I disagree voted but did not downvote. The reason is that while I find the behavior in this TIME article incredibly alarming and abhorrent, I do not think that centralization of EA contributes to the behavior nor do I find decentralization would help (EDIT: though actually this is oveconfident, see caveat added below). In fact, my rough guess is that centralization is fairly helpful here by allowing harassing behavior to be known and being able to effectively gatekeep events etc. from harassers (EDIT: though actually this is an overgeneralization, see caveat added below).
I do however find it clear this article points to some EAs being incredibly unthoughtful about power dynamics to a dangerous degree. I would like to see that fixed. I recommend reading Julia Wise’s article on the issue. People who abuse power dynamics in this way have no place in the EA movement I want to promote and would have no place at Rethink Priorities.
~
Edited to add some caveats after discussing this with two people independently: I do think that there are both important downsides and upsides to consider for each of more marginal centralization and more marginal decentralization. Most notably hearing about “the fact that the women would only speak under conditions of anonymity due to EA’s centralisation of power over funding and employment in a few (overwhelmingly male) hands” is very alarming and suggests I underrate how bad centralization can be.
I can certainly think of other ways this would be way worse if things were even more centralized (like imagine there was no independent Rethink Priorities or other groups and CEA controlled literally 100% of the jobs… also centralized groups can better protect and shield high-profile abusers) but I can also think of ways decentralization makes things worse (e.g., in my very limited past experience I’ve personally witnessed decentralization make an abuser be able to move easily from small group to small group without people knowing). This makes me think most that broad generalizations about centralization vs. decentralization are not the most productive axis of analysis here for solving the problem.
Thank you for the thoughtful reply! I think the kind of debate that you indicated to have had is exactly what we need to make sense of such emotionally difficult and complex topics.
Even if you come to the conclusion that other framings seem more useful to you, we can’t have confidence in such conclusions if they are made ex ante without deliberate engagement with the content. So thanks for doing that!
I personally think that we may need to take more time to really try to understand and explore the problem we are facing here, before focusing on solutions. I have the feeling that something broader than „isolated incidents“ of sexual harassment is the right way to frame the problem here. There have been „community-related“ issues after issues popping up over the last few months. We should step back and try to look at the whole picture here and try to understand the mechanisms and drivers that lead to such events. I think this is happening to some degree but it still feels like we could be doing this more explicitly and openly. I really think there’s a lot at stake regarding the future development of the movement.
I can also think of ways decentralization makes things worse (e.g., in my very limited past experience I’ve personally witnessed decentralization make an abuser be able to move easily from small group to small group without people knowing).
As an example of this, I’m involved in the contra dance community which is almost entirely organized at the level of individual dance groups. I can think of several examples of people who moved on to a different group after getting kicked out of their original group, sometimes multiple times.
Some of this is that the contra dance world doesn’t have any group with the role of CEA’s community health team (disclosure: my wife is on that team), but even if it did have one it would be very hard to coordinate with the hundreds of local groups around the country. In a decentralized system it’s very hard to share information about bad actors to the people who need to know it without making it essentially public, and making things public is often a large escalation that people who report problems to you don’t want.
But those are two very different communities / movements and I don’t think that the situations are similar. As you said, there is something like CEA and the EA movement also has the ambition to act in a somewhat coordinated fashion to solve the world’s biggest problems, whereas dance groups grow like wild flowers wherever and whenever enough people interested in dance come together regularly.
I am not saying that there is nothing to learn by comparing these different situations but this doesn’t seem to be an argument against the theory that centralization of power could have somehow contributed to creating an environment in which people felt badly treated or even harassed. Rather it seems to be more of an illustration that preventing such behaviors is a really difficult problem regardless of concentration of power.
My point here was that the conclusion that can be drawn from your example is orthogonal to the question of how concentrated power is. Your example did not provide much evidence against the claim that concentration of power may be a contributing factor to the issue here. Feel free to reread my prior comment.
I disagree voted but did not downvote. The reason is that while I find the behavior in this TIME article incredibly alarming and abhorrent, I do not think that centralization of EA contributes to the behavior nor do I find decentralization would help (EDIT: though actually this is oveconfident, see caveat added below). In fact, my rough guess is that centralization is fairly helpful here by allowing harassing behavior to be known and being able to effectively gatekeep events etc. from harassers (EDIT: though actually this is an overgeneralization, see caveat added below).
I do however find it clear this article points to some EAs being incredibly unthoughtful about power dynamics to a dangerous degree. I would like to see that fixed. I recommend reading Julia Wise’s article on the issue. People who abuse power dynamics in this way have no place in the EA movement I want to promote and would have no place at Rethink Priorities.
~
Edited to add some caveats after discussing this with two people independently: I do think that there are both important downsides and upsides to consider for each of more marginal centralization and more marginal decentralization. Most notably hearing about “the fact that the women would only speak under conditions of anonymity due to EA’s centralisation of power over funding and employment in a few (overwhelmingly male) hands” is very alarming and suggests I underrate how bad centralization can be.
I can certainly think of other ways this would be way worse if things were even more centralized (like imagine there was no independent Rethink Priorities or other groups and CEA controlled literally 100% of the jobs… also centralized groups can better protect and shield high-profile abusers) but I can also think of ways decentralization makes things worse (e.g., in my very limited past experience I’ve personally witnessed decentralization make an abuser be able to move easily from small group to small group without people knowing). This makes me think most that broad generalizations about centralization vs. decentralization are not the most productive axis of analysis here for solving the problem.
Thank you for the thoughtful reply! I think the kind of debate that you indicated to have had is exactly what we need to make sense of such emotionally difficult and complex topics.
Even if you come to the conclusion that other framings seem more useful to you, we can’t have confidence in such conclusions if they are made ex ante without deliberate engagement with the content. So thanks for doing that!
I personally think that we may need to take more time to really try to understand and explore the problem we are facing here, before focusing on solutions. I have the feeling that something broader than „isolated incidents“ of sexual harassment is the right way to frame the problem here. There have been „community-related“ issues after issues popping up over the last few months. We should step back and try to look at the whole picture here and try to understand the mechanisms and drivers that lead to such events. I think this is happening to some degree but it still feels like we could be doing this more explicitly and openly. I really think there’s a lot at stake regarding the future development of the movement.
Sorry, this got meta pretty quickly…
As an example of this, I’m involved in the contra dance community which is almost entirely organized at the level of individual dance groups. I can think of several examples of people who moved on to a different group after getting kicked out of their original group, sometimes multiple times.
Some of this is that the contra dance world doesn’t have any group with the role of CEA’s community health team (disclosure: my wife is on that team), but even if it did have one it would be very hard to coordinate with the hundreds of local groups around the country. In a decentralized system it’s very hard to share information about bad actors to the people who need to know it without making it essentially public, and making things public is often a large escalation that people who report problems to you don’t want.
But those are two very different communities / movements and I don’t think that the situations are similar. As you said, there is something like CEA and the EA movement also has the ambition to act in a somewhat coordinated fashion to solve the world’s biggest problems, whereas dance groups grow like wild flowers wherever and whenever enough people interested in dance come together regularly.
I am not saying that there is nothing to learn by comparing these different situations but this doesn’t seem to be an argument against the theory that centralization of power could have somehow contributed to creating an environment in which people felt badly treated or even harassed. Rather it seems to be more of an illustration that preventing such behaviors is a really difficult problem regardless of concentration of power.
That’s a pretty different claim from the “this is what happens when you centralize power so much” that started this thread!
My point here was that the conclusion that can be drawn from your example is orthogonal to the question of how concentrated power is. Your example did not provide much evidence against the claim that concentration of power may be a contributing factor to the issue here. Feel free to reread my prior comment.