Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI slop is killing online communities

rmoff.net

Post reply on HN
Print view View original

701–710 of 781 posts

Re: AI slop is killing online communities

#701

Post by ninkendo »

antisthenes wrote: post #468
ninkendo wrote: post #344

Earlier quoted context omitted.

Would you enjoy the experience of telling your LLM “make a HN-style comment thread on $subject with 200 comments, no trolls please”, and then actually spend time reading them? No? I’m imagining not at least. Because there would be no point to it. If you would enjoy it, then I’m surprised you’re here and not just simulating the experience with your LLM by yourself.

> Would you enjoy the experience of telling your LLM “make a HN-style comment thread on $subject with 200 comments, no trolls please” The reason I'm not simulating the experience with an LLM is because: 1. It costs more time to do so, because I have to prompt it to create a single comment. Multiply that by the typical number of an HN thread. 2. I suppose in a way you need bad takes to form your own view of a topic or…

> That's something you imagined that I claimed I want

This is a thought exercise. I phrased it as "If you would enjoy it", and you clearly have stated reasons why it would be insufficient, so, it doesn't apply to you. If I say "if X then Y, else Z" and you say !X, then of course, !Z applies.

But your original post says "so what" if the people posting here are actually LLM's. And you just answered that question above: "LLMs would also be unable to provide truly unique experiences." So that's what.

To continue the argument I'm making: We're all here because we want unique experiences talking to other humans. None of us would ever think it's ok if this site was just LLM's instead. If we did, we would open up our favorite LLM, not open hacker news. Right? Right. Seems like you agree.

Re: AI slop is killing online communities

#702

Post by mrkstu »

RankingMember wrote: post #683
ricardobayes wrote: post #638

Earlier quoted context omitted.

I always find the most difficult is to deal with users who are not actively breaking any rules but are toxic and hurting the conversation. Given your experience, what's your take on handling this?

Hackernews' approach is hellbans. FWIW that seems to work pretty well here at least.

The combo of quick bans and incrementally earned rights seems to be a pretty effective combo.

Re: AI slop is killing online communities

#703

Post by rgblambda »

armchairhacker wrote: post #653
robotswantdata wrote: post #609

Earlier quoted context omitted.

Slop as a letter is a thing already https://www.axidraw.com/

If someone buys or makes something like this just to send a letter, they a) put in more human effort than writing it themselves, and b) are probably interesting and would benefit your community.

I think what robostwantdata was suggesting was that people would supply a "Use my mechanical pen for a fee" type service, where customers send the text they want written and the address they want the letter sent to.

As I've already said in my direct reply, there's still the various costs involved in posting the letter, and if the community is based on a particular geographical area e.g. "Connacht Hill Farmer's Association", then it would raise eyebrows if the envelope had a Kazakhstani stamp on it.

Re: AI slop is killing online communities

#704

Post by rgblambda »

000000000001 wrote: post #639
account42 wrote: post #616

Earlier quoted context omitted.

Adding that much friction is also going to loose you many genuine users. Might be worth it depending on the community but if it makes newcomers fewer than your usual churn rate its a death sentence.

I think a better approach would be to upload an image to the sign up form with explicit instructions that isn't in text on the page to be easily read by AI. Like "Answer 2 in the Math captcha" even though the correct answer is 5, the AI will always enter the wrong number.

That would work for a time, like with Captchas. Then the AI would overcome it and you'd have to come up with a new idea.

Re: AI slop is killing online communities

#705

Post by RiverCrochet »

aliasxneo wrote: post #96
colechristensen wrote: post #36

Earlier quoted context omitted.

One of the paradoxical things that makes me hopeful is that there's going to be such an incredible amount of low effort AI slop content that it's going to drown out the low effort human-made content and generate a large amount of distaste for it. So much will be so bad that good taste and high quality will be rewarded more status as the people who will say and believe anything will be led astray and left behind. Mayb…

> So much will be so bad that good taste and high quality will be rewarded more status as the people who will say and believe anything will be led astray and left behind. I really want to believe this will be true. However, I also suspect there's some external driving force, that I cannot readily name, which is making people incapable of consuming anything except this low-effort content. I mean, obviously it's workin…

The "couch potato" force that in the past, made people sit on their couches and watch hours of TV simply because it was the easiest thing for them to do.

It's fine though. Low-effort people deserve low-effort content. Maybe the work of high-effort creators and the patrons that support their works will simply not be on the Internet in the future.

Re: AI slop is killing online communities

#706

Post by RiverCrochet »

mrhottakes wrote: post #58
colechristensen wrote: post #36

Earlier quoted context omitted.

One of the paradoxical things that makes me hopeful is that there's going to be such an incredible amount of low effort AI slop content that it's going to drown out the low effort human-made content and generate a large amount of distaste for it. So much will be so bad that good taste and high quality will be rewarded more status as the people who will say and believe anything will be led astray and left behind. Mayb…

Have you considered that (further) lowering the signal-to-noise ratio will make it much more difficult to find and distinguish a signal?

If every TV channel is static, watching TV is not productive, and the correct thing to do is stop watching TV.

Re: AI slop is killing online communities

#707

Post by donpark »

ricardobayes wrote: post #638
CrzyLngPwd wrote: post #55

I run a niche creative community, and we outlawed AI-generated content in 2022 as it was easy to see how corrosive it would be to the community. It hasn't been easy. We ban fake AI accounts daily and shrug off around 600 AI content creator accounts monthly. It's a lot of work, extra work that wasn't needed before AI content came around, and of course, that is an extra cost. I fear losing the battle.

I always find the most difficult is to deal with users who are not actively breaking any rules but are toxic and hurting the conversation. Given your experience, what's your take on handling this?

Let users flag toxic users. Use the resulting toxicity score to filter out messages based on each reader’s toxicity tolerance level. Both the toxicity score and the tolerance level are driven by user flagging, with score decay over time to account for improved behavior.

Re: AI slop is killing online communities

#708

Post by 81AB24FB »

card_zero wrote: post #623
lordfrito wrote: post #614

Earlier quoted context omitted.

> I dont know why people would want to be in a community where they arent wanted. This is standard predatory behavior. Child abusers hanging out with kids, weirdos hanging out near the women's clothing department, etc. It's usually a clear indication of the sort of people you don't want to associate with in your online community. They bring a net negative to the table.

What is "it"? Putting the two halves together, the sort of people who want to be in a community where they aren't wanted are the sort of people you don't want in that community. I guess I can't argue with that.

They are talking about social norms. Inversely, "creepers".

Most adults understand why men should not, generally, be hanging out in the women's clothing department. When accidental violations of those norms are pointed out, they apologize and correct. Creepers, OTOH, gonna creep.

For their own well-being, online communities should police repeated violation of social norms. Otherwise the normals leave and creepers take over.

Re: AI slop is killing online communities

#710

Post by kazinator »

ricardobayes wrote: post #638
CrzyLngPwd wrote: post #55

I run a niche creative community, and we outlawed AI-generated content in 2022 as it was easy to see how corrosive it would be to the community. It hasn't been easy. We ban fake AI accounts daily and shrug off around 600 AI content creator accounts monthly. It's a lot of work, extra work that wasn't needed before AI content came around, and of course, that is an extra cost. I fear losing the battle.

I always find the most difficult is to deal with users who are not actively breaking any rules but are toxic and hurting the conversation. Given your experience, what's your take on handling this?

You identify what it is that they are doing that is hurting the conversation and articulate it into a written guideline.

Why, examples of how to do this can be found in the guidelines for this very HackerNews site!

Post reply on HN