Announcing the Grantees from Round One of BERI's Project Grants Program

•

September 2024

This last June, BERI announced its first Individual Grants program: a Project Grants round. This post summarizes our process and announces the grant winners for this round.

Process

We received 55 applications by the end of June, when the round closed. We then spent a total of approximately 77 person hours reviewing the applications, contacting references, and consulting advisors.

In late August, the Project Grants Committee (comprising three anonymous individuals) finalized its recommendations and sent those to BERI's First Grant Fund Committee and Board of Directors for approval. We then emailed applicants notifying them of their status. We have spent most of September and early October setting up the logistics for the grants (drafting and sending out grant agreements, and disbursing funding). We have not yet finalized the disbursement of all grants.

Common considerations

We thought it might be helpful for future applicants to lay out some of the considerations that the project grants committee frequently found itself discussing. The following factors are presented in no particular order, and each committee member put different weight on each factor:

The applicant's general and specific competencies as evidenced by their past achievements (e.g., educational achievements, business achievements, social achievements, etc.).
The quality and similarity of the applicant's past work samples to the project being proposed.
The general and specific competencies of the applicant's references, as well as the content of the recommendations that the references submitted.
The applicant's track record of commitment to long-term futurist causes, or other signs of likelihood that the applicant cared about the long-term future.
The likely and potential impact of the project, as well as the particular cost-effectiveness of the project being proposed.
Potential downsides of the project.
Whether the grantee was a good fit for the parameters of this project grant round.
What would happen if BERI did not fund the project right now.

However, note that the committee may put more weight on a different set of factors in the future.

Mistakes & Challenges

There were several kinks in our process (we expected this, given that it was our first individual grants round). Some examples:

We sent out unclear recommendation forms to references, resulting in more back-and-forth about the forms than we had anticipated (and likely leading to fewer references being submitted than otherwise would have).
The project grants committee had difficulties scheduling meetings given summer vacation schedules and other events.
There were a number of internal steps that needed further development during the evaluation process, such as:
- Transforming information gathered by the application form into an easy-to-compare format for committee members,
- Collecting additional information on some applications for committee members, and
- Developing grant agreements and figuring out how to communicate to grantees about the tracking requirements for the grants.

These steps took more time than we anticipated, and developing them as we went (as opposed to ahead of time) likely led to somewhat sub-optimal solutions in many cases.

However, despite the challenges, we're relatively satisfied with how the grants round went—we managed to mostly keep to the schedule we set for ourselves and are excited about our selections. We now have a long list of improvements that we’d like to make for future individual grants rounds.

Grant Winners

Of the 55 applications, we selected 18 to receive funding (~33%). One applicant was awarded a vendor contract in lieu of a grant based on the project's legal structure and the availability of services offered that BERI wanted to purchase (not included in the grants totals BERI is reporting). We managed to stay close to the budget we initially set for this first round ($750k); in total, we plan to award $799,465 to selected winners (excluding the vendor). We are not sure how many applicants or grant winners to expect for future rounds—we predict (with low confidence) that we will receive more applicants, maintain our budget, and therefore accept a smaller fraction of applicants in the future.

We're excited about our grantees, and we hope you are too! We believe that having diverse projects all pursuing work on x-risk as an objective is important to helping develop a culture of focus on humanity's long-term survival and flourishing.

List of winners

2018-09-07: a $50,000 general support grant to the Institute for Philosophical Research (no website).
2018-09-24: a $20,400 grant to Luca Rade to research the implications of coarse-graining by an agent in a complex environment for AI alignment.
2018-09-24: a $1,900 grant to Jordan Alexander to host several meetings at Stanford EA and the Stanford Transhumanist Association.
2018-09-24: a $20,500 grant to Bryce Hidysmith to analyze global risks from technology through a geopolitical lens.
2018-09-26: a $25,000 grant to David Manheim to research aspects of Goodhart’s law, focusing on multi-agent dynamics.
2018-10-02: a $55,000 grant to Baeo Maltinsky to further his research on AI and technology trends.
2018-10-02: a $25,200 grant to Zoe Cremer to support her as a visiting fellow at CFI, where she will research disagreements about the amount and kind of structure required for AGI.
2018-10-02: a $51,000 grant to Colleen McKenzie to support her research on AI timelines and the processes that produce technical and scientific progress.
2018-10-02: a $49,532 grant to Roxanne Heston to work on a variety of AI policy projects in Washington, D.C.
2018-10-02: a $10,000 grant to Cambridge in America for the support of the Leverhulme Centre for the Future of Intellgience.
2018-10-09: a $100,000 grant to Stephanie Zolayvar to train herself in circling and host circles for people who are promising contributors to reducing x-risk.
2018-10-10: a $10,000 grant to Jessica Taylor to work on her research in AI alignment and other areas.
2018-10-10: a $150,000 grant to Ben Goldhaber to support his project (co-lead by Jacob Lagerros) to bring x-risk-relevant questions to popular prediction platforms.
2018-10-15: a $20,000 grant to the Association for the Advancement of Artificial Intelligence for the support of the 2019 conference on AI, Ethics, and Society.
2018-10-18: a $4,000 grant to Effective Altruism Sweden to support Markus Stoor’s project to coordinate two follow-up lunch-to-lunch meetings in Sweden for x-risk-focused individuals.
2018-10-23: a $24,000 grant to Sarah Spikes to implement improvements in support of the Rationality and Effective Altruism Community Hub (REACH).
2018-10-24: a $100,000 grant to Justin Shovelain to lead and develop “Convergence Analysis”, a new group focused on x-risk strategy research.
2018-10-31: a $12,000 grant to Sebastian Farquhar to attend conferences and purchase compute for experiments related to his PhD research on uncertainty modeling in neural networks.