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

it's a general issue with the way CFAR operates, building up intense social connections over the course of a weekend, then dropping them suddenly.

So, this is definitely a thing that happens, and I'm aware of and sad about it, but it's worth pointing out that this is a generic property of all sufficiently good workshops and things like workshops (e.g. summer camps) everywhere (the ones that aren't sufficiently good don't build the intense social connections in the first place), and to the extent that it's a problem CFAR runs into, 1) I think it's a little unfair to characterize it as the result of something CFAR is particularly doing that other similar organizations aren't doing, and 2) as far as I know nobody else knows what to do about this either.

Or are you suggesting that the workshops shouldn't be trying to build intense social connections?

Is this suggesting that top-tier Berkeley is even eating the seed corn of Berkeley and making everyone but its own top-tier depressed in its wake?

This was the experience in Vancouver after CFAR workshops, and the atmosphere persisted for a long time. It wasn't only me who was conflating "[big event] atmosphere" with "Berkeley Rationalist Community". Not just me but a lot of other people in Vancouver, and also how a lot of rationalists from elsewhere talk about the Berkeley Rationalist Community (I'm going to call it the Bayesian Area), it's often depicted as super awesome.

The first thing that comes to mind is a lot of rationalists from outside of Berkeley only visit town for events like CFAR workshops, CFAR alumni reunions, EA Global, Burning Man, etc. So if one rationalist visits Berkeley a few times a year and always returns to their home base talking about their experiences in Berkeley right after these exciting events, it makes the Berkeley community itself seem constantly exciting. I'm guessing the reality is Berkeley community isn't always buzzing with conferences and workshops, and organizing all those things is actually very stressful.

There definitely is a halo around the Berkeley Rationalist Community for other reasons:

• It's often touted 'leveling up' to the point one can get hired at an x-risk reduction organization or working on another important project like a startup in Berkeley is an important and desirable thing for rationalists to do.
• There's often a perception resources are only invested in projects based in the Bay Area, so trying to start projects with rationalists elsewhere and expect to sustain them long-term is futile.
• Moving to Berkeley is still inaccessible or impractical for a lot of rationalists scattered everywhere that (especially if their friends leave) it breeds a sense of alienation and being left behind/stranded as one watches everyone else talk about how they *can* flock to the Berkeley. Combined with the rest of the above, this can also unfortunately breed feelings of resentment.
• Rationalists from outside Berkeley often report feeling as though the benefits or incentives to moving to the Berkeley community are exaggerated relatives to the trade-offs or costs of moving to Berkeley.

It would not surprise me if this halo effect around the Berkeley rationalist community around the world is just a case of confirmation bias writ large among rationalists everywhere. It could be there is a sense the Bayesian Area is doing all this deliberately, when almost no rationalists in Berkeley intended to do this. The accounts of what has happened to the NYC community are pretty startling, especially as one of the healthier communities I thought it would persist. The most I can say is there is a wide variance in accounts of how much a local rationalist community feels or not pressure exerted from Berkeley to send as many people as possible their way.

But then I also read stuff like this post by Alyssa, who is from the Berkeley rationalist community, and Zvi's comment about Berkeley itself eating the seed corn of Berkeley sounds plausible. Sarah C also wrote this post about how the Bayesian Area has changed over the years. The posts are quite different but the theme of both is the Bayesian Area in reality defies many rationalists' expectations of what the community is or should be about.

Another thing is much of the recruitment is driven by efforts which are decidedly more 'effective altruist' than they are 'rationalist'. With the Open Philanthropy Project and the effective altruism movement enabling the growth of so many community projects based in the Bay Area, it both i) draws people from outside Bay Area; ii) draws attention to the sorts of projects EA incentivizes at the expense of focusing on other rationalist projects in Berkeley. As far as I can tell, much of the rationality community who don't consider themselves effective altruists aren't happy EA eats up such a huge part of the community's time, attention and money. As far as I can tell, it's not that they don't like EA. The major complaint is projects in the community with the EA stamp of approval are magically more deserving of support than other rationalist projects, regardless of arguments weighing the projects against each other.

To me a funny thing is from the other side I'm aware of a lot of effective altruists long focused on global poverty alleviation or other causes are unhappy with a disproportionate diversion of time, attention, money, and talent toward AI alignment, but moreover EA movement-building and other meta-level activities. Both rationalists and effective altruists find projects also receive funding on the basis of fitting frameworks which are ultimately too narrow and limited to account for all the best projects (e.g., the Important/Neglected/Tractable framework). So it appears the most prioritized projects in effective altruism are driving rapid changes that the grassroots elements of both the rationality and EA movements aren't able to adapt to. A lot of effective altruists and rationalists from outside the Bay Area perceive it as a monolith eating their communities, and a lot of rationalists in Berkeley see the same happening to local friends whose attention used to not be so singularly focused on EA.

[Brainstorming]

One idea is to try to differentiate the NYC 'product' from the Berkeley 'product'. For example, the advantage of Vancouver over the Bay Area is that you can live in Vancouver if you're Canadian. The kernel project attempted to differentiate itself through e.g. a manifesto. In the same way, you could try to create an identity that contrasts with the Bay Area's somehow (for example, figure out the top complaints people have about the Bay Area, then figure out which ones you are best positioned to solve--what keeps you in NYC?) Academic departments at different universities are known for different things; I could imagine a world where rationalist communities in different cities are known for different things too.

I'm sorry that happens to you.

That's interesting. I would expect that New York would be a large enough city that it should be possible to build up a strong community there.

CEA turned down my proposal because there were other, more established groups than REACH with clearer track records of success and better thought out metrics for success/failure applying for the same round of grants. I am working on building up a track record and metrics/data capture so that I can reapply later.

I had difficulty finding a word to get across what I meant, so I went with 'slighted' but, I didn't think that's what you meant. My interpretation of your original post was not that you think CEA owes Berkeley REACH in particular, but you think CEA ought to be the kind of organization more willing to consider community projects like the Berkeley REACH. Of course it's apparent now that isn't what you meant either. Thanks for clarifying.

One point I was getting across is there is a perception among the rationality community not the Berkeley rationality community itself, but maybe leaders or key organizations there should play the role for rationalist communities around the world you're hoping Berkeley rationalists play as patrons of the Berkeley REACH, as the Berkeley has received so much from the rest of the rationality community. Whether this sense of owed reciprocity is fair is another question entirely (I personally don't think it's the right question to be asking if we want to find solutions to problems the community faces. I'm still working on my thoughts on that though).

Thank you for making this comment. Of all the reactions to this post, this one best captures how I want rationalists outside the Bay Area to relate to it going forward. Of course it doesn't go as far as I'd like, but I'm unsure of how I want to take it. I've been reading some of Zvi's posts from last year, which are wrongly pessimistic not because they're a self-fulfilling prophecy preventing non-Berkeley rationality communities from achieving their values, but because it's a map of how rationality communities develop not matching the territory. (I'm aware things were more tense between NYC and Berkeley a year ago, and while I don't know all the details, I imagine Zvi had sufficient reason for how he felt, and may not endorse as strongly now everything he said then.) At the same time, not regarding inter-community dynamics, but the whole rationality movement, I feel like the Community has failed to uphold the Craft. This isn't the same as not devoting enough resources or doing so in the right way toward AI alignment or another mission. It's about the sense I got from reading posts like this one from last year, and my sense other rationality communities are like Berkeley now: rationalists have an aversion to the changes trying to level up might bring to their communities because it would disturb the local state of affairs too much.

In Vancouver, we never blamed the Bay Area for our woes. I think it partially induced our woes, but I don't think from the scale of the individual to the whole Berkeley rationality community, or any subset in between should be blamed for what's happened. We depressed ourselves with how inadequate we seemed relative to Berkeley, and to the extent the Berkeley rationality community perpetuates that mindset, they're preventing the expansion of the Craft. That nobody in Berkeley talks about that, and barely anybody who complains about Berkeley mentions this, leads me to think it's a huge blind spot for all of us. In the past there have been attitudes toward other rationality communities from Berkeley rationalists I've found upsetting because of the harm I think they cause, but to act as though there is an agency at the heart of the Berkeley rationality community conspiring to plunder others is pointless.

If rationality communities outside Berkeley are tired of losing people because it prevents them from launching projects which would advance the rationality happening in Berkeley, I want to help solve that problem. If rationality communities outside Berkeley are tired of losing people to Berkeley because they feel like there should be a more equitable distribution of bonobo rationalist cuddle puddles between Berkeley and everywhere else, that's not a problem I'm interested in putting much effort to solve. I've got nothing against bonobo rationalists, and if I had a magic button which would optimize cuddle puddles everywhere for any rationalist who'd want them, I'd press it.

I think different concerns about the relationship between Berkeley and other rationality communities are all bundled up, it's hard to tease them out, and so for now I'm treating them as though they're all equally valid. But to the extent complaints are less of the form "we're tired of Berkeley taking up too much of the bandwidth of the awesome projects in the worldwide community" and more "we demand a redistribution of warm fuzzies", the less they move me. Local communities need to give individual rationalists reasons to stay, or reasons to come, and they need to take more self-responsibility for that. That's the attitude I tried bringing to turning things around in Vancouver, it inspired others, and we got a lot done. It sounds like some of what NYC experienced is an outlier in that regard, but based on the other comments here it's my impression local rationality communities solving the problem of getting hollowed out might be getting organized. My experience has been it's 99% perspiration, 1% inspiration. Based on my experience in Vancouver and accounts from elsewhere I'm guessing we can identify some heuristic 'best practices' for rationality community organization/development. So if rationalists, wherever they are, are willing to perspire, I'm confident other, more developed rationality communities can provide them with the tools they need. But to get people to the point they're willing to perspire, I (we?) have to get that 1% inspiration right. And how to do that is what I'm stuck on now.

I agree with all of this, except for existential risk reduction and other potential goals of the rationality community don't fit with waiting rationality to coalesce into a world religion, which you've already acknowledged. Also, I feel like just because it's the rationality community we should find a way to create tighter feedback loops of coalescing into worldwide community in a shorter period than religions typically take. Personally I'm more motivated by the Craft than the Community, but I figure to rally the whole community both are necessary (and interdependent?), so I'm still trying to hack together a way to balance both while accelerating sustainable development of local rationality communities.

A healthy topology of the field should have approximately power-law distribution of hub sizes. This should be true also for related research fields we are trying to advance, like AI alignment or x-risk. If the structure is very far from that (e.g. one or two very big hubs, than nothing, than a lot of two orders of magnitude smaller groups fighting for mere existence), the movement should try to re-balance, supporting growth of medium-tier hubs.

Although my understanding of network science is abecedarian, I'm unsure of both whether this feature is diagnostic (i.e. divergence from power-law distributions should be a warning sign) or whether we in fact observe overdispersion even relative to a power law. The latter first.

1) 'One or two big hubs, then lots of very small groups' is close to what a power law distribution should look like. If anything, it's plausible the current topology doesn't look power-lawy enough. The EA community overlaps with the rationalist community, and it has somewhat better data on topology: If anything the hub sizes of the EA community are pretty even. This also agrees with my impression: although the bay area can be identified as the biggest EA hub, there are similar or at least middle sized hubs elsewhere (Oxford, Cambridge (UK), London, Seattle, Berlin, Geneva, etc. etc.) If we really thought a power law topology was desirable, there's a plausible case to push for centralisation.

The closest I could find to a 'rationalist survey' was the SSC survey, which again has a pretty 'full middle', and not one or two groups ascendant. That said, I'd probably defer to others impressions here as I'm not really a rationalist and most of the rationalist online activity I see does originate from the bay. But even if so, this impression wouldn't worry us if we wanted to see a power law here.

2) My understanding is there are a few generators of power law distributions. One is increasing returns to scale (e.g. cities being more attractive to live in the larger they are, ceteris paribus), another is imperfect substitution (why listen to an okay pianist when I can have a recording of the world's best?), a third could be positive feedback loops or Matthew effects (maybe 'getting lucky' with a breakout single increases my chance of getting noticed again, even when controlling for musical ability versus the hitless).

There are others, but many of these generators are neutral, and some should be welcomed. If there's increasing marginal returns to rationalist density, inward migration to central hubs seems desirable. Certain 'jobs' seem to have this property: a technical AI researcher in (say) Japan probably can have greater EV working in an existing group (most of which are in the bay) rather than trying to seed a new AI safety group in Japan. Ditto if the best people in a smaller hub migrate to contribute to a larger one (although emotions run high, I don't think calling this 'raiding' is helpful - the people who migrate have agency).

[3) My hunch is what might be going on is that the 'returns' are sigmoid, and so are diminishing with new entrants to the Bay Area. 'Jobs'-wise, it is not clear the Bay Area is the best place to go if you aren't going to work on AI research, and even if so this is a skill set that is rare in absolute terms amongst rationalists). Social-wise, there's limited interaction bandwidth, especially among higher status folks, and so the typical rationalist who goes to the bay won't get the upside the most desirable bits of bay area social interactions - when weighed across from the transaction costs, staying put and fostering another hub might look better.]

(I echo Chris's exhortation)

I don't suppose I could persuade you to write up a post with what you consider to be some of the most important insights from network theory? I've started to think myself that some of our models that we tend to use within the rationality community are overly simplistic.

I don't think I've ever seen an event with more people interested than able to attend in Berkeley. If anything, it's difficult to get people to come out for events.

The money in the Bay uses 'if you're not in the Bay you're not serious, and even if you are other Bay money won't take you seriously so I can't afford to' as a coercive strategy to draw people there. Parallel with the community issues. Giving in to such tactics makes the problem that much worse and it snowballs.

Yes, Bay tech money is bigger and more our flavor there, but there's lots in many other places, and we'd get more out of what money exists if we were spread out than if we all chased the biggest pile, even with that pile playing hostile negative-sum games on us.

1) If you think tech money is important, you need to be in the bay area. Just accept that. There's money elsewhere, but not with the same concentration and openness.

This is true. There are reasons other than community-building to not be concentrated in one place. I don't think trying to reverse the relatively high concentration of rationalists in the Bay Area is at this time a solution to common community problems.

2) Are you focused on saving the world, or on building communitiy/ies who are satisfied with their identity as world-savers? "bring them in, use them up" _may_ be the way to get the most value from volunteer sacrifices. It may not - I haven't seen a growth plan for any org that explicitly has many orders of magnitudes of increase while still being an infinitesimal fraction of the end-goal.

This strikes me as pretty unlikely. Often even moreso among EA organizations than ones in the rationality community, world-saving operations which try this strategy appear to have a higher turnover rate, and they don't appear to have improved enough to compensate for that. The Centre for Effective Altruism and the Open Philanthropy Project are two organizations which have close ties and are the two biggest funders in effective altruism, which also covers x-risk/world-saving rationalist projects. They're taking more of a precision approach building community/ties in a way they think will maximize the world-saving-ness of the community. Not everyone agrees with the strategy (see this thread [LW(p) · GW(p)]), but it's definitely more of a hands-on approach moving away from a "bring them in, use them up" model that was closer to what EA organizations tended to do a few years ago.

Many of the other comments on this post point to an issue of concern being a trade-off between a world-saving focus and rationality community-building, but my sense of why it is tense is because both are considered important, so the way is to find better ways to not lose community-building to world-saving.

One pattern I'm noticing is because of the fact that because of the relative comparative advantage of citizenship in other countries, and the relative difficulty of attaining permanent residency in the United States, the communities of rationalists abroad are more stable over time because of the practical difficulty of convincing people to move to the United States. For example, having post-secondary education that is more subsidized not just in undergrad but in graduate studies as well in countries aside from the United States keeps non-American rationalists in their home countries until their mid-to-late twenties. That's young enough I know rationalists who musing about moving to Berkeley to work on AI alignment or another community project someday, but I also know a lot of rationalists who have set down roots where they are by then, and aren't inclined to move. Another thing if is a rationalist doesn't have a university degree or highly in-demand skills (e.g., STEM) for big corporations, it's difficult enough to get health insurance and visas for a lot of rationalists emigrating to the United States it doesn't make sense to try. This first post I wrote is intended to be part of a sequence to be focused on finding solutions to problems apparently common to community organization both in Berkeley and elsewhere. I tried to end this post on a positive tone, intending to build up to optimism in the next one, with marked examples of recent success among rationalists around the world. Vancouver has a couple local EA organizations, and strategies we've implemented locally have dramatically increased our rate of meetups, doubled the number of rationalist houses (from 2 to 4, but in 6 months that is still significant. The same has happened in Motreal the last few months. Jan Kulveit is another rationalist who has commented on this post as well who has had reported a lot of success with local community organization in the Czech Republic, as has Toon Alfrink from the Netherlands. If we can integrate what worked for us into a single strategy for mobilizing resources in local rationality communities it could be excellent.

The good news is I think the possible global failure mode I pointed out of the rationality community being too heavily concentrated in a single geographic hub which may then collapse appears quite unlikely to come about for the foreseeable future. So while the experience of the NYC rationalist community may be similar to a lot of rationality communities, it's not universal. I don't know if that means much given the NYC community has lost so many people, but hopefully if something comes out of people sharing solutions we can find a way to help the NYC community as well.

So its hard to tell people to refrain from moving to Berkeley

I apologize for possibly/probably twisting your words a bit here, but I never have trouble telling people to refrain from moving to the Bay/Berkeley. I tell them I lived there for a few years and it’s a pretty unpleasant place, objectively, along any of ten different metrics relevant to comfort and peace of mind. I tell them I never actually developed any sense of belonging with the local Rationalist Community, so it’s not gauranteed that that will happen. I tell them I make a pretty good amount of money in many cities, but since I’m not a Comp Sci grad that doesn’t translate to a decent living in Berkeley. I tell them on top of that, Berkeley is one of the most expensive places to live in the world and if there were some kind of objective ratio of cost of living divided by objective comfort/quality/value-of-a-dollar, Berkeley would be near the top worldwide.

I also don’t find the proposition that you have to literally move to an expensive unpleasant overcrowded dystopian city in order to be rational to be particularly, uh, rational.

Thanks for your response.

I am sort of agnostic about whether the Berkeley community is a good idea or not. On one hand it certainly feels pointless to try to build up any non-Berkeley community. If someone is a committed rationalist they are pretty likely to move to Berkeley in the near future. In addiiton it is very hard to constantly lose friends.

So there is the 'Craft' and the 'Community', or at least that is sometimes how rationality is modeled. And the Community could be broken down into Berkeley and other communities elsewhere. But if rationality is also about a mission to ensure human values are carried to the stars, and right now that hinges on AI alignment, it makes sense to me the rationality community is significantly concentrated in the Bay Area. This or other mindsets of singular focus in the name of the Craft appear they might come at some expense to the Community in Berkeley as well. The last year has seen some people in the Berkeley community ask if the Berkeley community is good for the community. I think this might be part of a worldwide problem in rationality, which I only have half an idea of how to tackle. I might need to get a lot of thoughts down before I figure out where I'm going with them.

This post probably best captures the emotional reality:

"I have lost motivation to put any effort into preserving the local community – my friends have moved away and left me behind – new members are about a decade younger than myself, and I have no desire to be a ‘den mother’ to nubes who will just move to Berkley if they actually develop agency… I worry that I have wasted the last decade of my life putting emotional effort into relationships that I have been unable to keep and I would have been better off finding other communities that are not so prone to having its members disappear."

If you base your social life around the rationality community, and do not live in Berkeley, you are in for alot of heart ache. For this reason I cannot really recommend people invest too heavily in the rationalist unless they want to move to Berkeley.

There are stories of mixed success throughout the community in building a rationalist community outside of Berkeley, and it gives me some hope, but then I read about these experiences and I feel ambivalent. I'm afraid anything other local rationality community organizers might recommend is something NYC or another once-flourishing rationalist community has already tried before, and it didn't work. I'm also afraid if a new community takes advice on how to build up while retaining membership over time that worked somewhere else, and then fails, it will greatly discourage the new community someone tried launching. Ultimately I consider the struggles the community faces as hard optimization problems, and right now I'm holding off on proposing solutions until I've discussed the solution more.

On the other hand concentration has benefits. Living close to your friends has huge social benefits. As sarah says very few adults live on a street with their friends and many Berkeley rationalists do. Its looks likely there will be rationalist group parenting/unschooling. The Berkeley REACH looks awesome (I am a patron despite living on the other side of the country). The question is whether the Berkeley community is worth the severe toll it places on other rationalist communities. In the past i thought Berkeley had some pretty severe social problems. Alot of people (who were neither unusually well connected or high status) who moved their reported surprising levels of social isolation. However things seem to have improved a ton. There are now a ton of group houses near each other and the online community (discord/tumblr) is pretty inclusive and lets 'not high status' people make connections pretty easily.

Also arguably 'Moloch already won'. So its hard to tell people to refrain from moving to Berkeley.

Ideally we would find ways to create similar outcomes for rationalists in lots of different places, which I see as a hard optimization problem I'm holding off on proposing solutions to until I've looked at it from more angles.

Also, the story is basically: for a while there was a LessWrong meetup, but then this got dropped and transformed into an EA Meetup. Then there were only EA meetups for a while. Then I started RRG and brought rationality back as its own hub, creating the Seattle Rationality FB group as well. The rationality community grew. Now there are multiple rationalist group houses including a new hub. People did leave for Berkeley, but weekly RRG is still going afaik, and there is still an active community, although its composition is perhaps quite different now.

I've edited my post. Thanks for clarifying.

I'm working on building up a similar reproducible set of operating guidelines for REACH and would be very interested in comparing notes.

It's taking me longer than get everything written down, but I explained the tactics that have generated the most value for the Vancouver rationality community in the last 6 months in this reply to stardust's comment [LW(p) · GW(p)].

Creating some kind of volunteer organization like that is an end-goal I have in mind, and I've started talking to other people about this project. I've volunteered and been friends for a long time with a local EA organization, Rethink Charity, which runs the Local Effective Altruism Network (LEAN), which does exactly that for EA: advising newly started groups, producing materials the groups can use, and innovating ways to help groups get organized. So as part of a volunteer organization I could get advice from them on how to optimize it for the rationality community.

what would be wrong with this?

Conceivably a community other than rationality steering the trajectory of AI alignment as a field might increase existential risk directly if they were abysmal at it, or counterfactually increase x-risk relative to what would be achieved by the rationality community. By 'rationality community', I also mean organizations that were started from within the rationality community or have significantly benefited from it, such as CFAR, MIRI, BERI, FLI and BERI. So my statement is based on 2 assumptions:

1. AI alignment is a crucial component of x-risk reduction, which is in turn a worthwhile endeavour.

2. The rationality community as a coalition, including the listed organizations, form a coalition which has the best track record of advancing AI alignment with epistemic hygiene relative to any other, and so on priors the loss of relative influence on AI alignment by the rationality community to other agencies would decrease x-risk less than it otherwise would be.

If someone doesn't share those assumptions, my statement doesn't apply.