I do independent research on EA topics. I write about whatever seems important, tractable, and interesting (to me).
I have a website: https://mdickens.me/ Much of the content on my website gets cross-posted to the EA Forum, but I also write about some non-EA stuff over there.
I used to work as a software developer at Affirm.
This gap isn’t too substantial IMO. Compare to the gap on something like immigration (78% of Republicans support a border wall vs. 21% of Democrats) or abortion (20% of Republicans say it’s morally acceptable, vs. 78% of Democrats).
Thanks! I wouldn’t describe this as a negative result, more just that it’s not obviously a great idea, and there are some important downsides. Further investigation might reveal that it is, in fact, a great idea. (Probably not, but maybe.)
Yeah I was surprised in pretty much all the same ways as you!
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.
Data is not contaminated, as only the first attempt is counted. Feel free to take it as much as you want!
How many bullets were you able to get? I know it’s possible to get at least 12 but there might be a way to get more.
Couple reasons why I browse /allPosts instead of the front page:
Front page turnover is very slow—posts stay there for multiple days. I check EAF multiple times per day most days.
I want to at least read the title of very post. Many posts never show up on the front page.
On completing the quiz with no conflicts I was invited to take the quiz again to explore possible conflicts. Did you exclude repeat quiz takers from your data analysis?
Yes, if you take it multiple times, only the first one is included in the analysis.
Why does the Unjournal Bot keep replying to me? e.g. https://forum.effectivealtruism.org/posts/fFDM9RNckMC6ndtYZ/david_reinstein-s-shortform?commentId=9nuWSjdSBZZ9Fw5SH
Most respondents came from the EA Forum, where I suspect the total view is more popular than in academia. I think if you surveyed moral philosophers you’d see more diversity across named views.
I did think about ways to broaden the sample. I considered paying MTurk but that would’ve been kind of expensive and I’m not sure how to prevent people from using AI to take the quiz. I didn’t even think of trying to survey academic philosophers, I don’t know how I’d do that but I’d be interested to see the results.
The total view is on the rise*, and yet family sizes are getting smaller. What a strange turn of events!
*I have no idea if this is true
At least in some cases it’s clear what a disagree-vote means, and it’s often a useful signal to know how many people agree or disagree. Especially if there’s something I haven’t really thought about, and seeing the votes gives a quick vibe check.
For a lengthy post (i.e., this one) where someone clicks “disagree,” I guess it’s just with the top-line message? I don’t know if that assumption is correct.
Yes I share this frustration. I often see agree-votes or disagree-votes that I don’t know how to interpret. LessWrong has the ability to (non-anonymously) highlight specific passages and agree-vote or disagree-vote just those passages, which I like.
I disagree-voted, here is my explanation of why:
It is good for people to write a comment explaining their disagreement. But if you de-anonymize agree-voting, the main outcome won’t be that people explain their reasoning; it will be that people stop voting. Some people will not feel comfortable attaching their name to a dissenting vote.
I think it’s better to make it easy for people to leave cheap anonymous signals. Writing an explanatory comment should be encouraged but not required.
the Forum doesn’t seem great for AI discussion.
FWIW it’s very easy to cross-post between LW and EAF, so I post all my AI-related posts on both.
Note: It’s easiest from the EAF side because there’s a “Cross-post to LessWrong” checkbox.
That’s exactly why I want stricter rules: I want to read newcomers’ posts, but right now they are drowned out by AI posts.
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.
Congress Probably Won’t Pass a Kill Switch, Even Now
The market is currently at 30% which actually sounds really high to me? A year ago, the odds of a bill like this getting through Congress would’ve been, what, 3%? Even now, 30% sounds high, but I hope Metaculus is right.
I think a big reason for this is that a lot of people are focused more on AI risk than they used to be, and AI risk discussion happens mostly on LessWrong, and a lot of LessWrong power users do not like the culture on the EA Forum (for reasons I don’t fully understand).
One anecdote that comes to mind is that Holden’s post on Anthropic’s RSP v3 got only 8 comments on EAF compared to 83 comments on LW, and the top comment on EAF was just me copy/pasting some of the top comments from LW.
Also, most of the posts on EAF lately are bad bot posts (see my post from today). But if you just look at posts with >30 upvotes or so, I think their quality is still about as good as ever.
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.
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:
5 are definitely real posts from real people
3 are definitely fully AI-generated
3 are not flagged as AI-generated, but they’re written by one- or two-day-old accounts and I am struggling to extract any semantic content by skimming them
3 are written by much older accounts, but those accounts have only one or two posts ever and zero comments
That’s pretty much how it is every day lately.
Update: I have noticed a suspicious pattern with AI-generated user profiles. Profiles that fit this pattern are not always AI-generated, but it’s a red flag:
profile picture with a face
bio line always says “[senior or fancy-sounding title] @ [company/org]”
full bio has multiple paragraphs
like they’re treating EAF as if it’s LinkedIn