Hide table of contents


We're starting a coworking group for research related projects in EA Israel. We meet weekly, the work is mostly on individual projects I've curated or chosen in advanced by the volunteer researcher.

The goals are

  1. Making the process of getting into EA (and EA research) easier.
  2. Experimenting with a scalable structure for co-working groups and project management.
  3. Actually providing some direct value.

In this post, I aim to explain the structure of our group and the reasoning behind it.

I'll mention upfront that I'm very interested in making more connections with existing organisations and other local groups to share knowledge and to collect more research/project requests and other possible collaborations. See the last section of this post.


Existing bottleneck in EA community building - transition from moderate to high engagement

The recent EA leaders survey suggested that a key bottleneck in our EA community funnel is

More dedicated people (e.g. people working at EA orgs, researching AI safety/biosecurity/economics, giving over $1m/year) converted from moderate engagement due to better advanced engagement.

At the start of 2019, there was a lot of discussion in the forum about What to do with people, considering that it is difficult to get hired by an EA org. There seems to be a problem where many promising people interested in EA can't find a way to directly contribute.

This is arguably a major problem that directly slows down the EA endeavor, and more clearly it is a bottleneck in the growth of our community. It may be better if they earn-to-give or build career capital, but that may not answer the social need to be a more integral part of the EA community (which may help to improve dedication and reduce Value Drift).

Lack of infrastructure to support volunteers

EA is Network Constrained. Especially problematic are the low management and mentorship capacity, which is extra costly because EA is also Vetting Constrained. It is difficult to estimate how good a volunteer will be, and this makes it difficult and costly for EA organisations to use volunteers effectively, which also means that there are fewer volunteering opportunities.

Also, many EAs lack relevant experience to do direct work well. From the recent survey,

In the 2019 EA Survey, we asked respondents about areas in which they had at least 3 years of work experience or graduate study. [...] There were 462 EAs (38%) who selected at least one of the most common areas cited in the leaders forum survey [as talents most needed in the EA community].

These considerations makes it harder for people who are new to EA, and have a specific set of interests and capabilities to actively participate in the EA community. This is the place where we think that local communities can serve as a gateway.

Israel EA Population

We feel that the above is true of our community in Israel. We've seen that many people are very interested in EA, and many of them are interested in contributing. However, we've found it challenging to convert people from interested casual participants to dedicated community members, and a big part of that has been generating volunteering or research opportunities that we believe are both impactful but also provide personal fit to the individual.

In a future post we will describe EA Israel. For this post, I'll just mention that Israel has a strong research community (especially in ML, clean meat, game theory, decision theory, behavioural psychology, and math and CS in general).

Research projects

I'm using the term "research" quite loosely here, which can be for example anything which results in a suggestion to an existing organisation, a forum post, a directive for subsequent direct work.

Research questions and Project list

I have collected a small list of concrete research questions and projects, with an associated guide - found here. This may be of interest to newcomers to EA who do not have a clear idea of what they want to work on, or as a collection of projects to be used by organizations or local communities when trying to set a research agenda.

From my talks with researchers at various EA organisations (in and out of the academia), it seems difficult to have part time, non-expert volunteers actually contribute to an existing research agenda. One of my goals is to see if we can find ways to meaningfully contribute without taking much time from expert researchers.

The main motivation for that is to have a list of ready-to-use ideas for research projects which are self contained, actionable, clear and short-termed.

Coworking group

At each session, each participant works on her own project. Some projects might be in teams, but I expect that it would mostly be individual as there is a lot of different expertise and interests. The coworking group is open to EA-projects which are not necessarily research.

If we feel like it, we use Pomodoro schedule of 25 minutes work and 5 minutes break. We take an hour long break in the middle for lunch (funded by CEA) and a status update round.

Some concrete features and guidelines:

  • Start small with dedicated individuals. After several weeks, open it gradually.
    • We want the culture to be established by a carefully chosen set of core members who can do effective research work and are fun to be around.
    • Ideally, members from the core group will be more comfortable supplying guidance and leadership later on.
  • Before attending the group, the project is decided in advance in a one on one. This has two justifications:
    • Checking suitability to the group's ongoing work and to the EA agenda.
    • Preparing for a productive work time!
  • Meeting for 5 hours weekly on Friday 10-15.
    • We decided on 5 hour sessions after receiving multiple recommendations for 5 hours as the approximate minimal amount of time (of effective work) that is worth the maintenance cost.
    • Friday is the first day of the weekend in Israel, and public transportation is limited from Friday evening onwards due to religious laws in Israel.
    • I expect most participants will have hard time committing to come every week. As long as this does not affect long-term participation, this should only be a problem when working on projects that other people are dependent on, but in those cases we can set a higher commitment bar (say, to commit to work at some time during the week).
  • Outputs should be of interest to the community or with direct impact following an analysis from an EA framework.
    • Ideally, I'd love for the group to be able to tackle concrete needs from EA orgs and do that with collaboration with relevant experts.
  • We also aim at Cause Impartiality and to be fairly Cause Agnostic.
    • We explicitly want to allow any type of work that can be considered by the researcher a possible way to do the most good, even if it is not among the top cause areas.
    • If the project is not directly related to an EA org or community research request, then the researcher should typically start with an analysis of the impact of her research proposal using a common EA framework to publish on the forum or otherwise engage with the community.
    • We believe that the EA-framework and engagement with the EA community are more important at an early stage than direct work on what is currently considered the most important causes, and therefore this initial cause agnositicism allows us to engage more effectively with new members and expose them to cause prioritization frameworks.


The coworking sessions should be fun, productive and supportive. This starts with good people. Also important are good food and a comfortable space.

I'll personally try to make sure that everyone is comfortable, that they know each other and know what they want to do at each session.

Mentorship and management

We want the group to become a place where participants can get experience and to help them become more connected to the global EA sphere.

  • We'd like participants to volunteer/intern at an EA aligned organisation.
    • This gives the participant a chance to work directly in a structured environment with expert mentorship.
  • Some participants can be connected to local experts.
  • Some participants have domain expertise, with which we can advise and learn from one another.
  • Some participants have strong general familiarity with EA.
  • I'll do my best to connect participants with whoever can guide or mentor them, and to suggest projects and set goals.

I have some relevant research and management experience. This is useful, but I do not think that it is necessary in order to start such a group, as long as you have the right people and can foster a good environment.

What will be a success?

Some estimated metrics, evaluated as the results from now to the start of 2021, that help me convey a gut feeling of what I view as an OK outcome (the minimal outcome below which there is probably something very wrong) and a Great outcome (above which there is probably something very right!).

  • Number of core members (That attend >75% of sessions)
    • OK: 3
    • GREAT: 10
  • Total number of members (That attended at least twice and intend to continue)
    • OK: 6
    • GREAT: 30
  • Quality of total research output
    • OK: at least one project that each core member is proud of
    • GREAT: 5 Important Outcomes*
  • Transitions from moderate engagement to dedicated EAs(with counterfactual being the non-existence of the group).
    • OK: 1
    • GREAT: 5

*An important outcome can be something like the following: a small but visible change in views for some community leaders in EA or for a significant population, a useful framework/tool/information that impacts some decision makers, correcting a major mistake in a paper or important post, a major contribution to an organisation equivalent to two weeks of expert work, or perhaps $2K of 2022 Impact Prize 😉

I would also like to measure how much we (the global EA community) will learn from this experiment, the resulting infrastructures that we'd help create (say, an updating list of important research projects for newcomers), more volunteers and interns at EA orgs, connections we'll make with the global EA community, improved basic skills for us researchers and how much fun this will be. But I think of these as correlated or secondary to the above goals and I do not intend to maximise for them.

What will be a failure

  • A major lack of interest from the EA community for the research outputs.
    • This means that either the communication work is much harder than I thought (or that I'm bad at it).
      • In this case, it is possible to double down on communication efforts.
    • Or that our team specifically is lousy at picking and doing important research.
    • Or that it is genuinely very difficult to conduct interesting research in this framework, using non-expert part-time volunteers.
      • I believe this to be the most likely failure mode. In which case, it should be carefully considered whether work on direct research is the best use of people's time.
      • Alternatives might include limiting the group to researchers with relevant experience, more dedicated mentorships, requesting more time out of volunteers, choosing the easier projects, or forgo hope of impactful research which means that newcomers and part-time researchers should not carry on with doing research for it's own sake, but mostly for the learning opportunities and as a vetting mechanism.
  • A work atmosphere which is not be productive, fun and supportive.
    • If someone is causing harm to the work environment, I'll talk with him and find solutions or if needed take him out of the team. We will not accept people to the coworking group if they are not committed or damage the group's dynamic.
    • Even though we do not aim at being disruptive, and most work is on individual projects, there are some important lessons from Ingredients for Disruptive Research Teams that are relevant here. There are also some major differences for a part time group of non-experts. I'll personally make sure to do my best such that:
      • The logistics are as easy as possible (food, location and possibly transportation).
      • Participants have clarity to the impacts of the chosen research projects.
      • There is a good balance of socialising and work time.
      • Participants will have a feeling of autonomy while not feeling lost.
      • The research outputs will have as much impact as possible. By reviewing written material and fostering communication with organisations and people that the research is directed towards.
      • Participants are excited about the work that they and the others around them are doing.
      • That everyone has someone to talk with if they need help with their process, with maintaining motivation, with general career/life related advice or with any struggle that it may be my place to help them with.
  • Participants may have difficulties in committing.
    • It would probably be bad if there would not be many members of the core group that persist for a long time.
      • It may be absolutely fine if most people still attend, say, 50% of the time.
    • It may turn out that the time is not good for many people.
      • We can do it twice per week, two different groups.
    • Perhaps getting funds for salaries, bounties, grants or prizes might help.

What I want from you

  • Requests for research projects you would be interested in. Even if it will not be relevant to our group, I'll put these here.
  • In the happy flow where there is indeed a liquid ecosystem of capable but non-expert volunteer researchers, how would you restructure your workflow and how much time do you think one should put on mentoring and directing/managing.
  • Collaboration on basic infrastructure. I'm particularly excited about
    • An updating list of research proposals (and other projects). (There are many possible features there)
    • An updating list of volunteering/internships/part-time-work opportunities.
    • An incentive system for outsourcing such research proposals (and other projects).
  • I'm curious about other local groups who are trying (or have tried) a similar format.
  • Feedback.





More posts like this

Sorted by Click to highlight new comments since: Today at 12:21 AM

Thanks for sharing this! It seems like a really exciting project, and I hope you continue to post updates. Very cool that you have explicit success metrics.

A semi-research thing I'm interested in is putting more information on Wikipedia. I wrote a little bit about this here. I suspect that for people who are new to research, or aren't entirely sure what subject they want to research, making existing research accessible is a similar task which is also quite useful for the world.

Thanks! We do have that as a possible project on our project list, but it wasn't on my mind as one of the first things to go for. And you are right that especially for people without much experience this really makes a lot of sense. Added to our workflow :)

Curated and popular this week
Relevant opportunities