I 100% agree with every sentiment here, I just haven't seen this as a huge problem. I feel like the mods have done a pretty good job recently of removing AI slop. I only consider posts when they get past about 10 Karma too which might be saving me tho...
I always browse from the /allPosts page so I can see posts in chronological order. Somebody has to be the one to read and upvote a new user's post first, or else it won't make it to the front page. But it's exhausting when >75% of new users are likely-bots who write long and meandering posts that look at first glance like word salad.
Hi Michael.
Somebody has to be the one to read and upvote a new user's post first, or else it won't make it to the front page.
Are you implying you are often the one doing this? If so, have you considered just checking the posts on the front page? You would not lose much if you are not interested in the vast majority of the 1st posts from new users.
@Toby Tremlett🔹 , maybe mods could be tasked with voting on the 1st posts of new users.
Couple reasons why I browse /allPosts instead of the front page:
I personally think rate limiting new users would be a good way to help with quality of new posts generally (e.g. maybe you can only comment, at some limited frequency, when you first create an account, and only make top-level posts either via mod approval, or if you hit some reasonable threshhold of comment karma from other people, demonstrating you write things others find valuable to read).
Hi Abraham. It looks like the 1st posts from new users are already being restricted to some extent. Michael Dickens said "Somebody has to be the one to read and upvote a new user's post first, or else it won't make it to the front page".
I think it is better to let the karma system work instead of limiting the number of comments and posts.
Yeah, that's fair — I guess I think that while posts shouldn't be filtered on content ~at all, they should be moderated on stylistic things, and part of the reason I like rate limiting is that it helps people get used to what I view as good community norms (e.g. reasoning transparency, etc). So maybe that's why I am more sympathetic than others to heavy-handed moderation.
(this is at -1 karma when I left this comment). I'm interested in why folks downvoted this? I understand disagree voting, but was surprised by the downvote, so would be helpful to hear your perspective.
BTW, I didn't downvote your comment and I think I understand where you're coming from. It sounds to me like you're remembering what the forum used to be like and are thinking of ways to preserve or perhaps even bring back that quality.
I agree that some level of moderation is necessary and of course there is already moderation on the forum. But whether a post is worth someone's time could also be a choice left up to the reader, no?
So perhaps readers could be given more control over what they want to see rather than go the route of restricting newcomers from contributing. For instance, what if account settings allowed readers to filter their feeds by things like account age or author karma? Then that becomes a choice a reader makes, not the community as a whole. I feel that might allow people one way to manage their own attention without placing a lot of barriers on newcomers, some of whom may genuinely be interested in learning, understanding and becoming part of the community.
I think my main goal with this isn't necessarily about restricting users, but having new users show that the understand community norms (e.g. about how to be clear about reasoning, etc) prior to posting. I agree filters might help too! (I left the downvote comment because I want to encourage people on the forum to be transparent about why they downvote, myself included).
LessWrong gets a lot of these too, but they land in https://www.lesswrong.com/moderation.
As a daily reader of the EA Forum, I have found the experience significantly degraded recently. When I come to the /allPosts page, I am confronted with an onslaught of posts with a single upvote, mostly by brand-new users.
I want to give genuine new users a chance and read their stuff, and maybe upvote it. But when I see a pattern of "brand new user wrote long complicated post that currently has no upvotes", most of the time it's fully AI-generated, and it's a waste of my time to read it.
Even just reading through the titles wears me out and makes me want to leave. There will be maybe 15–20 posts for the day, most of which have long, kind of vague titles that take me a minute to parse. (Example title from today: "From AI Tool to Being? - Memory, Time and Reflection as a Possible Architecture of Artificial Identity")
In principle, I don't object to using AI to clean up your prose. But as far as I can tell, these posts usually have zero human thought behind them.*
Many of these posts are already automatically flagged as AI by Pangram. Rather than flagging and leaving them, these posts should be deleted. Or at minimum, the "AI-generated" flag should be visible from the /allPosts page so I know I can skip them. But really, nobody wants to read those, so what's the point in leaving them to clog up the Forum?
*It looks to me like there are bots roaming the web that are autonomously signing up for and writing posts on EAF.
Could you please give examples of these posts you seem to be saying are wholly without merit? Scrolling through my feed, it isn't evident to me to what you are referring.
From today:
There are also a few posts that are either AI/human hybrids or fully human-written, from one-day-old accounts, with no summary or thesis statement so I can't quickly tell whether they're worth reading. I am not sure what to make of those.
Out of 14 posts today:
That's pretty much how it is every day lately.
LessWrong mods have reported that, apparently, people with AI psychosis are often told by the AI to post their crackpot ideas onto LW, and the mods have to delete a lot of AI slop posts and bad posts by humans with AI psychosis. Could be something similar is going on here, except that the posts aren't getting deleted.
I think that there is value in telling people with AI psychosis about why their ideas are wrong. It helps to get unstuck (if you think that your idea is good, but nobody tells you why, then you are stuck). That helps both the individual and society, because when a person is stuck on wrong idea, then they pay an opportunity cost (and society pays that too).
I think people might underestimate the value of feedback.
Your response identifies these as AI generated, but does not explain why they lack merit. I skimmed the articles and it is not clear to me why these articles like these should be discouraged, let alone purged from the forum.
Beyond literalism explains the problem of narrow rewards and the need to design systems that internalize second order effects.
From AI Tool to Being discusses the role of memory and existing across time as something that brings about selfhood and indicates consciousness, rather than just intelligence.
The Widow Opportunity Map proposes a way to disseminate information such that resources can be directed to a vulnerable population.
With 50 karma to your post, it bothers me considerably how popular your position seems to be. I have been deterred, on the margin, from generating posts sharing my ideas. With a job at the moment that takes over 60h/week and not being a fast writer, for me it's often between doing a series of iterations with Claude/Chatgpt on an idea, or it not being done at all. I feel increasingly alienated from this community, which I have identified with for like 6 years or more, due to this ascendant view I find rather censorious and prejudiced.
I guess I just would ask that people read stuff and evaluate it before condemning it. If something seems to be AI and you don't want to read further, fine. But going on to condemn it simply on that basis seems very wrong to me. It seems very strange that in this community I am in a small minority thinking this.
If people have an interesting idea, they should post it as a quick take instead of feeding into an ai as a prompt and generating the huge post. If I’m interested in what AI has to say, I can feed that idea into an AI myself.
An issue with allowing and not marking AI-generated posts is that it consumes time asymmetrically: normal posts take much longer to write than they take to read, and so a post implies some amount effort that means that it’s a good idea by default for the community to engage with it. AI-generated posts break that, and now the posts are attention sinks that imply nothing about effort spent and consume the attention and time of community members, while also making it a lot less likely that good ideas from newcomers would be noticed.
Hi Brad.
I very much agree with your view. I think posts should be assessed on their merits, and that it is better to let the karma system work instead of limiting the number of comments and posts (e.g. written by AI or new users).
The trouble with this is that AI-generated posts tend to be very long, which makes it a huge time sink to assess their merits.
I do not think a post being difficult to assess makes its existence worse than nothing (which is the bar to meet for a ban). I believe they are fine as long as they do not become too visible, and cause too many people to read it without getting value out of it. I do not see much risk of this. If a post is difficult to assess, it will have little karma, and therefore not gain much visibility.
Decreasing the assessment cost would be good, but there are a few elements already in place to achieve it. The estimated reading time, automatically generated summaries for long posts without one, disclaimer about the extent to which the post was written by AI, and karma, and number of posts and comments of the author (low karma relative to the number of posts and comments will tend to be associated with a less valuable post). If you are finding the assessmt cost too high, you might want to consider asking LLMs to go through posts for you, and just let you know about the ones which are expected to be valuable enough for you to check (based on some criteria that you describe). You can also ask for borderline cases to see if they are calibrated (the borderline cases should not be very worth checking or clearly not worth checking).
Thank you for saying this. A newcomer here with very little karma appreciates your perspective on giving newcomers a chance.
That's exactly why I want stricter rules: I want to read newcomers' posts, but right now they are drowned out by AI posts.
For what it's worth, I appreciate you reading the posts of new users.