Centre for Enabling EA Learning and Research (CEEALAR)
A centre for talented individuals to upskill, collaborate and work on AI safety, existential risk and other EA cause areas.
CEEALAR is a research and career development hub that supports EAs to achieve high impact goals by providing:
- Financial assistance (free/subsidised accommodation, vegan food, stipends)
- Networking opportunities
- A productive workspace
- Learning and development opportunities
~80% of our current grantees are working on X-risks (mostly AI Safety) and we expect it to be at least that high going forward. It costs us just $1,400 to fund an individual's work for one month compared to an average cost of $5,500 to support an EA independent researcher.* This means we are a highly cost-effective way of funding more research on AI safety.
Our grantees are more productive and do higher impact work thanks to CEEALAR. Many would not be able to transition to EA work without the financial security we offer. See our Theory of Change for more detail.
We select our grantees through a thorough vetting process, and to date have supported 20,900 person-days working at our centre. We plan to grow our impact by supporting more people and by increasing the support we offer. For example we plan to:
- Complete renovations on an additional building (recently donated to us by Greg Colbourn), increasing our capacity from 15 grantees to 26.
- Launch an alumni network, enabling our grantees and alumni to support each other to amplify their impact
- Further increase our grantees' productivity by investing in ergonomic workstations, coaching support and workshops
*Based on our own fermi estimate: 76 grants were made to independent researchers over Q4 of 2022 and Q1 of 2023. The average cost per month of work was $5,500 (over 3.5 times our cost) and the median was $4,400 (over 3 times our cost). https://docs.google.com/spreadsheets/d/1QGRJ10iASrboPnaDt8ZeJrYVYQ5Ub1xKf1yX3Ul413g/edit#gid=1434805434 ; If we compared our costs to the average salary for independent researchers in developed countries only, we expect we are even more cost-effective by comparison, however it is difficult to find data for this.
General comments on the project.