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 -
Count. My good sir.
I know someone using an LLM when I see it. I know you've used one. The issue is that I do not know it you were lazy and simply typed a short prompt and got Claude (in all likelihood) to do all the work, or if you shared a draft and had it edit it for you. Or something in between.
Why? Because I use them everyday. I can smell it. I also know that some effort was made to not make it glaringly obvious, but it takes a thief to catch one, at least depending on how busy/lazy the police are.
I suspect you might deserve a warning or a ban for this. The only reason I'm not doing it is because I genuinely do not know if you have used it in the manner I use it, which is minimally and doing the bulk of the labor. There are no facts here to fact check, per se, at least not in an objective sense, or I'd look for hallucinations.
Untouched by the light as you are since birth, I still wish to extend you a little charity and benefit of doubt, though I worry even that's too much. I am too tired to even get into the weeds of this, or at least how I pick up on the tells. But not too tired to not throw your post in Pangram, which says this is 17% AI. Nor too tired to remind everyone that I have used Pangram on my own edited writing and in semi-formal experimentation, and noticed that it has few false positives but many false negatives, including underestimating the degree of AI contribution. I'd eat my non-existent hat if a mere 17% of the post came from an AI. The issue is that I do not know if was 20% or 90%.
Unlike you, I feel little need to feel morally unclean because I have a pretty clean track record of good-faith, high effort engagement well before LLMs, and also because I disclose my usage to whoever asks, if I've used one to any degree. You really ought to understand that we're not the same, at least in my eyes. You still benefit from my fear of hypocrisy or error.
When all of God's Creations sing in heaven, won't a shitposter be part of the choir?
So I won't do anything. At least not right now. If the other moderators step in and whack you, I will defer to them. Alternatively, if they're busy, I will assess general sentiment and act accordingly when I'm back. I am genuinely okay with anything from tolerance, to a warning, to any kind of ban. You belong in a museum or a zoo, maybe a zoological museum. I'd pay for entry.
I didn't read his comment as it seemed long+boring. After reading yours, I went back to skim it. Even with the knowledge he did use AI, I have a hard time activating my "AI writing radar"
I pride myself on being good at detecting LLM text. I will have to think about this.
Even better that you clocked it as Claude, nicely done.
I should use Claude more I guess, I clearly don't have a taste for its style like I do ChatGPT/Gemini.
More options
Context Copy link
Most of the earlier parts of his post has telltale signs of not being AI generated, it's only in the latter parts where it looks more like a mix between AI and human.
IMO the incisiveness is what really matters. AI is good at structuring sentences but bad at naturally coming up with incisive ideas on-par with the average higher quality substack or poster here. A little bit of AI use shouldn't trigger a modhat, though I'm fine with a post that looks entirely AI-made and which has no real ideas other than vague AIisms being modded.
More options
Context Copy link
Ah, well done. I've had this idea for months now and wrote up a semi rambling draft of the post around midnight yesterday after a long day, asked Claude Opus 4.6 to take it and give it structure but didn't like the result, it was too bland and read like a "safe unsafe" newspaper op ed you'd find from one of the better writers in The Financial Times, asked for something less corporate and "safe" but no matter how I phrased it all it did was take the text and make it more rightoid, which absolutely wasn't what I was going for, so in the end I decided to take the Claudified Newspaper Op Ed and build the final post upon it from there.
But yes I agree, if you squint the newspaper OP Ed still comes through, for example this paragraph is something I'd never write, or at least never frame in this specific way:
I'd instead have changed it to talk about how it still had a culture of accepting as its input lower human capital from the French upper middle class over genuine talent from the lower classes (however little there may be, it still exists). Again, not because they deliberately shun such people, or at least no right thinking person (and we can be sure the ENA was chock full of right thinking people) does so overtly, but rather because of structural failures meaning that lower class talent never even seriously considered applying to the ENA, and the institution was fine with that continuing state of affairs. But again, that's a problem with the filter applied to at the beginning of the ENA process, it doesn't have any bearing on whether the forge once you are inside the process is of poor quality, and we can see that even the mediocre spawn of the French UMC emerged from the ENA better for the experience. The issues with the ENA was the "American" part of it, not the continental part. A shame that the baby had to be thrown out with the bathwater to satisfy the baying mob, but the common people have never really understood nuance.
I did change many many paragraphs from what Claude gave me and rewrote them from scratch or at most only taking some rhetorical turns of phrase that worked well and incorporating them into my own paragraph. In my daze at 1am yesterday I missed that paragraph and looking at the post again this morning some other things that I'd phrase differently if were an editor and the above post had been given to me.
Think of it as someone who designed the blueprints for a building themselves, outsourced the skeleton structure to a contractor, wasn't happy with the result they gave and so decided to finish the rest of the building themselves rather than knock it down and start from scratch. I genuinely don't know what percentage is "AI" or even whether you can put a number on it given my original draft and how heavily I worked the Op Ed it gave me into something that was fit to print. But if you insist, my median guess is going to be something like 33%, give or take.
Now, what did I tell you about agreeing with or complimenting me? It keeps me up at night, or so I'd say, if my sleep cycle wasn't shot and if I wasn't already the consumer of a cup of strong coffee.
I mostly believe you. Mostly. And mostly because you are the honest kind of criminal, once we have you in the lockup - you're usually kind enough to submit a written confession before we need to pull out the nightsticks or stage a shoot-out.
It would be... nice, of course, if you did that proactively without needing the trip in the police van. Have you seen gas prices these days? But I can't hold that against you, because I only disclose the usage of AI when required or when someone asks.
Go on. Shoo. Live to see another sunset, or continue giving Anthropic engineers a headache. I think that if the other mods were likely to act, they would have acted. I leave that door open for them, if they so choose.
More options
Context Copy link
More options
Context Copy link
Reads to me as even more slop infused than your own forays, which as you know I think are counter to the current official stance on LLM-posting -- my question as usual is: if bot-posts are officially not allowed, but nobody is ever willing to do anything about it, in what sense are they not allowed?
(I'm aware that you personally think that they should be allowed, but AFAIK this is not in fact the position of the mod-team as a whole; the law of the land so to speak?)
The official position is that AI usage is allowed, but cannot constitute the substance of one’s comment. If it wouldn’t pass the “low effort” rules without the AI additions, we’re probably going to mod it.
@self_made_human has modded accordingly.
Unfortunately, our best examples of what isn’t allowed tend to stop at the new user filter. You’d be surprised how many psycho-political manifestos we get from first-time posters.
An actual, obvious slippery slope. Poor rulemaking imo.
What would you recommend?
I would go for what should at least have some chance of keeping the human text production factor high on this site. Prohibit AI for text production - on the honor code, same as with other rules for good conduct. Some AI help with idea creation is ok.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
Then there is no rule at all -- and the slop-lovers among us will continue to push the line until the forum is largely bots arguing with each other.
Bad way to go.
More options
Context Copy link
Figuring out where the line is here is a fun and impossible task. I like the "if it would be considered low effort if you removed the AI text" rule, but then if a comment has AI writing sufficiently integrated into it throughout, I guess that means it is just invalid period? But at the same time, a sufficiently integrated comment would either read as 100% AI or its edited well enough that no one would know. I love thinking about this.
I know you can't encourage them, but I'm sad we don't get to see these. Now that I don't use 4chan anymore my exposure to skitzo-slop is too low. I miss stuff like the The Philmarilion and the rock guy.
More options
Context Copy link
More options
Context Copy link
My understanding of current consensus is that entirely or majority LLM written posts are banned.
The problem is that there is no consensus beyond that. If we had a rule, informal or not, that a suspected X% of AI is the cutoff for action, I would enforce that, even if I think that the acceptable value of X is larger than most.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link