The Berkeley Community & The Rest Of Us: A Response to Zvi & Benquo

Background Context, And How to Read This Post

This post is inspired by and a continuation of comments I made on the post ‘What is the Rationalist Berkeley Community’s Culture?’ by Zvi on his blog Don’t Worry About the Vase. As a community organizer both online and in-person in Vancouver, Canada, my goal was to fill in what appeared to be some gaps in the conversation among rationalists mostly focused on the Berkeley community. Zvi’s post was part of a broader conversation pertaining to rationalist community dynamics within Berkeley.

My commentary pertains to the dynamics between the Bay Area and other local rationality communities, informed by my own experience in Vancouver and those of rationalists elsewhere. The below should not be taken be taken as comment on rationalist community dynamics within the Bay Area. This post should be considered an off-shoot from the original conversation Zvi was contributing to. For full context, please read Zvi’s original post.


I. The Rationality Community: Berkeley vs. The World

While I didn’t respond to them at the time, several community members commented on Zvi’s post they had similar experiences: that while some local rationality communities and their members perceive themselves in a zero-sum game with Berkeley they didn’t sign up for (and, to be fair, the Berkeley community didn’t consciously initiate as though it’s a single agency), and some don’t, a sense of what Zvi was trying to point appears ubiquitous. An example:

In my experience, the recruitment to Berkeley was very aggressive. Sometimes it felt like: “if you don’t want to move to Berkeley as soon as possible, you are not *really* rational, and then it is a waste of our time to even talk to you.” I totally understand why having more rationalists around you is awesome, but trying to move everyone into one city feels like an overkill.

Similar anecdata from local rationality communities around the world:

Melbourne. When I met several rationalists originally from Melbourne in Berkeley a few years ago, the way they talked about the exodus of the core of the Melbourne rationality community to the Bay Area, it was a mixed assessment. Melbourne is an example of very successful local rationality community outside the Bay Area, with the usual milestones like successful EA non-profits, for-profit start-ups and rationalist sharehouses. So that many rationalists from Melbourne left for the Bay Area passed a cost-benefit analysis as high-impact individuals it was obvious to them they should be reducing existential risks on the other side of the world.

In conversation, Helen Toner expressed some unease that a local rationality community which had successfully become a rationality hub second only to the Bay Area had had a whole generation of rationalists from Melbourne leave at once. This could have left open the possibility a sustainable system for rationalist development for years had been gutted. My impression since then is around this time the independent organization of the Melbourne EA community began to pick up, and between that and the remaining rationalists, the Melbourne community is doing well. If past or present members of the Melbourne rationality community would like to add their two cents, it would be greatly appreciated.

The rationality community growth strategy out of Berkeley by default became to recruit the best rationalists from local communities around the world at a rate faster than rationalist organizers could replenish the strength of those local communities. Given the stories I’ve heard from outside Melbourne being more lopsided, with the organization of local rationality communities utterly collapsing, only recovering after multiple years if ever, I’d consider the case of Melbourne rationality community surviving the exit of its leadership for Berkeley to have been a lucky outlier.


Seattle. The Seattle rationality community has experienced a bad case of exodus to Berkeley over the last few years. My understanding of this story is as follow:

    • Like with rationalists around the world, effective altruism came along and said “hey, while our communities have significant differences, we care about existential risk reduction and other common goals; we’ve got several billion dollars; and worldwide network of thousands rising through every kind of institution to coordinate the globe”. At the time, the whole strategy for AI alignment wasn’t much more than “read the Sequences and then donate to MIRI...?”, so at the time EA’s value proposition couldn’t be beat. In Seattle the organizers of the rationality community took off their rationalist hats and switched it for an effective altruist one, albeit while prominently placing a rationalist button on it. This is what started happening in Vancouver as well circa 2013. The Seattle rationalists started a successful Rationality Reading Group in 2015 which got through the whole LessWrong Sequences.

    • Things went swimmingly in Seattle until AI safety ‘went mainstream’, and as the financial resources flowed into the institutions of the Berkeley rationality community, the demand and pressure to acquire the resources that were distant rationalists and their skill-sets intensified. In a period of several months but less than two years, the Seattle rationality community lost at least a half-dozen members, including some local organizers and other veteran community members. The Rationality Reading Groups ceased as regular meetups for over a year, and local community organization was at best intermittent.

    • The excitement of EA brought many more Seattleites into the world of x-risk reduction, and the EA and rationality communities of Seattle effectively merged to survive. Since then, they’re thriving again, but Seattle is still gradually exuding community members to Berkeley. Because of its proximity to the Bay Area, and the excellence of the Seattle rationality community, I expect it might have experienced more absolute loss from leaking members to Berkeley more than any other. Due to its size, the Seattle community has sustained itself, so the relative loss of local rationality communities which totally collapsed may be greater than has been the case in Seattle. As with Melbourne, if any community members who have lived or are living in Seattle wish to provide feedback, that is encouraged.

Vancouver. The experience in Vancouver has in the past certainly felt like ““if you don’t want to move to Berkeley as soon as possible, you are not *really* rational”. The biggest reason Vancouver may not have exuded as many rationalists to the Bay Area as cities in the United States is the difficulty being Canadian poses to gaining permanent residence in the United States and hence moving to the Bay Area. A couple friends of mine who were early attendees of a CFAR workshop lived in the Bay Area for several months in 2013, and returned home with stories of how wondrous the Bay Area was. They convinced several of us to attend CFAR workshops as well, and we too returned home with the sense of wonderment after our brief immersion in the Berkeley rationality community. But when my friends and I each returned, somehow our ambition transformed into depression. I tried rallying my friends to try carrying back or reigniting the spark that made the Berkeley rationalist community thrive, to really spread the rationalist project beyond the Bay Area.

But the apparent consensus was it just wasn’t possible. Maybe the rationality community a few years ago lacked the language to talk about it, but rationalists who’d lived in Berkeley for a time only to return felt the rationality-shaped hole in their heart could only be filled in the Berkeley. A malaise had fallen over the Vancouver rationality community. All of us were still around, but with a couple local EA organizations, many of us were drawn to that crowd. Those of us who weren’t were alienated from any personal connection to the rationality community. I saw in my friends a bunch of individual heroes who together were strangely less than and not greater than the sum of their parts.

Things have been better lately, and a friend remarked they’re certainly better than a few years ago, when everyone was depressed about the fact it was too difficult for us to all move to the Bay Area. In the last several months, the local rationality community has taken on as our mission our own development, and we’ve not rebounded so much as flourished like never before. But it took the sorts of conversations about the Berkeley rationalist community last year Zvi and others had to break the spell we had cast on ourselves, that apparently Berkeley had running a rationalist community like a well-oiled machine down to an art and a science.


II. The Berkeley Community and the Mission of Rationality

Benquo commented on Zvi’s post:

This is a good description of why I feel like I need to leave Berkeley whether or not there’s a community somewhere else to participate in. This thing is scary and I don’t want to be part of it.

I think this is some evidence that the Rationalist project was never or only very briefly real and almost immediately overrun by MOPs, and largely functions as a way for people to find mates. Maybe that’s OK in a lot of cases, but when your branding is centered around “no really, we are actually trying to do the thing, literally all we are about is not lying to ourselves and instead openly talking about the thing we’re trying to do, if you take things literally saving the world really literally is the most important thing and so of course you do it,” it’s pretty disappointing to find it’s just another flavor.

Since he wrote this comment, Benquo has actually continued to participate in the rationality community. This conversation was mired in tension in the rationality community it must have been difficult to think about impersonally, and so a charitable interpretation would be while these problems exist, Benquo and others are generally not as fatalistic about the rationality community as they were the time they wrote the comments. While I and others in thread saw grains of truth in Benquo’s statement, precision nonetheless remains a virtue of rationality, and I felt compelled to clarify. I commented:

I’d say the rationality community started whenever Eliezer forked off LessWrong Overcoming Bias, which was around 2008 or 2009. That’s certainly not when it peaked. Even in a way MIRI never was, CFAR started out a project built by the rationality community. That was happening in 2012 or 2013. Above Sarah is also quoted as saying she thinks the Berkeley rationality community hit the right balance of focusing on being a welcoming community qua community, and aspiring to the whatever the core mission(s) of the aspiring rationalist project are.

Unless you’re arguing there was a latency effect where the MOPs overran the community in 2009, but the consequences of such were buried for several years, the period between 200809 and 201213 doesn’t constitute being “immediately overrun”.

I get you’re pessimistic, but I think you’re overshooting. Matching the map to the territory of what went wrong in the Berkeley rationality community is key to undoing it, or making sure similar failures don’t occur in the future.
FWIW, I’m sorry you’ve had to experience so directly what you feel like is a decline in an aspect of your local rationality community. As someone who connects with rationalists primarily online, I can tell you they’re everywhere, and even if there isn’t a meatspace community as developed as the one in Berkeley, there are rationalists who won’t let the Craft disappear everywhere, and they want meatspace communities of their own built up outside of Berkeley as much as anyone.

Other comments in-thread from community members who had been around longer than Benquo or I confirmed my impression from their own personal experiences, so unless Benquo would further dispute these accounts, this thread seems put to rest. However, Zvi then replied to me:

I think we need to realize the extent to which Berkeley is actively preventing the formation of, and destroying, these other communities. The majority of high-level rationalists who started in the New York community are in the Berkeley community, which caused New York to outright collapse for years before recovering, and they just now once again caused a crisis by taking away a pair of vital community members and almost wiping out the only rationalist group space in the process. From meeting other community leaders in other cities, I hear similar stories A LOT.

I do agree that Plan A for most members can and should be Fix It, not walking away, and that pointing out it needs fixing is the requirement for perhaps fixing it.

To respond to Zvi here, indeed it appears to be an uncannily ubiquitous problem. I’ve collected a few stories and described them in some detail above. Between that and several comments from independent rationalists on Zvi’s original post giving the impression members of their local communities were being sucked to Berkeley as though through a pneumatic tube and leaving a vacuum of community and organization in its wake, it appears these many local stories could be a single global one.

The original mission of the rationality community was to raise the sanity waterline to ensure human values get carried to the stars, but we’re still godshatter, so doing so can and should take different forms than just ensuring superintelligence is aligned with human values. If ever the goal was to seed successful, stable rationalist communities outside Berkeley to coordinate projects beyond the Bay Area, it’s been two steps forward, one step back, at best. Even if we assume for the sake of argument it’s a good idea for rationalists worldwide to view Berkeley as a nucleus and their own rationalist communities as recruitment centres to drive promising individuals to Berkeley for the mission of AI alignment or whatever, the plan isn’t working super well. That’s because the apparent rate of local rationalist communities sending their highest-level rationalists Berkeley is occurring at a much faster rate than those rationalist communities can level up more rationalists to replenish their leadership and sustain the local community at all.

The state of affairs could be worse than it is now. But it creates the possibility that if enough local rationalist communities around the world outside the Bay Area simultaneously collapsed, the Berkeley rationalist community (BRC) could lose sufficient channels for recruitment to sustain itself. Given the tendency of communities like all things toward entropy, communities decay over time. The BRC could not be rubbing any of its members the wrong way and we would probably still observe some naturally occurring attrition. In a scenario where the decay rate of the BRC was greater than its rate of replenishment, which has historically largely depended on rationalists from outside communities, the BRC would start decaying. If we were to assume the BRC acts as a single agency, it’s in the BRC’s self-interest as the nucleus of the worldwide rationality movement to sustain communities-as-recruitment centres at least to the extent they can sustainably drive their highest-level rationalists to Berkeley over the long-term.

While this worst-case scenario could apply to any large-scale rationalist project, with regards to AI alignment, if the locus of control for the field falls out of the hands of the rationality community, someone else might notice and decide to pick up that slack. This could be a sufficiently bad outcome rationalists everywhere should pay more attention to decreasing the chances of it happening.

So whether a rationalist sees the outcome of the primary purpose of rationalist communities acting as a recruitment centres for the Berkeley rationalist community as an excellent plan or an awful failure mode, there’s a significant chance it’s unsustainable either way. It appears a high-risk strategy that’s far from foolproof, and as far as I know virtually nobody is consciously monitoring the situation to prevent further failure.


III. Effective Altruism and the Rationalist Community

In another thread, I responded directly to Zvi. I commented:

While rationalists are internally trying to figure out how there community has changed, and they’re lamenting how it’s not as focused on world-saving, there’s a giant factor nobody has talked about yet. The only community which is more focused on the rationality community’s way of world-saving than the rationality community is effective altruism. To what extent is the rationalist community less world-save-y than it used to be because the rationalists whose primary rationalist role was “world saver” just switched to EA as their primary world-saving identity. I think as things have gotten less focused since LessWrong 1.0 died, and the rationalist diaspora made entryism much easier as standards fell, what you’re saying is all true. You might be overestimating the impact of entryism, though, and underestimating people who exited not because they had no voice, but for sensible reasons. If at any point a rationalist felt they could better save the world within the EA rather than through the rationality community, it’d internally make sense to dedicate one’s time and energy to that community instead.

The EA community doesn’t seem able to build bonds as well as the rationality community. However, the EA community seems better at making progress on outward-facing goals. In that case, I for one wouldn’t blame anyone who find more at home as a world-saver in EA than they did in the rationalist community.

Zvi replied:

Definitely an elephant in the room and a reasonable suspect! Certainly partially responsible. I haven’t mentioned it yet, but that doesn’t mean I’ve missed that it is in the picture. I wanted to get this much out there now, and avoid trying to cover as many bases as possible all at once.

There have been many (Sarah [Constantin] and Benquo among them) who have been trying to talk for a long time, with many many words, about the problems with EA. I will consider that question beyond scope here, but rest assured I Have Thoughts.

Since then Zvi and others have made good on their intentions to point out said problems with effective altruism. I intend to engage these thoughts at length in the future, but suffice to say for now local rationalist communities outside the Bay Area appear to definitely have experienced being ‘eaten’ by EA worse than Berkeley.


I never bothered to tie up the loose ends I saw in the comments on Zvi’s post last year, but something recently spurred me to do so. From Benquo’s recent post ‘Humans need places’:

I am not arguing that it would merely be a nice thing for Bay Arean EAs and Rationalists to support projects like this; I am arguing that if you have supported recruiting more people into your community, it is morally obligatory to offer a corresponding level of support for taking care of them once you are in community with them. If you can’t afford to help take care of people, you can’t afford to recruit them.

If you don’t have enough for yourself, take care of that first. But if you have more than enough to take care of your private needs, and you are thinking of allocating your surplus to some combination of (a) people far away in space or time, and (b) recruiting others to do the same, I implore you, please first assess—even approximately—the correct share of resources devoted to direct impact, recruiting more people into your community, and taking care of the community’s needs, and give accordingly.
[...]
The Berkeley EA /​ Rationalist community stands between two alternatives:
1.Pull people in, use them up, and burn them out.
2. Building the local infrastructure to support its global ambitions, enabling sustainable commitments that replenish and improve the capacity of the people making them.

It’s important for rationalists in Berkeley to know that from where they’re standing, to rationalists around the world, these statements could ring hollow. The perception of the Centre for Effective Altruism slighting the Berkeley REACH is mirrored many times over in rationalists feeling like Berkeley pulled in, used up and burned out whole rationalist communities. The capital of a nation receives resources from everyone across the land. If the capital city recruits more citizens to the nation, is it not morally obligatory for the capital city offer a corresponding level of support for taking care of them once they joined your nation? Is it not the case if the rationality community can not afford to take care of our people, then we can’t afford to recruit them?

The worldwide rationalist project stands between two alternatives:

  1. Seed new local communities, use them up, and burn them out.

  2. Building the global infrastructure to support its global ambitions, enabling sustainable commitments that replenish and improve the capacity of the local communities making them.

This isn’t about the Berkeley rationalist community, but rationalist communities everywhere. In reading about the experiences of rationalists in Berkeley and elsewhere, I’ve learned their internal coordination problems are paralleled in rationalist communities everywhere. The good news in the bad news is if all rationalist communities face common problems, we can all benefit from working towards common solutions. So global coordination may not be as difficult as one might think. I wrote above the Vancouver rationality community has recently taken on as our mission our own development, and we’re not recovering from years of failures past so much as flourishing like never before. We haven’t solved all the problems a rationalist community might face, but we’ve been solving a lot. As a local community organizer, I developed tactics for doing so that if they worked in Vancouver, they should work for any rationalist community. And they worked in Vancouver. I think they’re some of the pieces of the puzzle of building global infrastructure to match the rationality community’s global ambitions. To lay that out will be the subject of my next post.