Yeah, I guess I mean genuinely new projects, rather than new tokens of the same type of project (eg group organisers are running the same thing in different places).
As MacAskill points out, it’s pretty hard to run $1m+/yr project (or even less, tbh) without Open Philanthropy supporting it.
But, no, I’m not thinking about centralisation in terms of micro management, so I don’t follow your comment. You can have centralised power without micromanagent.
What does it mean to have centralized power without micromanagement? Like I could theoretically force a group organizer to use a different font, I just choose not to?
To take one of the top examples in the post’s centralization continuum, presumably the US military counts as having a highly centralized power structure despite the President and Secretary of Defense not micromanaging. People lower on the food chain exercise power delegated and re-delegated from those two, but they are the ultimate fount of power.
They have the right to control—and responsibility to supervise—the powers they have delegated downward. With some uncommon arguable exceptions like military judges, one in the military has or exercises power independently of POTUS and SECDEF. And people know that if they use their delegated power in ways that would anger those higher up, they won’t have that power for too much longer.
That’s how power ordinarily works in larger centralized contexts; big-company CEOs
refusing to delegate font-approval authority is very much the exception.
Thanks! Just to double check: this is agreeing with my example, right? Like you are pointing out that the US military is centralized because POTUS could theoretically tell a random private what to eat for lunch (but chooses not to), similar to how I could theoretically[1] force a group organizer to use a different font (I just choose not to)?
If so: I would be surprised if the average private says things like “I’m eating potatoes for lunch because it’s impossible to do projects without buy-in from POTUS”? I agree there some technical sense in which that’s true, but a new recruit who oriented to their work by thinking they can only do stuff by convincing POTUS would probably struggle to navigate the military, similar to how I claim that an EA who believes they can only do stuff by convincing “senior EAs” is going to struggle to navigate EA (even under the assumption that EA is as centralized as the military).
Whether you could get someone nominally ‘under’ you to do an arbitrary thing is not a good proxy for power.
CEA is a regular hierarchical company, but it would still go very poorly if you decided to, on a power trip, tell one of your employees what to eat for lunch. This mostly doesn’t matter, though, because that is a goal you are very unlikely to have.
As a co-organizer of the Boston Meetup, if you sent me an email demanding that we serve potatoes at the next gathering, I would be very confused. But you could get CEA’s groups team to come up with guidance on meetup food, heavily influence that process, and I could then receive an email advocating serving potatoes from people I trusted and who I was pretty sure had thought about it a lot more than I had. Which would have a decent chance of resulting in potatoes at the next meetup.
Power is always, in a technical sense, indirect: no one is pulling levers inside other people’s heads to get them to do things. There is always some amount of inspiration, persuasion, threat, or other intermediary. Sometimes this is formalized, sometimes “soft”, but that mostly only matters for legibility. Maybe a better measurement for power is something like, if there’s something important about the way things are currently done that you want to change, how likely and how much are you able to cause that change?
By that measure, OP has a tremendous amount of power: through a combination of employing highly respected people and having control over the funding of most EA work they can make large and deep changes to how the EA movement grows and what work is carried out under its banner.
Thanks! This feels like a reasonable definition, but seems different from what Michael was talking about? He said:
unless you have ‘buy-in’ from one of the listed ‘senior EAs’, it is very hard to get traction or funding for your project (I speak from experience). In that sense, EA feels quite like a big, conventional organisation.
If CEA tried to push some dietary standard I’m pretty sure there would be a ton of complaints and blowback. But even if we somehow kept going through all of that, I’m pretty sure you would still be able to run a potato-less meet up, which doesn’t feel consistent with the “need buy-in” claim.
(Whereas in “big, conventional organizations” if the CEO says “the cafeteria is going to serve potatoes” then the chefs don’t have much of a choice.)
Yeah, I guess I mean genuinely new projects, rather than new tokens of the same type of project (eg group organisers are running the same thing in different places).
As MacAskill points out, it’s pretty hard to run $1m+/yr project (or even less, tbh) without Open Philanthropy supporting it.
But, no, I’m not thinking about centralisation in terms of micro management, so I don’t follow your comment. You can have centralised power without micromanagent.
What does it mean to have centralized power without micromanagement? Like I could theoretically force a group organizer to use a different font, I just choose not to?
To take one of the top examples in the post’s centralization continuum, presumably the US military counts as having a highly centralized power structure despite the President and Secretary of Defense not micromanaging. People lower on the food chain exercise power delegated and re-delegated from those two, but they are the ultimate fount of power.
They have the right to control—and responsibility to supervise—the powers they have delegated downward. With some uncommon arguable exceptions like military judges, one in the military has or exercises power independently of POTUS and SECDEF. And people know that if they use their delegated power in ways that would anger those higher up, they won’t have that power for too much longer.
That’s how power ordinarily works in larger centralized contexts; big-company CEOs refusing to delegate font-approval authority is very much the exception.
Thanks! Just to double check: this is agreeing with my example, right? Like you are pointing out that the US military is centralized because POTUS could theoretically tell a random private what to eat for lunch (but chooses not to), similar to how I could theoretically[1] force a group organizer to use a different font (I just choose not to)?
If so: I would be surprised if the average private says things like “I’m eating potatoes for lunch because it’s impossible to do projects without buy-in from POTUS”? I agree there some technical sense in which that’s true, but a new recruit who oriented to their work by thinking they can only do stuff by convincing POTUS would probably struggle to navigate the military, similar to how I claim that an EA who believes they can only do stuff by convincing “senior EAs” is going to struggle to navigate EA (even under the assumption that EA is as centralized as the military).
I actually don’t think I could do this
Whether you could get someone nominally ‘under’ you to do an arbitrary thing is not a good proxy for power.
CEA is a regular hierarchical company, but it would still go very poorly if you decided to, on a power trip, tell one of your employees what to eat for lunch. This mostly doesn’t matter, though, because that is a goal you are very unlikely to have.
As a co-organizer of the Boston Meetup, if you sent me an email demanding that we serve potatoes at the next gathering, I would be very confused. But you could get CEA’s groups team to come up with guidance on meetup food, heavily influence that process, and I could then receive an email advocating serving potatoes from people I trusted and who I was pretty sure had thought about it a lot more than I had. Which would have a decent chance of resulting in potatoes at the next meetup.
Power is always, in a technical sense, indirect: no one is pulling levers inside other people’s heads to get them to do things. There is always some amount of inspiration, persuasion, threat, or other intermediary. Sometimes this is formalized, sometimes “soft”, but that mostly only matters for legibility. Maybe a better measurement for power is something like, if there’s something important about the way things are currently done that you want to change, how likely and how much are you able to cause that change?
By that measure, OP has a tremendous amount of power: through a combination of employing highly respected people and having control over the funding of most EA work they can make large and deep changes to how the EA movement grows and what work is carried out under its banner.
Thanks! This feels like a reasonable definition, but seems different from what Michael was talking about? He said:
If CEA tried to push some dietary standard I’m pretty sure there would be a ton of complaints and blowback. But even if we somehow kept going through all of that, I’m pretty sure you would still be able to run a potato-less meet up, which doesn’t feel consistent with the “need buy-in” claim.
(Whereas in “big, conventional organizations” if the CEO says “the cafeteria is going to serve potatoes” then the chefs don’t have much of a choice.)