Comments
Epistemic Status: 15 years working as a therapist, 6+ of them with rationalist and EA clients. 7 years teaching and directing at rationality camps and workshops, as well as being a member of several of this community's overlapping subcultures.
Disclaimer: I finished my first draft of this article and its two subsequent ones months ago and sent them out for feedback to various people before a bunch of stuff happened in my life that derailed me from finishing, and in the meantime a big ol' brouhaha started around Slutcon 2 and various frictions this article was meant to highlight within the rationalist and EA community surfaced in The Discourse.
All of which is to say this post isn't about any particular incident, but the underlying dynamics that make many of them arise. When I make claims about how things go wrong and why, I'm drawing on a mix of research, observations, and both clinical and non-clinical experiences within these communities and beyond them.
TL;DR: It's helpful to think of the rationality/EA community as a village (social scene, professional network, dating pool, housing market, all tied together by shared culture and communal events), but it's also ultimately misleading.
We have never had the stabilizing infrastructure or monoculture that helps people navigate dual roles and boundary crossings that traditionally occur in real villages, and we've grown big enough to split into subcommunities with their own social norms and expectations that cause friction at the interfaces.
Also, traditional villages kind of suck in a number of ways we don’t want to copy, but we should understand the failure modes inherent to this fairly unique confluence of circumstances to try and navigate them better.
Say you’re at one of those infamous Berkeley house parties. Around the kitchen table are someone who might fund your org, a writer whose blog got you into all this, your coworker and ex, and someone who’ll have a say in whether you get a room in a group house you’ve been eyeing. Most know each other well, a few of them have dated each other, and at some point one might invite you to cuddle, or ask you to go for a walk, or offer you something to drink, and you might think you know what that means, because you've experienced or heard about the community plenty of times, and you have friends in it.
For a lot of people in the Rationalist/Effective Altruist/Postrat/EA-adjacent/etc community (hereafter referred to as some variation of "the community" or "the village"), that’s a fairly ordinary part of the social landscape. But that landscape wasn't one any of us grew up learning to navigate, or even heard about, because it's something that's actually quite new and unique... and it's been changing quite a lot over recent years.
This isn't an article for newcomers, nor is it specifically about Berkeley: London, Boston, Berlin, EAGs, Vibecamp, and other community hubs sit at different points on the spectrum of "how village-like is your social scene."
If this first article does its job well it'll help everyone have a better sense of what this big, beautiful community that's grown around us actually is and isn't, and help us navigate our differences better for all involved... though that part will mostly be covered in the later posts.
But first, why the "village" frame at all?
Quick check: what's your "village quotient" for the community?
Modern urban life, which shapes the default culture for most Americans and Europeans, just doesn’t have protocols for a social system where employers, neighbors, friends, partners and exes, therapists/lifecoaches/priests, and so on all work and live on the same few blocks.
The standard advice to “keep your professional and personal lives separate” comes from a very modern culture that never intentionally decided to build separation into day-to-day life. Instead people just started becoming aware of many kinds of drama that can come from mixed roles and developed heuristics against them, even though the drama mostly came from other factors, and threw out whole categories of basic social interaction along the way... at least, when they could get away with it. There’s a whole area of rural-practice literature for dual-role-sensitive professions like lawyers and police and therapists because dual relationships there are unavoidable.
But real villages, and the small bands we evolved in before forming them, don't have those separation, and managed to be very functional, because role crossing is not itself a problem; the problem is how role crossings are handled, and real villages have stabilizing infrastructure to navigate them.
Infrastructure our community largely lacks, namely:
Sure, our village has some version of some of these things. We have gossip networks, but lack common knowledge and shared agreement about how they should function, which makes them hard to predict or participate in as needed.
Small towns still run on a version of that sort of intangible infrastructure, and they mostly work because there are stabilizers in rough proportion to the entanglement: shared norms, long memories, low churn.
Our young, cobbled together village never had those. Instead we have weak and perpetually contested enforcement of norms that no one ever formally agreed on, constant churn of members both by self-labeling and geography (including an annual influx of very smart young people from a variety of backgrounds and cultures), and a dating pool and job market overlap to our friendgroups that virtually none of us were raised to know how to socially navigate.
And all this might work if our shared ideology was a very conservative, tradition-oriented one. Instead we share an ideology that drives us to poke at Chesterton Fences and decide individually whether they make sense for us, rather than respect them by default.
To be clear, I don't mean to romanticize traditional villages. A big part of their cohesion comes from much less savory aspects: oppressive conformity shaming, selective silencing of victims, implicit (or explicit) might-makes-right adjudications, and lack of optionality for those within them to leave. Things like “trusted elders” are a double-edged sword if those elders turn out to be bad actors themselves, or even just end up shielded from normal consequences due to their status (more on that in section 6).
What I mean to highlight is that we have village-level entanglement with weak forms of their “shock absorbers.” And to make things worse, we’re less and less clearly “one village” anymore.
Or maybe four, or six, or whatever. "What makes someone a rationalist" has no single answer, and the word is often used interchangeably to label an aspiration and as a community tag. Same with Effective Altruist, and all the various post-s, -adjacents, -ishes, and so on.
Personally, I think of an “aspiring rationalist” as someone who puts reasonable effort into falsifying their beliefs, engages with ideas on their merits, tries reasonably hard to stop making mistakes, treats someone asking "wait, is that actually true?" as a signal of being on the same side rather than an attack, and so on and so forth.
Meanwhile, I think of the “rationalist community” as largely made up of people who enjoy some of the same blogs and fiction, care about some of the same ways of thinking about things, and prioritize focusing their time and attention on a lot of the same sorts of things.
But no one made me the culture czar, and shared media, ideas, and values are not a shared set of social norms. By default, people in and around the community are more united by the sorts of things they like to think about much more than how they want to live their life.
The tricky part is that the media, the ideas, the values, all feel like they screen for more. Shared canon gives us shared vocabulary, shared jokes, shared heroes, and that fast, unmistakable click of “oh, these are my people.”
I don't want to knock the click, it's what has often made a room full of strangers in another country feel like home in minutes, for me. And for a lot of us it was the first time we ever felt something like it.
But the click arrives well before you get a basic sense of social norms, including but not limited to:
What the community is to you. For many people it's just the blogs and fiction, maybe an occasional meetup. For others it's their job, their house, their closest friends, and their dating pool all at once. And for some it's even more, a group of people that evoke deep, genuine feelings of care and kinship.
Conversational norms. For some people being argued with is being respected: bluntness is a gift, Crocker's Rules are aspirational, and no claim is off the table. For others, decoupling is not intuitive or obviously correct, and walking away from a group without saying goodbye feels strange.
What touch means. In many rationalist subcultures, particularly outside the Bay Area, it's exactly what it is in the wider society they're embedded in; either reserved for your romantic partners or a brief thing for affection between friends. In others, cuddles can be entirely platonic and terminal, as sexually charged as board games.
What romance looks like. Monogamists, relationship anarchists, and every flavor of polyamory from parallel to kitchen table, each with different background assumptions about what flirtation implies, what disclosure is owed, and what a "date" is a step toward.
And so on. Every combination of positions across these categories exists, and most of them are represented at large events.
Clusters form, sure, in endless forms most beautiful, and as the community has grown, those clusters have become a bunch of adjacent and often overlapping subvillages. But any classification is just a heuristic that’s right often enough to create blindspots, especially where subvillages meet. And of course some people live in the overlaps, as comfortable in the career-focused group houses as in the ones where people spend their time doing inner work and psychedelics, so you can’t even infer someone’s norms from knowing some of their friends.
Consider two hypothetical people, both behaving by the rules of their home subculture:
Casey comes from the platonic-cuddle culture. To Casey, joining a cuddle pile is like joining a conversation—warm, communal, like sharing a loveseat. Casey has cuddled with dozens of friends across years and it has never once escalated into something else without conversation.
Albert comes from a scene where physical affection is an escalation channel. In Albert's experience, sustained one-on-one physical closeness is how mutual interest gets communicated without anyone having to risk saying it out loud. Albert has learned, accurately, within that scene, that starting with cuddling, then casually escalating, is what works... and if the other person isn't interested, they'll say so, and no harm done.
Casey and Albert end up cuddling at a party. Casey experiences a nice, ordinary evening with a new friend. Albert experiences what is to him a clear signal of interest in more, and acts on it. Now Casey feels violated and Albert feels blindsided, despite each of them following their own culture's rules.
This particular instance can play out even if neither person is predatory or stupid or inebriated or naive or whatever! All it takes is running incompatible protocols through an interface that both people think they're familiar with. The same structure recurs everywhere.
Two things follow:
First, we need genuine grace at these interfaces, because a single crossing is not evidence of a bad actor by itself, and mistakes like this happen in every social scene, even those without a tangled mess of protocol confusions.
But second, individually innocent crossings can still aggregate into an unpleasant environment, even if they don't become violations.
We have a community where a young woman entering it may fall in with "her" crowd right away and get clear signals about social spaces or events with different norms, while her friend might encounter her fifth guy who didn't realize cuddling wasn't an escalation ladder within her first month, and both will be baffled by each other's experiences.
However innocent each individual misunderstanding, an atmosphere can be created in which she has to run threat-assessment at events that other women tell her are safe and wholesome. If your model of the community only operates at the level of individuals you know in it, you will likely keep being confused about why some people describe an environment that you, personally, have never once witnessed making anyone uncomfortable.
And all this gets complicated further when you take status into account.
Status is among the least measurable things we routinely hold people responsible for poorly tracking. You cannot check a mirror, or stick a thermometer in your mouth, or look up an account statement and discover "huh, I have more than I thought." You learn about it from interactions with others, but it's always context-bound, often transfers badly between communities or subcultures, and sometimes inverts entirely between them.
Imagine, for a moment, that you started a podcast years ago that a lot of people in the community enjoy. Maybe some people outside of it, but mostly not. Or maybe you wrote a sequence that got curated on Lesswrong, or you work at an org everyone in the community has heard of. In the rest of your life, nobody treats you as anything special: the people at your job don’t care, the barista does not recognize you, your parents probably don't read your substack. Even inside the community, you aren't a celebrity to yourself. You could easily name 20 people more famous in it, people you admire, events you wish you were invited to. Your self-model says, more or less, that you’re fairly ordinary.
But you would still have some fans, and those fans might encounter you at an event and treat you as someone high-status. Hell, those fans might feel you’re higher status than some of the people you yourself admire, or just assume you’re all roughly in the same orbit, and you would have no idea.
This isn’t a claim that the default for people is to be totally blind or confused to their social standing. Research on face-to-face groups (e.g., Anderson et al., 2006) finds that people track their status pretty accurately, and many err modest, since overclaiming status is one of the fastest ways to lose it.
But I expect that finding to apply very poorly to communities like ours. It holds where status evidence arrives continuously, face to face, but a large share of status in our community is parasocial. It lives in thousands of readers of blogs and tweets, or listeners of podcasts and videos, who regard someone from "offstage," and the evidence of that regard largely does not reach them until one of those readers walks up at an event.
A group-house leader who throws lots of parties, or an org director who speaks at a lot of events, gets constant face-to-face data about their standing; someone whose fame comes from the stuff they create gets almost none, until they attend an event where some portion of the attendees recognize them. So the writer's "ordinary person" self-model isn't motivated blindness, just a reasonable inference from genuinely missing data.
And most people's reference class naturally funnels upward. The podcaster calibrates to a few dozen people above them, who calibrate against the dozen above them, and so on, until you get a community where a surprising number of people have an interior mismatch between how much of a celebrity they are, and to whom. Even if you run events or give talks at conferences, many are open floor enough that you can easily feel like you're just one of the crowd even while, to most people, you stand out as more active and visible in the community than they are.
And admiration drives approach. The stranger who seeks you out after your talk and lingers in conversations is drawn disproportionately from the people for whom you were a name before you were a person. In this way, status differences actually generate interactions that are prone to being complicated by them.
The existence of subvillages makes all this even more complicated, because some people in the community have status in all of them, while others have status in just some. If you're used to being at Lighthaven for various events throughout the year, you might find yourself the center of attention at an ACX meetup in some random city even if your status in Berkeley is on the lower end. You could be practically invisible to most of the community, while being a micro-celebrity in the subvillage that you spend most of your time with.
Status is fairly intangible, but there are more concrete gradients between people (the slopes along which help or harm can flow more easily in one direction than the other) that often are associated with it, and that association is often what matters, whether true in any particular case or not.
In a community where some people are grantmakers and others are independently wealthy, money is definitely one. Employment is another, where your date today could easily become a professional reference next month. Housing, where the person pursuing you romantically might have a say in the group house you want to join next year. Even general access to the community itself, where a falling-out with someone can cost large chunks of your social world.
Not all of these are unique to village-like communities, of course. Normal friendgroups all over the world regularly involve messy dramas and falling-outs that sometimes have knock-on effects in their social, romantic, or even professional lives. But the likelihood is higher and the impact bigger the more interconnected everything is, which means it’s more predictable that social interactions and relationships might have extra complicating factors.
Alongside that predictability is the knowledge of what your relative position to someone else can mean to them. While you might genuinely not know how much social status someone else might perceive you to have, you cannot fail to know that you're their temporary host, or their potential employer, or twice their age. Roles are self-legible in a way status isn't, which I suspect is precisely why the traditional guardrails attach to things like roles, money, and age rather than to status.
That distinction, and the question of which end of a gradient you're standing on, is something I talk more about in the next post.
Consider a version of the Casey/Albert problem scaled up from a cuddle pile to the whole surface area around a subvillage.
Say people from one of the subvillages have gotten super into something new, or a new subvillage has started to form around their thing: it could be meditation, or psychedelics, or sex parties, or circling, or forecasting, or maybe it’s just day trading on stock markets.
They talk about it all the time, they publicly share how much they make on their trades, they insist that people are wasting their money and future if they have most of it in a savings account… and a lot of them do seem to be doing well. Maybe they start throwing lavish parties or retreats. They start funding stuff in the community. They become, in many ways, a bigger part of the community’s face, both internally and externally.
If it becomes loud enough, this has a number of effects.
None of these is inherently bad, but the subvillage’s successes and mistakes don’t stay inside it. Sometimes that’s literal, like funding parties or investment tips. Other times it’s reputation: people outside the community might start treating you or your org differently because of something a subvillage you’re only loosely connected to did.
More importantly, none of this is really avoidable. People will always want to follow their interests and find others who share them, and that will at some point hit a critical mass. It doesn’t even require anyone to start something new, since passive growth alone can push an existing interest over the threshold. If 10% of a 100 person community is into day trading, that’s ten people who mostly just talk to each other about it. In a community of 1,000 it’s a hundred, which is more than enough for a thriving group chat, regular meetups, and a shared set of norms and activities that might spill into broader community events.
And even if they give the standard disclaimers (Having brutally honest conversations with people while circling is emotionally risky! Don’t put money in stocks you’re not okay with losing!), that advice will obviously not stop everyone who, as it may turn out, are not a good fit for meditation retreats or psychedelics or orgies or whatever, or people who feel FOMO if they’re looking for more social connection, or are worried about feeling socially left behind.
And there doesn't exist a filter I know of that presorts people who are just a bad match for a particular version of the subvillage, instead of the one that they might be used to or have imagined. The usual interpersonal dramas that happen around book clubs or kink scenes are all still going to happen, and often times blame will be put on the distinguishing difference as if base rates don't matter.
To be clear, this is also not always a status thing! Making lots of money is a fairly straightforward desire that most people would be happier to be doing. Being seen as sexy is a bit more complicated for some, but is also mostly desirable, particularly for men. Even just seeing people form deep emotional connections around them can naturally make someone feel like they’re missing out if they don’t try whatever is causing it.
But I want to stress that even if it’s just a status thing, or things downstream of status, those are normal, understandable emotions for people to have. Emotions Make Sense. Not feeling strong emotions that lead to status seeking can be advantageous in many ways, but also often come with downsides.
So we might easily have a situation as follows:
Ravi is on a two-year grant with no runway, from Subvillage B, where most people work in nonprofits. Dana is from Subvillage C, where talking about money and encouraging people to invest is how people look out for each other, and the norms against picking stocks strikes Dana as exactly the kind of inherited fence worth tearing down.
Ravi sees that several friends are better off for having listened to Dana, and starts to reconsider standard advice about financial risk tolerance. Dana thinks she’s been helpful by pointing out a number of investments he should be holding if he takes his beliefs seriously; Ravi walks away feeling that the thing separating him from the people throwing the parties is that he’s not brave enough. So maybe he does something financially risky, something he wouldn’t normally do if he wasn’t in regular contact with people like Dana. If it works out, great. If not...
Different perspectives would place blame on different people here. And blame can, sometimes, be a useful framework. But only if centered in a shared understanding of what went wrong, why it went wrong, and who's responsible for what.
What I want to emphasize here is that the above scenario is both a collective challenge for the overarching village, and for individuals in it.
One of the stabilizers on the earlier list was rites of passage, specifically the one for marking "adulthood." Meanwhile, one of the hallmarks of the modern, Western urban lifestyle has been the falling away and watering down of not just rituals, but actual signifiers, for what marks “adulthood.”
Adulthood is fundamentally about responsibility. We are considered an adult when we are deemed old enough to be responsible for our own actions; the younger we are, the less society will punish us for our mistakes, and the more they will look to our guardians to explain our behavior. The older we get, right up until a relatively arbitrary age cutoff, the more people expect from us, and the more they expect us to take responsibility for our actions.
This essay is already too long to get into some of the many ways that rural and urban communities treat their children, and how this creates very different “maturity tracks.” Obviously this is not a strict rural and urban divide, and many other factors like socioeconomics and family culture and so on play a big part too.
But traditional villages had rites of passage that marked the point where you were expected to know your own mind and answer for your choices. Not only does our village not have this, our society has been fairly confused about it too, on a number of issues. So all the people arriving in the scene do so at very different points along that track, and with different background assumptions. Some have spent years figuring out what they want, who they are, and more importantly who they aren’t, while others are figuring it out in real time at parties.
Our community, more or less, wears the label of rationalist. We call our forum LessWrong. We praise being agentic, yes, but only with serious attempts to predict second order effects and avoid unnecessary risks. We praise people betting on beliefs in part because it builds in a natural responsibility for our potential mistakes.
And we admire people who are willing to say reasonable but unpopular things, even at potential personal cost, because we know that just going along with the consensus view of the crowd risks creating an echo chamber.
When you take these all together, it becomes clear that an important part of the actual practice of rationality includes not just putting effort into falsifying your beliefs, but holding your beliefs against social proof.
And this includes beliefs about yourself, such as decoupling "everyone treats X as good" from "X is good for me.”
Reasoning through what is true or false will almost always occur while embedded in social systems. Our own wants and preferences are the one thing we have privileged access to, which makes them the cheapest place to practice. Holding a preference against social proof is the easier version of the move we need to be able to make for all sorts of other beliefs and questions where we don't have a privileged inside view to use.
This is why I stress “knowing what you want” as one of the most fundamental parts of being a rational agent, or having healthy relationships, or whether I’m talking about executive function or whatever. You cannot optimize your actions for your preferences and values if you do not know what they are, or cannot hold to them.
And in a collection of subvillages, this can be quite hard to consistently do well.
A lot of that stuff requires time and effort, and young people in particular are less knowledgeable about their wants and preferences, or how to even figure that out. Experimentation is important for figuring that stuff out!
But experiments can be expensive, and it makes sense that not everyone has the same amount of time to try things, or knowledge on how to experiment safely. Also, the village environment is going to incentivize certain experiments more than others, usually the ones the most visible and high-status subvillages are running, and we should all be aware of the failure modes that come with those experiments.
Not to put too fine a point on this, but many more people play golf than enjoy golf, because their social and professional circles reward them for it. There are very few activities we should expect to be immune to this effect, and I personally know of people who have admitted to going to things because they wanted to be part of the social scene, or even connected ones, rather than because they fit the target audience, and they’re generally not going to admit that to the people at the center of it.
A big part of the value of being in a rationalist village is that if you surround yourself with people putting the effort in to think more clearly and reduce negative actions, even just sometimes, you can pick some up purely from osmosis. But at the end of the day, reading the right blogs and surrounding yourself with people who properly decouple “that person is sexy” from “that person is correct” is only a protective buffer; it isn’t the mental motion itself, within your own mind, of reminding yourself to separate “people like that person for X” and “I should do X.”
Social pressure isn’t often an argument you can examine. It’s a feeling; that everyone else is more relaxed about this than us, that our hesitation is a limitation rather than a preference, that we'd be more at home here if we did what the others do, etc.
Feeling FOMO at a party where everyone seems to be doing something we are not is data about how we relate across an interface between our home-culture’s norms and the one we’re observing. So social pressures in villages are real, but they don't reach in and set our preferences. What they do sometimes is introduce inner conflict when a thing we want (money, sex, prestige) seems gated behind a process we don’t enjoy or trust.
But we can still examine whether the want is actually ours or not.
We can examine whether the feelings of worry, shame, or disgust are actually ours, or not.
We can explore our boundaries while holding them as boundaries, or not.
But if we try to perform one particular subvillage’s status markers, even if it’s the most visible or most prestigious subvillage (for the moment, anyway), the identity of the ur-village itself and what it was built on is going to be more difficult to navigate.
It would be convenient to believe boundary problems are something unusual communities import along with their unusual norms. The evidence points the other way: they are the default output of ordinary community structure.
The workplace harassment literature is the most-studied case. Meta-analytic work (Willness, Steel & Lee, 2007) finds the strongest predictor of harassment incidence is organizational climate: whether complaints are perceived to carry risk for the complainant and consequences for the perpetrator. The 2018 National Academies report on harassment in academic science reached the same conclusion.
Various professional communities learned the ubiquity of this in their own way. The clinical ethics literature on dual relationships (Schank & Skovholt, 1997) wanted to examine whether role-mixing could simply be prohibited, and rural and small-community practice forced the field to admit it can't be. The field's mature position is that "crossings" are constant and unavoidable, and could even be beneficial. The ethical work lies in preventing crossings from becoming "violations,” which are harmful. (Gutheil & Gabbard, 1993)
Also, all communities seem to subconsciously protect their valued members by default. Status buys what Hollander (1958) called idiosyncrasy credits: accumulated deviation allowance that lets established members break norms newcomers would be sanctioned for. Mission-driven institutions, whether political or non-profit or religious, compound this by treating accusations against important members as attacks on the mission (Smith & Freyd's work on institutional betrayal documents the pattern and its costs to victims).
Nobody decides on all this. It seems emergent from various fundamental parts of group psychology, and can be seen in all walks of life, from churches to kink communities, political coalitions to school boards. And having explicit, signposted rules is definitely not a silver bullet: the NCSF consent-violations survey in kink communities shows that explicit norms coexist with substantial violation rates.
All of this matters insofar as it helps us understand the shape of the problems we might struggle with. We are in many ways a special case, but only in how many different situations and contexts that can be found elsewhere have been stitched together. We have a number of completely ordinary problems running on several aggravating conditions at once.
I have a lot of thoughts on how to navigate all this better, but to keep this post from becoming a novella I’ll save most of it for the next posts. But I want to highlight that I think this is a genuinely hard problem, in the realm of Molochian ones, because it’s actually multiple hard problems stacked together and exacerbating each other.
There is no playbook here, and I’m skeptical of anyone who says there are obvious solutions; usually they’re underestimating the scope, or taking for granted that the tradeoffs are worth it for everyone, or easy to implement without stifling or killing something about the community that others find quite rare and valuable.
For better or worse, there's no culture czar or village elder to make such decisions, and as usual, we're going to have to figure this out as we go.
Meanwhile, there’s things I want people to keep in mind:
The next post is about what each of us owes from wherever we happen to be standing, which I frame as “grace across interfaces and lines on gradients.” The one after that is about the stuff individuals can’t really fix on their own: shared spaces, starting new subvillages, what to do when something ambiguous happens to you, and how individual experiences can add up to patterns someone can act on.
Hopefully nothing else in the community explodes in the meantime.
