This post was cross posted from LessWrong

TL;DR: 

• If you enjoy LessWrong (and/or do EA/AI safety field building), you should consider starting a rationality reading group at your university. It’s fun, cheap, and a good way to meet like-minded friends, and it can reach rationalist-ish people who’ve never encountered EA or AI safety.
• Starting one can be as easy as grabbing a couple friends, picking a time, place, and reading, sending an email blast, and bringing food. The ACX Meetups Everywhere call (organizer form due October 9th) can be a good way to find an initial group.

Context: this post was loosely based on a memo originally written by Noah Birnbaum, and I (Carolanne Jiang) most of the current version. “I” in this post refers to Carolanne, and “we” refers to Carolanne and Noah.

Also, UChicago Rationality is still going! If you are a student at UChicago interested in joining, meetings this quarter are on Wednesdays from 6 :00 – 7:00 PM in the Reg. Noah and I are both no longer at UChicago–for more information, please reach out to Will Kyle at [email protected]

UChicago Rationality was a huge success: it was very fun, a great place to make like-minded friends, and responsible for getting various people^[1] into AI safety and/or EA. If you're a frequent reader of LessWrong, an organizer for your university's EA or AI safety group, or are otherwise interested in rationality as a university student, it is possible that you'd have a great time running a rationality group at your school. This post is our strong encouragement for you to consider doing so, and guidance on how to do so effectively.

The backstory of UChicago Rationality

A few years ago, Noah Birnbaum was interested in rationality and field-building for EA and AI safety. He noticed that many rationality-curious people at UChicago had never really encountered EA or AI safety, so he started a reading group, partly to meet like-minded people and partly in the hope that the communities would start to overlap.

This took form as a weekly reading group of about 20 people, where we read the Sequences, Scott Alexander's blog posts, and other rationalist and adjacent writing. The ACX Meetups coordinator gave us funding to get sandwiches at a nearby restaurant, Scott kindly agreed to promote us on his blog, and the group grew so quickly that we sometimes had to cap attendance. For much of the club's history, the group hosted very high-quality conversations and brought together people who otherwise wouldn't have met, several of whom went on to get involved in EA and/or AI safety, including me (Carolanne).

More reasons on why you should consider running a rationality group

• It is fun (and low cost).^[2] You get to read and talk about ideas you find interesting with friends and like-minded people, and you can get funding for e.g. food at club meetings (most clubs can get funding through their school, but there are other sources of funding to help make ambitious clubs possible).
• You can build the community you wish existed! There are often more rationalist-ish people around than you'd expect,^[3] and a group gives them a way to find each other. At UChicago, the group surfaced many of these people, and I met lots of new friends there.
• If you're also interested in EA or AI safety community building (which Noah argues that you should be), a rationality group is a great way to reach people who might get a lot out of EA or AI safety but haven't encountered them yet.^[4] 
  • Initial outreach in community building can be understood as sets of overlapping but distinct selection effects. Many forms of EA and AI safety outreach reaches people who are already somewhat interested, while a rationality group can attract people who share many of the same habits of mind but have never had a reason to look into EA or AI safety.
  • "Come meet friends and talk about interesting ideas over food" is also a much lower barrier to entry (& much lower commitment/investment) than an EA fellowship or AI safety event, so the invitation can go out to far more people.
  • If you intend to start a group with the primary purpose of serving as a pipeline into EA or AI safety,^[5] we strongly encourage you to make this intention clear to people, and recognize in advance that (& be okay with the fact that!) many folks will enjoy the rationality content but won’t be interested in EA/AIS.^[6]

Q: I am sold. How do I go about starting one?

• There is a call for ACX meetups everywhere organizers right now. If you’re interested in this round–you should fill out the organizer form by October 9th. This is a great opportunity to attract an initial group of people!^[7] (The post includes also useful information about how to host your first meetup)
• You could also start by thinking of it as “a group of friends hanging out and reading interesting things together.” It doesn’t have to be some big, ambitious club: just grab a couple of friends, pick a time, a place, and a reading, send out an email blast, and bring food. That’s pretty much all there is to a minimal (and excellent!) rationality club.^[8]
  • (If you want to be more ambitious, there are lots of general resources about starting university groups. Here is one I find particularly helpful and clear. It was originally written about starting forecasting groups, but lots of content here can be adapted to rationality groups, including how to find co-organizers, general outreach guide, general tips for great event logistics, how to send emails, etc.)
• If you’re interested in starting a group and/or have more questions, please feel more than welcome to reach out to me or Skyler directly.

Appendix 

Counters arguments to why a rationality group might not contribute to field building for AI safety and/or EA

At the time of the UChicago Rationality group, LessWrong had somewhat less AI safety content compared to now, so perhaps most people interested in reading LW have already gone into AI safety.

Empirically, this does not seem to be the case; there seems to still be lots of talented rationalist-ish people not in AI safety. However, there may be reasons why they’re not in AI safety unrelated to the lack of opportunities, but I don’t think this should be a major concern. In addition, I believe this effect is likely to be minimal, as many college aged students have not fully decided what they want to do with their careers.^[9]

The discussions were particularly good because there was good facilitation/talent scouting, which may not be easy to replicate.

We have some thoughts about how to make discussions go well, (e.g. the discussion should be very focused on specific points & argumentation), and would be happy to write more about this if there is interest.

Maybe UChicago has a disproportionately large number of people interested in rationalist-ish things

This seems relatively unlikely. To the extent that the concentration of rationalist-ish people may vary across universities, we expect it to depend more on factors like school size, academic culture, and the strength of related communities, etc. than on anything unique to UChicago. In addition, even if interest is somewhat concentrated, a good group could work from a handful of engaged people.

Thanks to Skyler Crossman and Saul Munn for comments and suggestions, and Noah Birnbaum for the initial memo and feedback. Errors remain my own.

1. ^{^}

   Over 2 years, we had ~20 people at most meetings (with frequent new faces), and ~150 on the mailing list. 

   By Noah’s rough estimate, the group was responsible for something like 10–30% of the people who got involved in UChicago EA, people who likely wouldn’t have found it otherwise, while taking only about 1.5–2 hours a week of the main organizer’s time.

   More broadly, about 15% of EA Survey 2024 respondents first heard of EA via ACX or LessWrong, and about 20–25% cite each as important for getting involved (somewhat lower if you exclude respondents who found the survey through ACX). In Coefficient Giving’s 2023 survey of people working on (or likely to work on) AI safety and other catastrophic risks, university groups were the most common top-4 career influence (25%), and Eliezer Yudkowsky’s writing was listed by 11%.

2. ^{^}

   UChicago rationality took 1.5–2 hours a week of the main organizer’s time. This mostly looked like: picking a piece of writing, maybe re-reading it, booking a room, sending an email to the group, personally texting folks to join (which is very effective), and making a sandwich order. 

   Further more, if you have two concurrent reading groups that are 2h each and require 30m of preparation from you, that’s ~5h/week of total work (plus maybe another 30-60m of admin work, eg. paperwork with the university). 

3. ^{^}

   Empirically, LessWrong, ACX, and related blogs have lots of readers, a non-trivial amount of whom are currently in university, and there aren’t that many universities out there. So, it is probably quite likely that there are a few at your school.

   Anecdotally, we just keep finding rationalist-ish people and Scott Alexander readers who didn’t know about any related groups at UChicago. They were mostly all excited to come to the rationality group.

^{^}

Counter arguments: see appendix

^{^}

Noah started the group explicitly with the goal of getting people into EA while stating his intention as such.

^{^}

If this is your goal, you may want to make sure that your successors are aligned to these goals. Alternatively, if you just want to enjoy running a reading group and don’t care about encouraging people to take action, that is fine as well. (and reading groups are great fun!)

^{^}

We hosted 2 years of ACX meetups at UChicago, both of which were very well attended.

^{^}

Less related to why you should start a rationality club in particular: consider gathering friends to read whatever you’ve been meaning to get around to, or whatever you enjoy and want to talk about with others! 

eg. Future of Humanity Institute reading group (cc Lydia Nottingham & Saul Munn), Joe Carlsmith reading group (cc Saul Munn), Global Priorities Institute reading group, Forethought reading group, Sequences reading group (cc Lighthaven), or a book club in general (cc Jo Jiao), etc.

^{^}

Skyler adds: many people find the rationalist community (or AI Safety or EA) in university, around 18~22. It's typically a time when one starts reading more broadly and bumping into more ideas and interest groups

Show all footnotes