The assumption I had is we defer a lot of power, both intellectual, social and financial, to a small group of broadly unaccountable, non-transparent people on the assumption they are uniquely good at making decisions, noticing risks to the EA enterprise and combatting them, and that this unique competence is what justifies the power structures we have in EA.
Is this actually true right now? People donating to EA Funds seem like an example of deferring financial decisions, but I don’t have data how EAs donate to the Funds vs. decide themselves where to donate.
Or do you mean decisions like relying on GiveWell recommendations as an example of ‘deferring financial power’?
I am also not sure how the EA Community compares to other movements. Is your claim that EA is worse at this than comparable movements or that we should hold ourselves to a higher standard?
I have mixed feelings about your post overall. If people defer decision-making power to “the leadership” then it’s good to ask these questions. But mostly I see individuals making decisions for themselves. If others think the decisions are bad, they don’t have to admire “the leadership” for it.
The vast bulk of funds in EA (OpenPhil and, until last week, FTX Future Fund) are controlled by very few people (financial). As is admission to EA Global (social). Intellectual direction is more open with e.g. the EA Forum, but things like big book projects and their promotion (The Precipice, WWOTF) are pretty centralised, as is media engagement in general.
The FTX Future Fund had a large regranter program. They didn’t fully let regranters do whatever they wanted with funds, but I think it’s incorrect to say that it’s controlled by very few people.
Ultimately the Future Fund had veto power over regranters (even those with their own pots), [edit:] so I think it’s inaccurate to say that the regranters had control of the funds (influence, sure; but not control).
I’m somewhat perturbed by the ratio of karma on these comments (esp agree karma; although low sample size—mine has only 1 vote on agreement (5 votes on karma); see pic below for time of writing this comment)[1]. We’ve just found out that we’ve in general been way too trusting as a community, and could do with more oversight etc (although I guess it’s open to discussion how much decentralisation of decision making is ideal; see below). The fact that regranters could influence the Future Fund on their grantmaking was great, but we shouldn’t confuse that with actual control. What ultimately matters is what is true from a mechanistic legal perspective—where the buck actually stops, and who is actually in charge of authorising grants. For the Future Fund, that was 5 people (who presumably in turn could still have been vetoed by the 4 on the board).
The next step for a regranting program in terms of actually distributing control would be to actually give the regranters the money, to do whatever they saw fit with it. I can imagine many people screaming in horror at the thought, especially those in central positions who think that they are the best experts on avoiding the unilateralists curse, but that illusion has been shattered now. I have to say that writing this, I’m torn in that I still think the unilateralists curse is a big problem, given the vulnerable world hypothesis etc. I don’t know what the solution is.
Although maybe it’s in part due to the fact that Neel is mentioning the regranting program that I didn’t mention (I would’ve been better mentioning it explicitly in my original comment, but pointing out that regranters didn’t have control, perhaps in a footnote); in which case, fine.
I was reached out to by a regranter and got the vibe immediately that they were stressed about providing grants that might be accepted and basically just optimizing for what they perceived to be the most likely things for the team to give the ok.
Now again I only talked to one person but if they were just shooting ideas at the team to be processed similar to how they were processing general apps the regranter program serves more as a marketing tool to increase applicants and a slight filter of awful apps than it does change who has the power. I would be very interested to say the data on how many regrants were given / how many regrants were suggested compared to the normal funds.
I was a regranter. I did not have my own pot, but could make recommendations for grants. 52% of my regrants (11/21) were approved (32% by $ value). I understand that those with their own pots allocated to them had a lower bar for acceptance so probably had a better success rate for approvals.
I’ve been trying to find people willing and able to write quality books and have had a hard time finding anyone. “Doing Doing Good Better Better” seems one of the highest-EV projects, and EA Funds (during my tenure) received basically no book proposals, as far as I can remember. I’d love to help throw a lot of resources after an upcoming book project by someone competent who isn’t established in the community yet.
Even the forum is organised so as to promote posts from people with large networks of high-upvoted people, which de facto means that core network of people pretty much get auto-highlighted for posting their shopping list.
Is this actually true right now? People donating to EA Funds seem like an example of deferring financial decisions, but I don’t have data how EAs donate to the Funds vs. decide themselves where to donate. Or do you mean decisions like relying on GiveWell recommendations as an example of ‘deferring financial power’?
I am also not sure how the EA Community compares to other movements. Is your claim that EA is worse at this than comparable movements or that we should hold ourselves to a higher standard?
I have mixed feelings about your post overall. If people defer decision-making power to “the leadership” then it’s good to ask these questions. But mostly I see individuals making decisions for themselves. If others think the decisions are bad, they don’t have to admire “the leadership” for it.
The vast bulk of funds in EA (OpenPhil and, until last week, FTX Future Fund) are controlled by very few people (financial). As is admission to EA Global (social). Intellectual direction is more open with e.g. the EA Forum, but things like big book projects and their promotion (The Precipice, WWOTF) are pretty centralised, as is media engagement in general.
The FTX Future Fund had a large regranter program. They didn’t fully let regranters do whatever they wanted with funds, but I think it’s incorrect to say that it’s controlled by very few people.
Ultimately the Future Fund had veto power over regranters (even those with their own pots), [edit:] so I think it’s inaccurate to say that the regranters had control of the funds (influence, sure; but not control).
I’m somewhat perturbed by the ratio of karma on these comments (esp agree karma; although low sample size—mine has only 1 vote on agreement (5 votes on karma); see pic below for time of writing this comment)[1]. We’ve just found out that we’ve in general been way too trusting as a community, and could do with more oversight etc (although I guess it’s open to discussion how much decentralisation of decision making is ideal; see below). The fact that regranters could influence the Future Fund on their grantmaking was great, but we shouldn’t confuse that with actual control. What ultimately matters is what is true from a mechanistic legal perspective—where the buck actually stops, and who is actually in charge of authorising grants. For the Future Fund, that was 5 people (who presumably in turn could still have been vetoed by the 4 on the board).
The next step for a regranting program in terms of actually distributing control would be to actually give the regranters the money, to do whatever they saw fit with it. I can imagine many people screaming in horror at the thought, especially those in central positions who think that they are the best experts on avoiding the unilateralists curse, but that illusion has been shattered now. I have to say that writing this, I’m torn in that I still think the unilateralists curse is a big problem, given the vulnerable world hypothesis etc. I don’t know what the solution is.
Although maybe it’s in part due to the fact that Neel is mentioning the regranting program that I didn’t mention (I would’ve been better mentioning it explicitly in my original comment, but pointing out that regranters didn’t have control, perhaps in a footnote); in which case, fine.
I was reached out to by a regranter and got the vibe immediately that they were stressed about providing grants that might be accepted and basically just optimizing for what they perceived to be the most likely things for the team to give the ok.
Now again I only talked to one person but if they were just shooting ideas at the team to be processed similar to how they were processing general apps the regranter program serves more as a marketing tool to increase applicants and a slight filter of awful apps than it does change who has the power. I would be very interested to say the data on how many regrants were given / how many regrants were suggested compared to the normal funds.
I was a regranter. I did not have my own pot, but could make recommendations for grants. 52% of my regrants (11/21) were approved (32% by $ value). I understand that those with their own pots allocated to them had a lower bar for acceptance so probably had a better success rate for approvals.
I’ve been trying to find people willing and able to write quality books and have had a hard time finding anyone. “Doing Doing Good Better Better” seems one of the highest-EV projects, and EA Funds (during my tenure) received basically no book proposals, as far as I can remember. I’d love to help throw a lot of resources after an upcoming book project by someone competent who isn’t established in the community yet.
I wrote a quick shortform post.
Yes, this is brilliant.
Even the forum is organised so as to promote posts from people with large networks of high-upvoted people, which de facto means that core network of people pretty much get auto-highlighted for posting their shopping list.
Yea I’m not really sure why the default isn’t democratic voting, with the option to toggle karma-weighted voting if you want.