Better models for EA development: a network of communities, not a global community
[EDIT: I recommend to first read this comment which clarifies the raison d’être of this post. The post should make more sense afterwards.]
Whenever this post mentions ‘we’, it refers to EA Geneva.
This article aims to contribute to developing a better understanding of ‘community building’. It is the result of many discussions indicating that some considerations regarding terminology, professionalization, and group dynamics have not yet been sufficiently talked about or documented. That seems in part due to (i) a lack of people with time to think about EA network strategy and (ii) the fact that the people who do think about it exchange mostly informally.
We hope to contribute to the model-building process by presenting some of our current thoughts—not as conclusions but to nourish discussion.
Communities are groups of humans with a shared identity who care for each other.
The EA network is a set of organizations, communities and individuals who share a mission but cannot all care for one another.
Building real communities is done through psychological safety and common growth, growth is facilitated by shared work.
Developing the network is done by aligning incentives and professional coordination.
We try to disentangle first responsibilities and goals to improve network development and clarify what community building entails.
Understanding and defining ‘community’
In his 2018 EA Global opening speech, William MacAskill compares the Spartans to the Athenians. Spartans get stuff done through centralised enforcement of conformity. Athenians hope for the best argument to win through the open-ended celebration of argument. A neat set-up to then define EA through its methods, values, norms and culture.
In his talk, MacAskill refers to those who identify with the definition of EA as the EA ‘community’. But this broad understanding of ‘community’ is causing confusion because much of what is currently called community does not always look alike. By using ‘community’ as a generic term, an essential part of a powerful community is diluted:
A community is a group of humans with a shared identity who care for each other.
EAs aspire to care for everyone, of course, but in practice, no single individual can take care of everyone else. Our limited time and energy pose hard constraints on even the most caring humans. The above definition brings about the realisation that a global community is just nothing that current human beings can achieve.
Clear distinctions can help to better define activity portfolios, responsibilities and to set realistic expectations. We suggest the following definitions:
An aspiring effective altruist (EA): an individual that identifies with EA principles.
An EA community: a group of aspiring EAs who take care of each other.
An EA organization: a group of aspiring EAs collectively contributing to EA’s mission (anything from local groups to EA Czechia to specialised orgs like the Open Philanthropy Project).
The EA network: the global set of individuals and organizations who identify with EA principles.
One could illustrate the definitions hierarchically as follows:
But it is probably more correct to assume that (i) not all aspiring EAs have found a community within the EA network; (ii) not all EA organizations exclusively need aspiring EAs as employees; (iii) EA communities can contain minimal amounts of not-entirely-aligned individuals; and that (iv) some organizations function without a community at their core:
Whether we call people ‘aspiring EAs’ or just ‘EAs’; whether we want to say ‘EA network’ or ‘system’ or ‘nexus’ - the point is that the EA network needs a clear understanding of what it means by the term ‘community’, and what others understand by it.
The rest of this article will further illustrate the case for defining ‘EA community’ along the lines of ‘a group of aspiring EAs who take care of each other’ and for labelling the other things in very distinct manner.
Breaking down current ‘community building’ efforts
A vast array of activities has so far been labelled ‘community building’ - to many community builders’ confusion. The term seems to have originated from this line of thought:
It is good to have people who:
Thoroughly understand EA related topics;
Act on that knowledge to do the most good; and
Stay on top of the collective thought development.
Bringing about more of (1) is good.
A decent definition of the EA network one would want to ‘build’ thus seems to be ‘the set of people and organizations who seek to maximally satisfy all three criteria of (1)’. And anything that remotely resembles (2) has since been labelled as ‘building’:
CEA (incl. 80’000 Hours) ‘builds’ globally by trying to be the go-to organization for anything EA; creating online content; compiling resources & knowledge; coaching and placing individuals; coordinating with key organizations; encouraging donations; supporting local groups & projects; and hosting Schelling events for the actively engaged segment of the global network. Some of these services are also offered by LEAN.
Other more geographically focused organizations’ activities often appear to be similarly broad: from small social events; to freshers’ fair booths; to career planning and EA concepts workshops; to personal coaching; to representing EA at important events; to organising large-scale conferences; to translating key resources into the local language; all the way to—yay, meta! - supporting other local groups.
Sorting these activities into more self-explanatory categories gives a better understanding of what ‘community building’ entails. There seem to be roughly five different parts to ‘building’:
All five types of activity can take place, in varying forms, at the global, organizational, and community levels. Disentangling what is and should be happening how, why, when and where is another step in optimising the coordination of the EA network.
Professional ‘network development’
In CEA’s current ‘community building’ grants round, the sole quantified metric to assess recipients is how many people they direct towards ‘priority roles’. An understandable metric, as it should prove value alignment and the short-term effectiveness of a group.
Despite explicitly leaving room for various other achievements by grantees, it skews CEA’s implicit definition of ‘community building’ towards one single activity: recruitment. That’s not a bad thing at all—EA Geneva itself agreed to CEA’s terms because we think they make sense. It just does not have much to do with community.
And it is one of several pointers in a general direction: organizations without paid staff are generally unlikely to do an impressive job at at least four, if not all five, of the activities we defined: marketing, coordination, recruitment, and training.
A look at the private sector suggests that most of the short-term value that local groups and the like are currently expected to produce is best reaped by specialised professionals.
Alternatives to letting organizations grapple with these activities would be to centrally employ/organize:
Professional representants (lobbyist-ish) to build influence in relevant networks
Professional EA coaches to support aspiring EAs in their work or studies
EA training and seminars to skill up individuals and institutions to do more good
Recruitment tours through unis
Speaker tours through conferences, foundations and companies
Targeted outreach to outside experts in related fields for input and feedback
These roles are attractive also because staff could be trained up and coordinated more effectively, providing much higher marginal return than leaving it to each local community or organization.
The value from organizations and communities will then come through activities that require (semi-)permanent local presence. Local anchorage in relevant regions gives access to more networks and will likely improve the EA network’s impact potential.
The network can do a great job at conveying the drive of its members. But the roots of the excitement most people experience at their first EA Global conference are developed by tight-knit communities that work on EA organizations. What makes such communities?
Limits to community size
A well-known objection to EA claims that many people care more for their close circle than the rest of the world. It seems to conflate two types of care. One type regards moral patienthood and often is meant to imply some form of responsibility. The other is the practical question of “who do I invest my relationship time in?”. Aspiring EAs sometimes seem to make the same error.
Caring for all beings in the abstract does not change that humans are limited by their time and their number of neocortical neurons when it comes to those who they can take care of personally. In return, the number of people who can take care of them is similarly limited.
Luckily, it’s not necessary that everyone personally takes care of everyone else’s well-being. A sense of moral concern is enough for EA communities of ~30 people (cf. size of bands) to coordinate effectively as a network.
Psychological safety as a key
To provide a true sense of community, members need to know that they are safe. This visceral trust is maintained through strong relationships, structure and participation. You want people to really know that you care for them? Well, everyone has to put in some quite a lot of work.
An alleged lack of psychological safety within the EA network appears to put off even theoretically hardcore EAs from taking a job at certain organizations. We postulate that the underlying problem is not, as often put forth, some form of discrimination or a lack of diversity. Instead, there seems to be a lack of the visceral sense of care and belonging that no network can provide.
Calling a network a ‘community’ creates a false sense of belonging that can only be upset in the long run.
Nowadays, most Westerners can freely choose their primary community. It does not even have to include your biological relatives. That does not mean, however, that it makes things easier or allows for bigger communities. On the contrary, the absence of century-long traditions only makes community building more difficult, already on a small scale. Humans are still just monkeys.
A human’s primary community is not going to be much bigger than a band. The band has to actively choose who to include and inclusion means commitment—which will be limited by the amount of time and energy community members can invest.
EA hubs, like Berkeley and Oxford, seem built around professional organizations that have come out of very small communities (e.g. close friends, couples, colleagues).
An explanation for small communities nourishing impactful organizations could be that co-working on your relationship(s) and/or projects brings people closer together. Professionalization bonds even more intensely. Output generated by work attracts other similarly dedicated individuals who often search for just such a community.
A focused and intense-yet-caring community can inspire others to branch off and develop their own community or helps them to find their spot in an existing one. Dynamics which maintain a virtuous, or at least self-sustaining cycle for hubs that have surpassed a certain critical mass.
Organizations without a paid community at their core seem much less impactful and sustainable than organizations built around a mission and a core community. Of course, it is a hen-and-egg problem, but understanding that seems important to spot whenever there is either a community that should be working on something or when there is something to work on that could shape a community.
Groups of friends might want to try to figure out how they can work together on a promising project. Volunteer organizations should try to figure out whether they can build a community by figuring out what they could work on professionally. It’s the ultimate stress test and with the right systems and output, it provides immense learning value and proof to the network.
The EA network has one mission: do the most good. EA organizations, communities and individuals aim to maximally contribute to that mission, but they also have to make many trade-offs to take care of things that are only indirectly related to the core mission (e.g.: maintaining their health, education, epistemic standards and local integration). The network has to leave room for adaptation to preserve its own resilience but has to unmistakably focus on pushing its core mission by incentivising maximal contribution.
To do better than just one community or organisation, a network needs to coordinate well. To coordinate well, one needs to have the means to build alignment and professionalize. To build alignment, communication and training systems need to be installed. To develop the necessary systems, people need to understand the network and its components.
To develop the network, EA needs people whose job it is to build the systems necessary to effectively align communities and organizations. Essentially, these people would get paid to dispose of the necessary relationship time to be a part of two or more communities.
‘Network developers’ then have distinct responsibilities from ‘community builders’. Community builders can focus on maximising their community’s utility (and will do a better job if employed by the community), while network developers ensure alignment maintenance when employed by a ‘network development organisation’. Both roles are then incentivised to cooperate optimally as the network’s impact and the community’s relevance depend on each other.
What we propose could look something like: CEA becomes the ‘network developer’, maintaining the networks quality and incentivising maximum contribution. LEAN focuses on supporting actual communities in developing their members, teams and organizations. Organisations like the EA Stiftung, EA Czechia, or EA London support CEA locally by complementing its efforts with services that are difficult to tailor to local realities from afar. Local groups develop the capacity to optimally feed into the network depending on their biggest possible value add. Communities take care of its members on a human level and build the badass, specialised core teams that EA needs for maximum impact work.
Of course, this is fully conditional on finding the right people, communities and organisations that should be part of the network. That’s a discussion for another time.