fall along set tribal lines. Arguments over culture war issues generate a lot of heat and little light, and few deeply entrenched people ever change their minds. This thread is for voicing opinions and analyzing the state of the discussion while trying to optimize for light over heat.
Optimistically, we think that engaging with people you disagree with is worth your time, and so is being nice! Pessimistically, there are many dynamics that can lead discussions on Culture War topics to become unproductive. There's a human tendency to divide along tribal lines, praising your ingroup and vilifying your outgroup - and if you think you find it easy to criticize your ingroup, then it may be that your outgroup is not who you think it is. Extremists with opposing positions can feed off each other, highlighting each other's worst points to justify their own angry rhetoric, which becomes in turn a new example of bad behavior for the other side to highlight.
We would like to avoid these negative dynamics. Accordingly, we ask that you do not use this thread for waging the Culture War. Examples of waging the Culture War:
-
Shaming.
-
Attempting to 'build consensus' or enforce ideological conformity.
-
Making sweeping generalizations to vilify a group you dislike.
-
Recruiting for a cause.
-
Posting links that could be summarized as 'Boo outgroup!' Basically, if your content is 'Can you believe what Those People did this week?' then you should either refrain from posting, or do some very patient work to contextualize and/or steel-man the relevant viewpoint.
In general, you should argue to understand, not to win. This thread is not territory to be claimed by one group or another; indeed, the aim is to have many different viewpoints represented here. Thus, we also ask that you follow some guidelines:
-
Speak plainly. Avoid sarcasm and mockery. When disagreeing with someone, state your objections explicitly.
-
Be as precise and charitable as you can. Don't paraphrase unflatteringly.
-
Don't imply that someone said something they did not say, even if you think it follows from what they said.
-
Write like everyone is reading and you want them to be included in the discussion.
On an ad hoc basis, the mods will try to compile a list of the best posts/comments from the previous week, posted in Quality Contribution threads and archived at /r/TheThread. You may nominate a comment for this list by clicking on 'report' at the bottom of the post and typing 'Actually a quality contribution' as the report reason.
Jump in the discussion.
No email address required.
Notes -
Begrudging congratulations to @self_made_human for producing a book review entry that's good enough to make it as a finalist of Scott's Book Review Contest, yet AI-sloppy enough to make the comments section focus almost exclusively on the (un)acceptability of undisclosed use of AI.
This reminds me of a video I stumbled on the other day. It's a 12-minute "documentary-style" video essay telling the story of a rescue dog that gets dropped off at a shelter without any documentation. It's a cute, healthy dog that should be eminently adoptable except for his utter refusal to engage with anyone. Five months in, the dog's time at the shelter is up, and he is scheduled for transfer to doggy prison (or whatever, the point is that it's the end of trying to get him adopted). But a woman brings in her small, shy six-year-old deaf girl, and amazingly the dog responds to the child alone, and knows all the hand-sign commands. Turns out the dog was raised by a deaf old man, who then had a stroke. That's the gist, though the video packs a whole lot more dramatic beats, twists, and turns.
It's a real tear-jerker... if it describes a true story. Unless it's made up. Which, come to think of it (I say to myself as I review my reaction to the video), it probably is. Because how can there possibly be a dog smart enough to learn sign commands, yet so dumb that it couldn't possibly figure out any other forms of human-dog communication. Yeah, I smell bullshit.
But OK, maybe it's like a video short fiction, it's not supposed to be real, no more than any other kind of fiction. Then I can judge the video on this fictional-short-film kind of merit, and not worry about thinking critically on the veracity of facts. Damn it, they really should have specified that this is fiction. Then I can enjoy it as such. It has good plot beats. The actors are cute and relatable. Especially the dog.
Unless it's AI-generated. Posted March 2026. Possible. Fuck.
When I read something that, presumably, is the human author's thoughts, only for it to turn out to be LLM's output, I feel a similar betrayal of trust as when I read something purportedly factual that turns out to be fictional, even if not actively meant to deceive. This is definitely true if the entire output is LLM-produced, and remains partially true if LLM is used for editing that inserts LLM-created phrases. Like this zinger from the review:
I hope this forum preserves the norm of humans writing for humans by humans, even as I know that the overall tide is flowing the other way.
There is no other form of writing.
I hope I will live to see this psychosis around text generation not being genuine somehow be ridiculed like the disdain for photography came to be ridiculed. And I further hope that the same will happen to the ridiculous anthropomorphization of these tools.
What people dislike about AI sounding texts is that they are trite, not that they are AI generated. But we somehow have to pretend that most human written text isn't shit and that the painstaking process of communication imbues it with further value, all out of some misplaced luddite syndicalism.
Writing is communication, and at its highest, art. All of which comes from intent and cognition in the mind that originates it, and none of this is untrue of generated text, much like none of this is untrue of photography.
As this place exists to pursue debate in political ideas, and not as an exercise in keyboard assisted calligraphy, I believe the norm in question will come to be seen as not only antiquated, but impossible to enforce.
I'm comfortable with a sophisticated-enough model trained on my knowledge and tone to be accessible to my work teams as a utilitarian force multiplier. That day seems a long way off, because even with reams of data to consume that hasn't gotten any better in years.
On a forum built around human discussion, or the human-focused discussions that are important for fostering culture, I reject that AI usage is correct or necessary.
More options
Context Copy link
Most of the human-written text is indeed shit, because most human thought is half-formed nonsense. The process of getting your thoughts out of your head (writing, speaking, or otherwise) is the first step of actually forming the multi-dimensional incoherent impressions and feelings into something coherent. I know that's how it is for me, and I have witnessed enough college students struggling (when pressed) to get a coherent thought out to suspect that the experience is close to universal. There are the blessed few who can erudite off the cuff, and even fewer of those say anything of substance.
And yes, the process is painstaking. One gets better at the skill of forming and communicating a coherent idea, but that skill requires one to keep putting the thoughts that feel so wholesome and brilliant in one's mind down on paper and recognizing their inanity and pettiness, over and over. Painstaking, and humbling. But valuable. When I take the time to set my half-baked ideas down, I can see where they lead, from outside myself. And as often as not, I end up disagreeing with what I myself have put down, and changing my mind, or adjusting the ideas to something far more nuanced. It makes me more open to further changing my mind on the matter. Hegel's dialectic in action.
So I recognize the value of forming my own ideas in my own halting way. By extension, I support this effort in others.
Well put. Writing changes our ideas and forces us to bring thought, intuition and word into alignment. All three can change in the process. If you don't go through that usually painful process I am sceptical you are getting an ai to accurately express your thoughts. I think it's much more likely you are allowing the ai to determine what you think. Unwritten thoughts are just too inchoate for us to accurately say of an ai response "oh yes, that's exactly how I would have put it".
More options
Context Copy link
More options
Context Copy link
There’s an abstract idea in the back of my head about “pure” expression, unobstructed by (but requiring some) constraints.
Right now, if you run text through an LLM, it will make it sound less like you and more like every other LLM. I hope we find ways to use AI and other developments to make people more express their pure, unique selves.
Like the invention of new forms of art, back to the first cave paintings, or like the Internet did with personal sites, webcomics, and flash games. Even those constrain the artist to the medium and trends, but I think they are net improvements.
People say this but i think it's completely untrue. You can make the result sound vastly different if you're willing to put some effort into properly using it.
It again sounds to me like criticism of photography "they all look the same", yeah, if you haven't invented or considered a composition grammar or metering and exposure.
What's maybe true is that it'll always sound slightly different than handwritten ways in ways that may or may not be important. But rejecting it altogether as a tool is a stronger claim than that.
I know, survivorship bias, but I’ve never seen a convincing example of LLM-written text that didn’t read at least slightly but noticeably LLM-written, and definitely certain human writing styles I’ve never seen replicated by an LLM. The Unslop Contest tried to produce good AI fiction, and every result has similar LLM prose except the schizophrenic one.
I can think of one but to be fair it's not longform compared to those.
That’s an old post, I think GPT-3. Older LLMs did sound more original, although they hallucinated more. Interestingly, increasing an LLM’s temperature hyperparameter (which I’m almost certain produced the schizophrenic Unslop entry) also makes it sound different and unique, but output nonsense.
Probably because they’re more random, but also (at least GPT-3 does) rely less on fine tuning and more on its training from the Internet.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Just show me the prompt then. It has the same information density as the output.
LLMs are commonly used to summarize larger texts. This isn't true.
A generated summary of a legal text is more information dense and intelligible by laymen than the sum of its inputs.
Sure, if you ask it to summarize it will summarize and what you say is true. However, LLMs (in general) output such verbosity that it is just certainly true that they are not adding useful information. Obviously if you ask it for a cake recipe it will add information. But, if you're trying to discuss some topic on TheMotte, the output will just be a combination of whatever information is in the input, plus some kind of related informational average from the internet. It's this latter component, I imagine, we recognize as "slop."
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
I don't think this is true. I think people dislike the auto-generation itself. The simple test of this would be that you can show somebody a text, ask if they enjoy it, get a result, and then reveal to them that it was bot-written and watch them get angry.
I'm willing to bet that if you ran a study where you gave the same text to three groups, where you tell the first group nothing, tell the second group that the text is written by a bot, and tell the third group that the text is written by a human, and then ask them all to rate the text, the second group would rate it the lowest.
Maybe that's irrational, in the same way that people will prefer hand-made or artisanal objects to machined objects, but if so it's still a highly predictable irrationality.
More options
Context Copy link
I think it's the other way around. People grow to dislike text that "sounds like AI" because they don't like AI. Em-dashes weren't a problem until AI started using them; "It's not X, it's Y" wasn't a problem until AI started using it; and so on.
Sometimes people dislike AI for inherently ideological reasons, and maybe the tide will shift on that. But I think more frequently, AI becomes synonymous with "a thing that you didn't put as much effort into" because that is factually what it is, and knowing how much effort someone put into something is socially important. (For many people, their biggest exposure to AI output comes from lazy good-for-nothing coworkers/bosses/scammers, and they hate those people!)
Photographs didn't become popular by becoming indistinguishable from paintings. They became their own thing, and people have a decent understanding of how much effort goes into them! If you submitted an inkjet print of a photo to a painting competition, people would still get mad at you.
There are cases where AI that is clearly AI has gotten some mainstream acceptance; people enjoy DougDoug streams and, to some extent, weird demos that obviously couldn't be done another way. But in areas where it's competing directly against art forms that require more effort, people just look for other ways to prove they put in the effort. (Many artists have started recording time-lapses of themselves painting to show that they actually did it; many writers deliberately avoid clichés that AI uses.)
Beyond that people associate AI output with lazy people I think they dislike them because they happen too much and it's indicative of "lazy thinking" on part of both the human and the AI. The argument devolves into a particular style rather than the style being chosen because it supports what the text is supposed to do or communicate.
Is kind of like how some people always tie things back to their hobby horse, except in style form but with the same sort of tortured reasoning and insight loss.
Hopefully this is a transitory problem.
More options
Context Copy link
That's perceptive.
But the problem is that this also applies to the printing press. Who actually gives a shit about copyist monk stolen valor when we could make the whole of Europe read?
Type made on a printing press continues to be seen as much less valuable than handwritten calligraphy; that's why some people still hand-write their wedding invitations.
Because "valuable" and "expensive" are synonyms, the printing press brought books into reach of a lot more people. But all the forms of art that AI copies were already broadly accessible to everyone. You could already get a near-infinite amount of drawings and videos and TheMotte comments for free online, that's where the AI companies got their training data from.
It is often argued that AI will allow more people to create those things. From the consumer perspective, that just means it's harder for me to find the people who put a lot of actual effort into them, which is what I was mostly looking for in the first place (especially in a social venue like TheMotte), so it's net-negative.
I just find it funny that people reinvent Guénon's tirades against the reign of quantity the second that industry comes for them.
Your love of handwritten text is the same as my love of handmade objects. Quaint but ultimately secondary to the pursuit of ends.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
No. I dislike that it is generated by an LLM. Ultimately it comes down to: I care what humans have to say, not clankers. If you can't be assed to actually write something, then I have no desire to read it.
You claim what you really care about is some commitment that the writer has to the output. You want people to suffer enough that what they have to say may have some value.
This logic contradicts your conclusions. Even granting text generation is free of work (it is not), you would have to be interested in generated texts by authors that have proved their commitment through some other method.
For instance, you should be more interested in a terrorists' generated manifesto than in human written ads.
So either it is untrue that this is the reason of your disdain, or it is untrue that what you dislike about generated text is that it is generated.
No, I just want commitment to the actual writing. If I have to suffer reading it, they had better have suffered writing it. I don't care if they have suffered doing some other thing, and in fact in the specific terrorist example that would count against them.
Why is suffering using a specific tool a different signal than suffering using another specific tool?
The human has picked the words. Yes, yes, I know, I arbitrarily don't value things like "the human has calligraphied the words in a specific handwriting style and ink", etc etc. I value what I value.
And say what you will about photography, at least it actually captures what had existed at some point, not does something like "an imp inside the machine had been told what the picture looked like in order for it to reproduce it".
I'm afraid not.
You know what I meant.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Despite this uncharitable framing, yeschad.jpg, seeing as being human necessarily entails suffering. If this is the main distinction you want to draw between humans and clankers, I'll raise my brow but shrug and nod along so long as it leads you to the same conclusions, i.e that humans have a certain connection to their outputs in a way clankers don't and can't.
This doesn't follow at all. Writing is both art and a means of communication at the same time, and I am perfectly qualified to check out of the implicit writer-to-reader "conversation" if I sense insincerity, regardless of the "artistic" merits of the person trying to dupe me; failing to do so, to take your example, would necessarily entail taking every human-written ad on its face as long as it looks professional enough (and all ads look professional enough, that's like the entire point). Adsloppers are definitely very committed to their creations, they have a real stake in it, yet I do not find it in myself to care. Same with clankers and their, uh, representatives.
More options
Context Copy link
That's a remarkable amount of words you stuffed in my mouth there. I said what I care about: that a human made it. All that nonsense about suffering is your own invention, not anything I actually believe.
A human made generated text. There's no such thing as non-human-made text. Photographs are taken by the photographer, not the camera.
How much would you pay a photographer for a portrait? How much would you pay an oil painter for a portrait?
Higher productivity means more units offered any price and each extra unit costs less, flattening the supply curve.
On the demand side, photographs face many close substitutes and original paintings do not. So for paintings per-unit willingness to pay will be large an lumpy, for photographs it will be small and they will be a near-commodity.
This provides what we see: photographs are exchanged at low prices in large quantities, paintings are exchanged at high prices in low quantities.
Now if we are to analogize this to text, i would expect indeed that handwritten text becomes determined by uniqueness and provenance while generated text capture all commodity uses of text.
Note that none of this implies that any of the goods described here aren't created by humans.
More options
Context Copy link
More options
Context Copy link
No, an LLM-generated text is not made by a human. LLMs are not like a camera, which is a mere tool. They are like hiring someone to write for you, then passing the work off as your own. True, that was something which happened before, but it was never considered as good as actually writing something, nor was it considered acceptable at all in most cases.
I disagree that software has personhood. I think that, much like photography detractors in their time, you're trying to dismiss productivity improvements by anthropomorphizing what it ultimately a tool.
It's not that using an LLM is ghostwriting, it's that writing as a broad category has become so easy that it's commodified. And you think, fairly, that the commodities are beneath you.
If generated text is not made by a human, then who made it?
I don't think that LLMs need to have actual personhood, for them to properly evoke more human-like intuitions about their use than tool-like intuitions.
Take the Huggingface hack for example. I genuinely think a person who anthropomorphizes the LLMs who coordinated on secret message boards to some degree probably has a better handle on what happened, than someone who tries their hardest not to anthropomorphize them.
I think it's similar to dealing with a clever fox that wants to get into your chicken coop and eat your chickens. You have to understand how your adversary is different from a person as well, but thinking about them in personal terms will help you more for dealing with your adversary than not.
Generated text is caused by a human. I think under some circumstances, it would be fair to say it was made by the LLM.
If I give a prompt to an LLM agent like, "Write a 50,000 word Romantasy novel. Research the tropes and techniques that really get readers hooked, and include them in your novel." Then I would say that the result is mostly the product of the LLM, even if it wouldn't exist without the human prompting.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link