This update covers CEA’s work in the second quarter of 2021.
CEA brings people together to discuss how we can help others as effectively as possible.
Program progress in Q2
You can find more details for each program further down in the update.
Key output metrics remained roughly stable quarter-on-quarter, as we focused on hiring.
Aside from hiring, we focused on generating CBGs (Community Building Grants) at focus universities. To do this, we planned a US group leaders retreat for August, a UK group leaders unconference for August, and launched a remote summer program for community builders.
We handled 60+ cases, including some important media cases.
While we’ve exceeded our targets overall for events that we’ve run to date, several events have been delayed or scoped down due to COVID, so we’re not on track to make as many new connections as we wanted to this year. We’re reassessing our events calendar for the rest of the year to try to get back on track.
We recently hosted our first in-person event since the pandemic started: the EA Picnic in San Francisco. We are still collecting responses to our feedback survey, but LTR was 8.48 at the time of writing (with a 21.5 % response rate).
We ran a well-received virtual training course for EAGx organizers.
Hours of engagement on the EA Forum are up by 124% year-over-year.
The Editing Festival led to the tagging of hundreds of older posts, making it easier to search for pre-2019 content.
Internal progress in Q2
Our operations team improved CEA’s legal structure, began building new customer relationship management software, improved cybersecurity, and built a new internal wiki/staff handbook.
We made four full-time hires:
Rob Gledhill (CBG Programme Manager)
Sarah Cheng (Developer)
Jonathan Mustin (Developer)
Litawn Gan (Finance and Data Lead)
These updates focus on the object-level progress of our front-facing programs. Hiring, personnel, and operations are covered in detail in the next section.
We help EA group organizers by advising them, providing resources they can use, and creating online spaces where they can share resources and support each other.
Recruitment: 140 people started an In-Depth Fellowship this quarter, compared to 245 who started last quarter; however, this decrease is expected due to the university trimester system.
Retention: 36% of highly-engaged EAs have a funded group organizer in their city (vs. 33% in Q1). This increase comes from a newly funded organizer in Washington, DC.
Broad support: 54 calls with group leaders in Q2 vs. 58 calls in Q1.
We maintained our broad support for group leaders and started adding some team capacity.
We had around 54 calls and 77 in-depth email / Slack exchanges with group leaders — approximately the same as last quarter. We received positive feedback on the calls (average likelihood to recommend of 9.28/10).
We launched a role for scalable university groups support and received >50 applications.
Some impact stories:
We encouraged and supported EA LSE (London School of Economics) to organize a Summer Fellowship program with the intention of recruiting future leaders at LSE, University College London, Imperial College London, and King’s College London.
Catherine helped a relatively large EA group to identify a new leader after there was a gap in succession.
Community Building Grants (CBGs)
We hired a new CBG Programme Manager, made 6 grants, and piloted activities to improve the CBG university pipeline.
Joan did extensive outreach, which resulted in 4 of our 5 top candidates applying.
We made the following grants:
Renewal for EA Israel (0.5 FTE, 12 months)
New grant to EA DC (0.25 FTE, 6 months).
University (all new, some covering summer programs):
Swarthmore (0.25 FTE and 0.5 FTE for 2 people, respectively, for 7 months)
Yale (1 FTE for 6.5 months)
Stanford (0.5 FTE for 4 people for 2 months)
Brown (0.75 FTE for 2 months, 0.25 FTE for 2 months).
Program feedback: We received helpful critical feedback from a group of CBG recipients about changes they would like to see in the program.
We added this to our mistakes page, and followed up with a survey to prioritize improvements.
In Q3, we plan to work on website updates and conduct a focus group related to city / national group strategy.
The new CBG Programme Manager plans to follow up on some of these issues when he joins in the fall.
Our hypothesis is that funding organizers is the highest leverage way to increase the impact of a university group. As a result, we piloted programs to see if we could increase the CBG pipeline at focus unis.
Will Payne (contractor) launched an 8-week remote program for leaders at focus universities. Students work 4-16 hours / week on community building projects. They also attend weekly workshops and discussions related to community building with pre-assigned reading and activities. 34 people are participating, including students from Caltech, Berkeley, Columbia, and Stanford.
Jessica McCurdy (contractor) is running an EA university groups retreat in the US in mid-August which has 40 university leaders registered. She is also planning an unconference / retreat in the UK in mid August.
The Student Careers Team (SCT) is running a ‘Summer in Oxford program’, where 20 university group leaders are living in Oxford and working together on community building projects. The SCT has owned all the programming, and we’ve provided some additional support by making several grants to enable attendees, meeting with attendees and hosting them in our office.
Notable impact stories:
We will have attendees from the law schools at Harvard, Georgetown, Columbia, Yale and UChicago at the US leaders retreat; as it stands, EA groups at law schools are generally bottlenecked by leadership capacity.
Between Jessica’s work and Will’s work, we will have 11 dedicated group leaders taking part from Columbia, Caltech, and Georgetown, university groups that were recently started and may benefit substantially from additional encouragement, connections and strategic advice.
543 people started a fellowship in Q2, which is about half as many people as in Q1, mostly because groups normally start their fellowships in Q1.
All data is based on the number of people who started a fellowship in the listed quarter. Fellowships last for 8-10 weeks. The average completion rate in Q1 was ~81%.
Notes on this data:
The Q4 2020 intro fellowship had 150 participants from focus universities.
We ran the Q4 2020 intro fellowship in collaboration with EA Stanford.
We ran the Q1 2021 in-depth fellowship in collaboration with EA Oxford, and the Q2 in-depth fellowship in collaboration with EA Stanford.
We ran both rounds of our The Precipice reading group in collaboration with EA Stanford.
In Q2, we confirmed participant numbers with all fellowship organizers. In Q1, we assumed that fellowships for which we didn’t know participant numbers had 12 participants (the median for fellowships for which we knew the numbers).
Other notes on the fellowship programs:
Attendees’ average likelihood to recommend the fellowship to a friend is very high at ~9/10.
This quarter, she hired Yi-Yang Chua for operations support.
EA Virtual Programs ran two facilitator trainings for 27 new facilitators for the upcoming July-August round.
Max Daniel developed a new curriculum for the In-Depth Fellowship and ran two pilots: one with Oxford, and an ongoing pilot at Brown. One of his tentative observations is that participant selection might influence program quality significantly.
Some of the In-Depth participants went on to do the Stanford Existential Risk Initiative research fellowship this summer.
Emma Abele used the virtual programs to help start an EA group at Georgia Tech.
In the Q1 update, we mentioned that we were working with Yale to help seed the Georgetown group. As a result of the fellowship, 7 of the 14 fellows from Georgetown are keen to get involved in organizing a group, and the current co-organizer is participating in one of the summer programs mentioned above.
The Community Health team aims to preserve paths to value and mitigate key risks to the future of the EA community. Our work includes fostering a healthy culture, improving diversity, mitigating harm done by risky actors, protecting the brands of EA and longtermism, and identifying risks to early field-building.
Julia Wise went on parental leave in the middle of June, and will return full-time in early August.
Sky Mayhew is transitioning into working for CEA as a part-time consultant, instead of as a core staff member. She plans to hand off her current responsibilities and take extended leave.
Hiring was the team’s major focus for this quarter. Because community health work requires an unusual combination of skills, we expect to take hiring slowly and leave these positions open for much of this year rather than definitely making a hire. We listed 4 roles as ‘expressions of interest’ and expect to hire 0-2 new staff to the team for those roles. The open roles are:
We’ve also hired a personal assistant (contractor) for Nicole Ross, to help maximize Nicole’s capacity to focus on high-priority work.
15+ inquiries or cases regarding media stories about EA.
5 concerns around interpersonal problems such as sexual harassment.
~25 cases where we advised on situations in early field-building (geographical areas or academic/professional fields where EA is just getting established).
6 “organizational health” cases where we advised organizations or projects on conflicts of interest, conflicts between staff, best practices in HR, etc.
9 other situations where we advised groups, organizations, and individuals on situations like conflicts between staff or group members, online conflicts, personal or mental health problems, and improving diversity, equity, and inclusion.
Examples of different types of cases from this quarter:
We coordinated and coached spokespeople for a press article about EA and longtermism.
We provided support and advice to 3 sets of community members concerned about EA colleagues struggling with personal/mental health problems.
Key project areas
Given low capacity on the team this quarter, we’ve mostly prioritized epistemics and DEI.
Diversity, equity and inclusion:
We advised an EA organization that asked for guidance around better supporting their staff and recruiting a more diverse staff.
We mentored and met with some community members who were interested in serving as early-career role models for other EAs from underrepresented groups, in their fields or in the EA community more generally.
Progress made here was focused on hiring.
We conducted an expedited trial for a candidate, but it ended up not working out. We are now focused on completing our standard hiring process to find other plausible candidates.
We assessed 8 potential public spokespeople for EA and longtermism. We plan to offer 2 or 3 of them media training.
We worked with organizers of camps and multi-day workshops in the EA and rationality spaces to develop a guide to community health for these events (for example handling alcohol, interpersonal conflict, and dating at camps or workshops). We created a Slack channel for organizers of EA camps and multi-day workshops — many of whom had not been in touch with each other before, despite having similar goals.
Events enable attendees to make new connections, learn about core concepts, share and discuss new research, and coordinate on projects.
Going forward, the team plans to focus more effort on making high-quality connections rather than curating content We chose to focus on this metric because connections have generally been the source of our strongest case studies, responsible for most of the impact recorded in a survey of highly involved EAs, and the main thing that attendees report to be their most valuable experience at our events.
We still need to figure out exactly how we’ll measure this, but we’re broadly looking for connections that were positive for both people involved, and were useful for at least one of the parties. The paradigmatic example would be connecting a junior person to someone more senior who can give them advice, mentorship, and/or further connections. However, we think that peer-to-peer connections (between people with similar levels of experience) might also be useful.
This means that developing the stewardship program is likely to be a key focus in the future. We might have a smaller selection of high-quality talks at future events (to structure the event, and convey key messages).
Here’s information on how many connections we estimate were made through our events from 2019 through 2021. A proxy for this metric is the number of estimated total new connections:
|Year||Estimated total new connections|
|2021 (to date)||6,289 (to date)|
The 2021 figures do not include figures from EA Global: London, which is scheduled for later in the year. However, we think we’ll likely need to add more programming or expand events (e.g. by adding a virtual element to EA Global London) to match last year’s level. The main reason we’re falling short is that we had to turn EA Global: San Francisco into the smaller EA Picnic because of COVID, and because most of the planned EAGx events have been delayed until 2022.
EAGx organizer training
The EAGx online training aims to increase the confidence and connectedness of EAGx organizers.
As expected, the overall feedback for online training was not as good as the feedback for in-person training in 2020, but was still pretty good.
9 attendees filled in the survey after the training:
9 out of 9 said they feel confident that they can perform their role to a high standard.
8 out of 9 said that they have a clear understanding of the strategic goals for their event.
Average likelihood to recommend was 8.8 (in comparison to the in-person one in 2021 with 9.6).
They felt that they could reach out to an average of 4 additional organizers or CEA staff after the training.
EA Picnic: San Francisco
Much of this quarter was spent planning the EA Picnic (July 11th), our first in-person event since the beginning of COVID.
This event focused on increasing the numbers of connections between highly-engaged community members:
Number of registered attendees: 191 (We estimated 200 attendees).
Gender: 70% of accepted applicants identify as male. (About 72 % of applicants responded to the application’s gender question.)
Under-represented groups: 32% of registered attendees who reported ethnicity are from under-represented groups. (About 66% of applicants responded to this question.)
We hosted a session for attendees from under-represented groups with the help of Zachary Robinson from Open Philanthropy
We aimed to increase the number of new connections per attendee by at least 50% (compared to EAG Reconnect), but in fact it was about the same (4.8 new connections vs. 5 for EAG Reconnect).
We are planning to run the EA Coordination Forum for 40-60 attendees in Napa (California) on September 27-30. We also hope to be able to hold EA Global: London in person on October 29-31.
Amy is now being managed directly by me (Max).
The Effective Altruism Forum (EA Forum) aims to be the central place for collaborative discussion about how to do the most good.
Hours of engagement hours are up by 124% year-over-year as of July 1st. (This is above our target of 100%, and means we’re on track to double engagement hours this year.)
Progress this quarter
We released Auth0, a new login system on the Forum.
In order to tell whether community members continue to engage with our projects, we need to be able to identify users across projects.
This system was already being used for events registration, EA Funds, and Giving What We Can.
This unifies accounts across those platforms, which should make it easier to track actions by highly engaged people across the platforms and will hopefully give us more insight into retention opportunities.
Thanks to the Editing Festival, 100% of the articles with at least 20 karma on the EA Forum have been tagged.
We hit our goal of >15,000 words of edits on the wiki from people other than Aaron and Pablo.
Charter Cities intervention report—David Bernard, Jason Schuhkraft (mentioned on Astral Codex Ten and Marginal Revolution)
Career choice for longtermists – Holden Karnofsky
EA needs consultancies – Luke Muehlhauser
What gives me hope – Michelle Hutchinson
April Fools’ Day posts
The Operations team aims to provide the financial, legal, administrative, grantmaking, logistical, and fundraising support that enables CEA, 80,000 Hours, Forethought, and Giving What We Can to run efficiently.
Finance Lead hire: Litawn Gan joined us as Finance Lead in June. Louis Dixon, the outgoing Finance Lead, is tapering down his hours until his departure in mid-September.
Operations Associate hiring: We are evaluating candidates who applied to our Operations Associate role.
Improvements to financial systems:
EA Funds is now processing such a large volume of transactions that manually coding each one is inefficient.
We worked with Sam Deere at EA Funds to add a batch coding system that saves 20+ hours of bookkeeping time per month. The coding within our books has also been overhauled to make sure that the system is scalable and accurate.
Customer Relationship Management software:
We scoped out plans for the software, with a focus on supporting the Groups team. We tested some prototypes.
We selected our development partner—OpenTent—who specialize in CRMs for community building. We are now beginning to develop the CRM.
Legal restructure of CEA: The restructure of CEA was finally concluded after two years. The new framework is a formal parent/subsidiary relationship between CEA UK and CEA USA. The board was also reshuffled, with all current board members appointed to the parent (UK) board and two new board members (Nicole Ross and Becca Kagan) appointed to the subsidiary (US) board.
Feedback: We sent out a mid-year survey asking our clients (at the orgs we support) for feedback, and used the results to prioritize future plans.
More people began to use the refurbished Oxford office this quarter. They report that it’s a pleasant and productive working space.
We’ve finished allocating costs for this project, which has been a significant challenge given the complexity of the arrangement.
User queries: We improved the efficiency, response time, and phrasing of some of our standard responses to customer queries (e.g. on Funds). We estimate that these changes will save us ~10 hours of time per month.
Our internal file storage system (Google Drive) was insecure and unorganized. We removed all legacy users, and established a new, more secure structure.
CEA moved to a new password manager (1Password) which is more robust than the system we were previously using (LastPass).
Staff completed a number of tests to ensure they were aware of common phishing techniques used by hackers.
Despite this, the risk of a cyber attack still seems like a real and present threat given the uptick in cyber crime over recent years. We will continue to seek improvements and ways to reduce risk throughout the year.
We have developed an internal wiki on Notion, covering our strategy, values, norms, and policies.
We also moved the staff handbook for the whole legal entity onto Notion. We simplified many policies and made the book easier to skim, so that it’s less overwhelming and legalistic for new hires.
We set key proxy metrics for each department. In most cases, we were already tracking these metrics, but we were also tracking other metrics/factors. We’re now more focused on these metrics.
Forum/online: Hours of engagement
Events: High-quality connections
Groups: Number of people who have completed In-Depth Fellowships
Operations: User satisfaction scores
People operations: CEA team morale
The Community Health team isn’t using a key proxy metric at the moment.
We plan to supplement these metrics (which are Goodhart-able and don’t capture all of the paths to value) with annual objectives focused on supplementary/qualitative/quality-focused targets. For instance, groups might have an objective about improving the quality of In-Depth fellowships or providing extra training for grant recipients.
We are also running a hiring round for an EA Strategy Coordinator. We had a work trial with one candidate who declined our offer, and we are now close to beginning further work trials.
Previously, we’d been thinking about ourselves as an Oxford-based company with remote staff. We now think of ourselves as a remote organization with a hub in Oxford. We think that this will allow us to support remote staff more effectively, and attract remote talent.
Harri Besceli has been carrying out hiring support for the two open Groups positions during Q2. In Q3 he will hand over CBG responsibilities to Rob Gledhill (new hire), and try out other roles at CEA (most likely as a research assistant to Max). If those roles aren’t a good fit, he’ll leave CEA.
We hired four staff this quarter, thanks largely to the efforts of the hiring managers (Joan, Ben, Josh).
Robert Gledhill (CBG Programme Manager)
Sarah Cheng (Developer)
Jonathan Mustin (Developer)
Litawn Gan (Finance and Data Lead)
We also brought on some new contractors, including:
Hillary Grills (part-time People Operations support)
Yi-Yang Chua (part-time EA Virtual Programs support)
Marisa Jurczyk (part-time Groups resources support)
I’m hopeful that we can hire around 5 people in Q3.
We’re in the process of finishing recruitment rounds for the following positions (none of which are still open to applicants):
Community Events Manager
EA Strategy Coordinator
Scalable University Support
We’re also trying to expand the community health team by running expressions of interest for the following roles:
Early Field-Building Specialist
Epistemics Project Manager
External Communications Specialist
Sky Mayhew (Community Health) will move from being a core team staff member to working as a part-time consultant.
The average morale reported for Q2 was 7.03/10, compared with 6.45/10 in Q1.
We’re hoping to hold our first in-person team retreat next quarter, pending UK COVID restrictions.
We have continued to host weekly games, as well as alternating between socially-focused All Hands meetings and content-focused All Hands meetings to support team connectedness.
We added a more detailed, anonymous survey about managers’ performance to our set of bi-annual reflection tools.
Amy Labenz returned from parental leave. Julia Wise and Josh Axford each took some full-time parental leave and have now returned part time or will do so soon.
n = 25. ↩︎
There are currently 19 focus university groups (chosen primarily based on the expected influence of an average graduate, with some weight also placed on the groups’ track records and leader quality. Current focus university groups in no particular order: Harvard, Oxford, Cambridge, Stanford, Yale, Princeton, MIT, Hong Kong University, Georgetown, Swarthmore, London School of Economics (LSE), Caltech, Berkeley, Chicago University, Columbia, Penn, and the law schools of Harvard, Stanford, and Yale. ↩︎
This is true across group-run and Virtual Program fellowships. However, there is likely significant sample bias in the data, as the survey was only filled out by ~30% of EAVP participants and ~66% of local participants, all of whom self-selected to give us feedback. ↩︎
This is the number of attendees for each event multiplied by the average number of new connections survey respondents said they made. We expect survey respondents to be more engaged in the event than average, so we think that the absolute numbers are too high, but the differences between events/years might be more meaningful. ↩︎
This figure excludes EAGx events. ↩︎