The EA Forum does a lot of things. It’s a place to talk about EA community issues, share jobs, updates from organisations, ask for support with new projects, meet people etc…

All of this is good, but as we all know, not all good things are equally good.

One of the most impactful things the EA Forum does is enable us to have important conversations that could meaningfully shift the focus of funding and talent in the community^[1]. As EA grows, I want the Forum to maintain that function, and, indeed, get even better at it.

There will be a number of interventions in this direction in the near future.

The first is the ‘featured page’, launching today. This is a new homepage for the EA Forum which allows us to highlight the posts we (the EA Forum team and moderators, or, for this experimental period: me) think are likely to be the most impactful.

If this works it will a) reward the authors of particularly valuable posts, who will have more readership and have their writing elevated and b) make it easier for readers to find the very best of the EA Forum, however often they turn up.

This page is experimental, we will be collecting data and may revert changes in a few weeks. More details below.

How it works

The regular old frontpage will be re-titled ‘New & Upvoted’. You can access it at the top of the homepage. You will be able to toggle between ‘Featured’ and ‘New & Upvoted’:

 When logged in, clicking on either heading will set it as your default homepage. Next time you visit the site, it's the first page you'll see. 

New comments and posts on the featured page will be highlighted in EA teal. Based on the last month, I'd guess that there will be an average of 6 posts added by the mods every week, and an average of 2 by the Forum audience. Details on how posts are selected is in the 'What goes on the featured page?' section. 

What is the featured page for?

Getting (a) more attention for and (b) incentivising the creation of (c) posts which could  meaningfully update (d) an impartial altruist trying to do the most good with their resources.

(a) more attention for: curating a post massively increases the attention that it gets, this will be another lever like that (as long as people find the featured page useful enough).

(b) incentivising the creation of: curation or inclusion in the digest, especially when done by another human being, feels great! The featured page will bestow prestige on posts and therefore authors, and (as long as the criteria are legible) will encourage people to write for it.

(c) meaningfully update: we want people to be changing their minds, for good reasons. The ‘for good reasons’ part depends on the quality of content created by the audience, and selected by the mods.

(d)  an impartial altruist trying to do the most good with their resources: This category contains early career EAs who are still more cause neutral, people in long term EA roles, and the decision-makers and leaders in the movement. 

...well that's the theory. There are also a few other motivating reasons for this page:

• We get more high quality posts than I can curate, as well as high effort and good posts that fall short of curation, and I'd like to reward them in some way.

• CEA is growing the EA movement, and more and more text is being AI generated. Both are reasons to expect an incoming wave of low-average quality content. I want to make sure the Forum has the features it needs to deal with this. I don't love the default, where the Forum audience has to search through the slop. Selecting the best is less resource intensive than removing both the worst and the not-great. Hence... Featured page.^[2] 

• Some users we really don't want to lose (i.e. the decision-makers that many authors on the Forum want to influence) want a higher signal version of the Forum. The Featured page will be pre-sorted page, where they won't miss the best ideas coming out of the Forum. 

What goes on the featured page?

The current one sentence explanation at the top of the featured page is:

‘Selected posts to help us answer: how can we do the most good with our resources?’

We’ll be looking to highlight posts which are a) good, and b) could help an impartial altruist meaningfully update their thinking about how to do good with their resources.

We aren't doing an independent consequentialist analysis for the value of each post. The aim is for 'featured' to be a transparent category that the moderation team (and ideally the Forum audience) is aligned on. 

To avoid mod biases, blindspots or unpopular opinions stopping particular kinds of content from making it onto the featured page, we’ll also include ~all^[3] posts above 100 karma.^[4] 

Plus, you can dm me if you think a post should be featured that isn't - if this happens a lot, we'll make some easier way for users to flag posts as suggested. 

Defining 'featured'

Necessary conditions to be on the featured page

• The post must be good: engaging, clear and honest.
  • Engaging: i.e. if you post the contents of an arxiv paper rather than a summary, it likely won’t be featured. A featured post should be readable.
  • Clear: not super jargon filled or AI-written to the point of unparseability.
  • Honest: Not clickbait, not using any dishonest persuasive tactics (exaggeration, excessively emotive or angry language, etc…)
• The post must be ‘on topic’, in other words:
  • It makes an argument or shares information that could change how an impartial altruist prioritises — across causes, within a cause, or in how they deploy their resources, including their career.
    • This is EA, so I’m including a few layers of meta in here; philosophical arguments and quite abstract ones can be included.

  • It is written from or conversant with the effective altruist frame^[5].

    • I include ‘conversant with’ to indicate that you don’t have to agree with the EA frame to be on topic. This bullet is just ruling out well written things on EA topics that are scope insensitive, don’t seem impartial etc… (as more and more AI-related stuff will likely be).

What we exclude from the featured page:

Note that ‘community’ ‘announcement’ ‘opportunity to take action’ and ‘organisation update’ posts will be excluded from the >100 karma inclusion. Mods can still feature them manually.

• Basically all ‘community’ posts — community posts are far from valueless, but my take is that they are incentivised enough already. I wouldn’t want an EA Forum with no community posts, but a good community post already gets a bunch of karma, readers and comments, partially because they are much easier to engage with than an object-level post.
  • For an example of posts that would be featured despite being community, I’d look at some of @Mjreard's posts (like this one), which make or kick off prioritisation arguments, but stay very EA-community-coded.
• Bare organisation updates — no job postings, RFP postings etc… unless they contain independently good analysis that meets the bar.
• Newsy and quick to depreciate calls to action: "Do X now!" I want these on the Forum, but the featured page should be more evergreen.
• Announcements — of a new fellowship, new video, new podcast etc… Again, these are fine to have on the Forum, but not on the featured page, unless they are independently substantive posts.

This is an experiment

The featured page is a pretty big change to how the EA Forum works, and maintaining it will take a lot of curation work. As such, I want to make sure it’s useful before we make it a permanent feature^[6].

Based on data we collect over the next few weeks, we’ll decide whether to revert changes, or double down on the featured page. To help us make the right decision, please:

• Give us some feedback.
• If you normally are logged in to the EA Forum, make sure you still are. This helps us measure the time spent on featured posts vs other posts.

I would genuinely love your feedback

Please share any takes, concerns, suggestions, or encouragement in the comments, via dm, or through this form (which is optionally anonymous):

I am fully expecting to tweak the visuals, the inclusion criteria, etc… So don’t feel like anything is out of bounds. I'm ready to debate inclusion criteria in the comments if people are keen. 

Acknowledgements

Massive shout out to @Ollie Etherington for building this page the old fashioned way (typing code himself) and to @Sarah Cheng 🔸 and @Agnes Hasselblad 🔸 for several rounds of mock-ups and feedback. Mistakes or bad decisions are my own. 

1. ^{^}

    Me and @Sarah Cheng 🔸  have been working on the EA Forum’s theory of change a lot this year. I’ll have something to post on this fairly soon, but feel free to ask questions in the comments.

2. ^{^}

   I'm not saying that everything not on the featured page is bad - that is not the case. In a world where we had to fight a lot more slop, I'd have to find a solution for org updates, community posts etc... as well. 

3. ^{^}

    Excluding those tagged community, organisation update, announcement and opportunity to take action unless overridden by a mod.

^{^}

You might wonder specifically about whether critiques of CEA and the Forum would make it onto the featured page. My guess at this stage is that many wouldn't because they would be firmly community posts (of interest to an EA but unlikely to shift any resources). However, critiques like "The Forum team/CEA is failing in X way, we need an alternative organisation/platform" would make it on if good, because they could in principle affect the flow of resources (they might lead to people funding or founding an alternative). Call me out if you think I'm getting this call wrong. 

^{^}

 Key aspects of the effective altruist frame to bear in mind: cause neutral, scope sensitive, broadly impartial, welfare maximising.

^{^}

 Also, if we want to stay relevant in an age of AI writing and Substack popping off (and we do), we need to move fast. I want to be more experimental with the Forum product, without risking the core values of the Forum. Bounded and transparent experiments are a nice way to do that.

Show all footnotes