I noticed that part of me feels defensive of EA when there are criticisms. That part says something like “but EA is doing massively better than most social movements regarding steering”. But of course, that can be true while at the same time being true that “EA needs to become better at steering and invest marginal efforts into it rather than rowing”.
An interesting question is “how sensitive are outcomes to steering vs. rowing?” (Are we navigating towards a continent over open sea, or towards a treasure hidden between treacherous cliffs?) because that informs how obsessed we should be with either.
the absolute value of an outcome isn’t necessarily in proportion to the amount of power: it can be the case that if we don’t achieve a certain amount of power, key decision will go uninfluenced/wrong (e.g. if AGI decisions will be made centrally, we need A LOT of power to influence that)
I’d be curious to hear examples of movements/institutions that are/have been really good at steering, and how EA compares
I’m very interested in seeing more work and debate around what the optimal distribution of power in EA should be. I can see arguments both for narrow concentration and wide concentration being better for steering.
If I can expand the metaphor, I think EA is better thought of as a fleet than a single boat. This allows better for concepts such as chain of command, localized decision making, spreading out vs. staying together. One failure mode this could identify is “spreading out so much that inevitably some boats will do tremendous harm”
Good post!
Some thoughts I had:
I noticed that part of me feels defensive of EA when there are criticisms. That part says something like “but EA is doing massively better than most social movements regarding steering”. But of course, that can be true while at the same time being true that “EA needs to become better at steering and invest marginal efforts into it rather than rowing”.
An interesting question is “how sensitive are outcomes to steering vs. rowing?” (Are we navigating towards a continent over open sea, or towards a treasure hidden between treacherous cliffs?) because that informs how obsessed we should be with either.
the absolute value of an outcome isn’t necessarily in proportion to the amount of power: it can be the case that if we don’t achieve a certain amount of power, key decision will go uninfluenced/wrong (e.g. if AGI decisions will be made centrally, we need A LOT of power to influence that)
I’d be curious to hear examples of movements/institutions that are/have been really good at steering, and how EA compares
I’m very interested in seeing more work and debate around what the optimal distribution of power in EA should be. I can see arguments both for narrow concentration and wide concentration being better for steering.
If I can expand the metaphor, I think EA is better thought of as a fleet than a single boat. This allows better for concepts such as chain of command, localized decision making, spreading out vs. staying together. One failure mode this could identify is “spreading out so much that inevitably some boats will do tremendous harm”