On average, respondents estimated that EA resources over the next 5 years should go 45.2% to Longtermist (including GCR) causes, 23.8% to Near-term animal-focused causes, and 22.6% to Near-term human-focused causes.
Evaluating more fine-grained causes, respondents estimated that 26.8% should go to work focused on AI, 15.7% to Global health, 14.7% to Farm animal welfare, 10.2% to building EA and related communities, and 8.2% to Biosecurity (in addition to smaller percentages to many other causes).
Have you considered asking about the marginal cost-effectiveness of each of the areas as a fraction of that of GiveWell’s unrestricted funding? I think it would be valuable to ask about this, and funding allocation (as above) to Rethink Priorities’ (RP’s) worldview investigations team. I see their members as the closest we have to impartial experts, so I guess their views may influence the cause prioritisation of donors. I am more pessimistic about their views impacting the cause prioritisation of organisations, which often have little unrestricted funding they could direct to the causes they deem most cost-effective at the margin. To save time, you could ask each member of the team to spend at most 10 min answering a few questions, and then publish the results in an unpolished way on EA Forum, flagging it is not up to RP’s usual standards (as disclaimed in some other RP’s publications).
Have you considered asking about the marginal cost-effectiveness of each of the areas as a fraction of that of GiveWell’s unrestricted funding? I think it would be valuable to ask about this, and funding allocation (as above) to Rethink Priorities’ (RP’s) worldview investigations team. I see their members as the closest we have to impartial experts, so I guess their views may influence the cause prioritisation of donors. I am more pessimistic about their views impacting the cause prioritisation of organisations, which often have little unrestricted funding they could direct to the causes they deem most cost-effective at the margin. To save time, you could ask each member of the team to spend at most 10 min answering a few questions, and then publish the results in an unpolished way on EA Forum, flagging it is not up to RP’s usual standards (as disclaimed in some other RP’s publications).