Effective Altruism Infrastructure Fund: March 2024 recommendations

By Linch, calebp, Harri Besceli, Tom Barnes @ 2024-05-27T21:11 (+88)

This payout report covers the EA Infrastructure Fund’s grantmaking from June 16th 2023 to March 31st 2024 (9.5 months). It follows our previous June 2023 payout report.

• Total funding recommended: $1,697,882
• Total funding paid out^[1]: $1,386,854
• Number of grants paid out: 41
• Acceptance rate (excluding desk rejections): 49/173 = 28.3%
• Acceptance rate (including desk rejections): 18.4% (49/266)
• Report authors: Linchuan Zhang (primary author), Caleb Parikh (interim fund chair), Harri Besceli, Tom Barnes
• Funding breakdown^[2]
  • EA Groups: $456,780 granted across 18 grants
  • EA-related Groups: $312,304 granted across 6 grants
  • EA Content: $78,216 granted across 3 grants
  • EA Services and Infrastructure: $179,957 granted across 6 grants
  • Effective Giving: $53,000 granted across 2 grants
  • Research: $216,535 granted across 6 grants
  • [Total]: $1,386,854

8 of our grantees, who received a total of $521, 206, requested that our public reports for their grants are anonymized (they’re written below). 1 grantee, who received $2,500, requested that we do not have a public report at all (You can read our policy on public reporting here). Our median response time over this period was 27 days, and our average response time was 42 days. For paid out grants, our median and average turnaround times are 57 and 61 days, respectively.

Highlighted Grants

Below we’ve highlighted some grants from this round that we thought were particularly interesting and that represent a relatively wide range of EAIF’s activities. We hope that these reports will help donors make more informed decisions about whether to donate to EAIF, as well as help the wider community understand our work.

Rethink Priorities Worldview Investigations Team ($168,000): Stipend to improve their Cross-Cause Cost-Effectiveness Model, including a portfolio builder to help individuals and foundations prioritize their philanthropic spending. [Grant type: Research]

Note: This grant, while approved, has not yet been paid out, pending due diligence. 

This project fits in well with the EA Infrastructure Fund's tentative reorientation towards Principles-Focused Effective Altruism. 

The fund managers were highly impressed by the ambitious scope of this endeavor. Despite the EA movement existing for over a decade, there were no other publicly available cross-cause models with comparable breadth and an EA-informed perspective. The gap suggests that creating such a comprehensive model is much more challenging than it might initially seem. The fund managers admired the team's intention to produce a practical tool that funders could realistically use, rather than (e.g.) purely theoretical work on cause prioritization.

However, some fund managers were concerned about the default values used in the Cross-Cause Model, which sometimes appeared insufficiently principled or overly conservative. Caleb Parikh, the primary investigator for this grant, provided more detailed thoughts in a comment.

Overall, while excited to grant this project presently, the fund managers believe continued excitement for renewing the grant or offering similar grants hinges on a few key conditions: 

1. The methodology employed in Rethink Priorities' Cost-Effectiveness Model should be broadly reasonable, and easy for EAIF’s fund managers to endorse
2. The project should demonstrate potential to genuinely influence decision-making among major funders. 
3. We should broadly believe the team's proposed improvements to the model are likely to be useful.

Despite the high expected value, the fund managers acknowledge that real-world grantmaking decisions often involve holistic, contextual factors that may limit the direct impact of even thoughtfully-designed theoretical models. Consequently, they view funding this project as implicitly a low-likelihood, high-payoff bet that the models would end up directly influencing large grantmaking decisions.

Rethink Priorities: Survey of talent gaps ($9,000): Staff costs for a survey of talent needs in the effective altruism community. [Grant type: Research]

Rethink Priorities requested $9,000 to conduct a survey of EA organizations regarding their talent needs within the effective altruism community. The fund managers considered the Rethink Priorities survey team competent and likely to execute the project well. The fund managers believed this information could help capacity-building groups like university clubs and 80,000 Hours prioritize their outreach efforts more effectively. Notably, such groups had provided positive feedback on the proposed survey.

However, one downside risk was the concern that community builders, especially more junior ones like student group leaders, might place too much weight on the survey results without properly contextualizing them. This may lead them to (e.g.) focus too much on recruiting for specific talent needs highlighted by the survey, and neglect other important factors suited to their local communities and recruitment pipelines.

Ultimately, the fund managers assessed that the potential to provide clearly useful and actionable findings outweighed the risks, leading them to favor funding the survey proposal from Rethink Priorities.

Rethink Priorities’ 2023 EA Survey ($15,000): Staff costs to run and analyze a followup to the 2022 EA Survey and write a report on the results. [Grant type: Research]

Rethink Priorities' survey team planned to conduct two surveys assessing the "climate of EA" across EA organizations and the community at large, covering topics like post-FTX issues, trust in leadership, and harassment concerns.

The fund managers believed the findings could prove highly valuable for both individuals and organizations within the EA community. This was especially true if leadership organizations held systematically inaccurate assessments of broader community sentiments. When we made the grant, several specific organizations had already expressed interest in this information.

While the applicant originally requested $29,600, the fund managers deemed it fair to offer half that amount - $15,000. The decision captured the intuition that survey results constitute a diffuse public good, with benefits widely distributed across the community.

However, the fund managers hoped that partner organizations collaborating with Rethink Priorities would contribute the remaining funding needed. Their rationale was that entities with specific research questions stood to benefit more directly compared to the community at large.

(Note that Linch was also employed by Rethink Priorities when the above three grants were made. He was not in any way involved with the decision to make those grants). 

Astra Kamratowski ($7,716): Resources to produce 2 animated YouTube videos on common EA themes. [Grant type: EA Content]

We funded Astra Kamratowski to produce videos on EA themes. Many fund managers believe popular YouTube channels have strong potential for effectively communicating EA ideas. Further, they view video as a somewhat neglected medium currently, though less so than a few years ago.

At the time of evaluation, Astra's channel is very new, with only two existing videos of decent quality but lacking significant viewership. By supporting two additional videos, the fund aims to help Astra build more of a track record. This will allow better assessment of their videos' reception and potential for further funding.

Initial cost-effectiveness modeling left fund managers fairly pessimistic that Astra's videos would meet the quality and viewership thresholds necessary to be directly cost-effective. Consequently, this grant is valued primarily for its information value. It represents a bet on the possibility, while small, that the fund managers' pessimism ends up being incorrect and Astra's videos turn out unusually successful or impactful enough to merit scaling up production.

EA Netherlands ($23,000): Subsidies for an office space for the effective altruism/effective giving community in the Netherlands, for 12 months. [Grant type: EA Services and Infrastructure]

The grant covers approximately 40% of the costs for an office space in Amsterdam, which also functions as a community hub. The office has seen 735 visits from 75 external visitors and hosted 16 events. The fund's theory of change for supporting this project is twofold: firstly, that office access will boost the productivity of its users, and secondly, that connections formed at the office will prove to be highly impactful in the future.

Modeling based on valuing a worker-hour in the office at €50/h suggests the office will generate around €30,000 in impact.  The connections made at the office seemed valuable, but have not yet produced any impact stories that the fund managers consider to be "slam dunks". Another potentially impactful route for the office is through functioning as an event space. However, if that is the primary theory of change, it might be just as beneficial to subsidize event space rentals rather than office space, which is much cheaper. So the fund managers did not seriously analyze its value as an event space.
Further, survey results regarding the office were fairly positive, more so than might be expected for this reference class. 

Linch’s personal views:

(I provided some of my personal views on some of the highlighted grants, to help a) shed light on how some grantmakers think and b) personalize these grant reports to make them less dry and factual. This is an experiment; I hope it’s helpful to at least some people.)

I’m uncertain about EA Funds sponsoring projects like office spaces, from a procedural standpoint. Often, the individuals working in these spaces are already associated with organizations or projects funded by either EA Funds or other EA donors. An alternative , and possibly better, approach could be for funders to provide more money directly to the organizations, allowing them to make their own decisions about office costs. Having funders micromanage spending decisions of individual organizations, rather than evaluating them holistically based on impact or outputs relative to costs, may be problematic and inefficient.

On the other hand, it’s plausible that well-run coworking spaces generate community benefits and positive externalities that are not captured by individual organizations. For instance, collaborations formed in coworking spaces may lead to new organizations, and such spaces provide newcomers with an accessible way to observe how professional EAs work and interact, etc.

Anonymous (~$16,000): stipend for ~3 months for a medical doctor to pivot their career into founding a new impact-focused organization or joining an existing EA organization, likely in biosecurity. [Grant type: Individual Funding]

A medical doctor with extensive research experience is considering either founding a new impact-focused organization or joining an existing EA organization.

The fund managers believed the candidate is  likely quite competent and possesses a rare combination of experiences and skills within the EA community. However, due to the fund's own time constraints, they were unable to thoroughly vet some details regarding the grantee's theory of change and community connections. 

The grantee requested to remain anonymous. In accordance with our stated policy, the fund is willing to make anonymous grants if they would otherwise pass the rough cost-effectiveness bar. 

I, Linch, made the decision to include an anonymous grant on this list of highlighted grants in order to provide more transparency around anonymous grants for donors and other community members. The fund has previously received complaints that the decision-making process behind anonymous grants is more opaque than desirable, and we hope that writing more about anonymous grants can help shed more light on such decisions.

Effective Altruism Sweden ($9,218): Funding for a three-day EAGx-inspired annual conference for 100 people in the Swedish EA Community [Grant type: EA Groups]

We previously supported this national group for other projects before, and believe that their past track record has been pretty good.

The fund managers consider this grant to be highly cost-effective, with a cost of $92 per attendee, which is quite low for conferences (conferences can be quite expensive!). In particular, the cost per attendee, including food, is significantly cheaper than for EA Global.

Even when using conservative estimates that account for potentially reduced event quality and lower impact per attendee compared to EA Global, this grant still appears to be much more cost-effective than marginal donations to the Centre for Effective Altruism (CEA), which organizes EA Global.

Given that marginal donations to EAIF are likely not substantially more cost-effective than donations to CEA, this creates a fairly simple argument that supporting EA Sweden's conference should be a good bet for our marginal resources.

Richard Chappell ($6,050) : Subsidy for a one-course buyout of Fall 2024 teaching duties to improve EA ideas through public philosophy, ethics theory research, and EA community support. [Grant type: Research]
Note: This grant, while approved, has not yet been paid out, pending due diligence. 

Note: The legal recipient of this grant is the University of Miami, Richard’s workplace.

Richard Chappell (Forum) wanted a course buyout so he could spend more time working on EA-relevant projects including philosophy research, public philosophy such aspublishing on utilitarianism.net, and EA community support, for example, blog posts either simplifying philosophical research or analyzing relevant community issues. 

A fund manager highlighted several reasons for supporting this grant:

1.  Richard's research seems quite strong. 
2. The project aligns closely with the fund's vision of a "principles-first EA" community, we’d be excited for the EA community’s outputs to look more like Richard’s.
3. The risks associated with this grant seem quite low
4. Since it only covers a single course buyout, there is no obligation to provide further funding if the results are not substantial.

Linch’s personal views:

One concern I have is that the benefits of such work are likely to be highly diffuse and difficult to track, compared to (relatively) more tangible meta-EA efforts like recruitment or intervention research. This means that the fund managers may not be able to confidently assess the effectiveness of their decision, and it is also comparatively harder for donors and other community members to assess our track record. That said, this might just be an inherent cost of moving towards a "principles-first" approach, as concrete metrics that align well with final impact may be harder to obtain in this new paradigm compared to other EA domains.

Other Updates

• (Tentative) New strategy to focus on “Principles-First EA”: We’re in the middle of a strategic revamp, where we want to focus our grantmaking efforts more on Principles-First EA. This means that we’ll be more inclined to make grants to research on cause prioritization, cause-neutral projects that build communities focused on impartial, scope sensitive and ambitious altruism, and epistemic infrastructure. Compared to previous years, we’ll be less inclined to make grants to cause-specific movement building or infrastructure.
• Open Philanthropy and EAIF distancing: To increase funder diversity and independence, EAIF and Open Philanthropy implemented some distancing measures (such as Open Philanthropy employees stepping down from leadership positions). This means that EAIF will likely rely more on individual donors (maybe including you?)
• Donations to date: We’ve received ~$1.2M in donations from individual donors since our last update on August 31st 2023. ~$1M of that was matched 2:1 by Open Philanthropy, for a total of $3.2M in the last 9 months. We currently have about $2.3M in reserves. Thank you to everybody who donated! 
• Leaving Effective Ventures. Along with other organizations, we plan to leave our fiscal sponsor (Effective Ventures). We do not have a timeline for this, but will keep you posted. We do not expect the change to post significant inconveniences for either our donors or grantees.
• Grants database: Though it was technically launched last year, we now have a grants database for all of the non-confidential grants we’ve given out. 
• Update from Linch: I’ve joined EA Funds full-time as the second full-time employee, leaving my role as Senior Researcher at Rethink Priorities. I joined EA Funds because I think that now is a critical period to improve EA funding. 
  • Note that my primary role for EAIF is communications, rather than grantmaking.

Other writings

Below are some writings that are relevant to EA infrastructure work, including posts by EAIF fund managers, other EA infrastructure grantmakers outside of EAIF, or myself, which might be helpful to readers:

• What do Marginal Grants at EAIF Look Like? Funding Priorities and Grantmaking Thresholds at the EA Infrastructure Fund: A detailed discussion of “grantmaking thresholds” for marginal grants at EAIF. Essentially, given that we have limited resources and many good projects to fund, how do we choose which grants to make per $X we have? The post covers different projects we might want to fund at different thresholds ($X per 6 months).
• LTFF and EAIF are unusually funding-constrained right now: Our fundraising post in September. The post may help you decide whether EAIF (or LTFF) are good donation options relative to your next best alternative. (Please note that the Open Philanthropy match mentioned is now over).
• Relationship between EA Community and AI safety: Tom Barnes, an EAIF fund manager, on how the EA community could relate to the AI safety field going forward 
• Caleb Parikh on AI consciousness: Caleb Parikh, Project Lead of EA Funds and interim EAIF fund chair, shares some brief thoughts on why he thinks the broader EA community is underinvesting in research projects working on AI consciousness.
• Meta Charity Funders: Summary of Our First Grant Round and Path Forward: Meta Charity Funders, a new funding circle for EA meta funding discusses their grantmaking process and grants they’ve made in their first grant round.

Appendix

Other grants

We’ve categorized all of our grants during the payout period by type. For ease of understanding, we attempted to make the grants categorization mutually exclusive. There are many cases where a grant spans multiple categories, and for these we’ve made a judgment call to decide on how to categorize it.

EA Groups: Funding for EA groups, communities and programs. 

EA Related Groups: Funding for EA related groups, communities and programs. These may or may not be ‘EA branded’, and have a particular focus such as an EA policy group, or a rationality group. 

EA Content: Funding for the creation of EA content for a public audience, including video content, podcasts and translation for example. 

EA Services and Infrastructure: Funding for services and infrastructure which enable and support people in having a greater impact, for example including office spaces, job boards and coaching.

Individual Funding: Funding that enables people to develop career capital or explore impactful opportunities. For example, funding for a period of career exploration or funding to attend a conference in a relevant area. 

Effective Giving: Funding for projects that aim to raise funds for effective charities. 

Research: Research related to effective altruism, including cause prioritization research and research about the effective altruism community itself, for example. 

1. ^{^}

   Exact numbers may be off, e.g. due to currency conversion changes or grantees partially returning grants. I'm collating more information and will try to keep these numbers updated

2. ^{^}

   See categorization in the appendix here.