I agree that if you choose at random from EA org and non-EA org jobs, you are more likely to have more impact at an EA org job. And I agree that there is work involved in finding a high impact non-EA job.
However, I don’t think the work involved in finding a high impact non-EA org job is hard because there are so few such opportunities out there, but because finding them requires more imagination/creativity than just going for a job at an EA org does. Maybe you could start a new AI safety team at Facebook or Amazon by joining, building the internal capital, and then proposing it. Maybe you can’t because reasons. Either way, you learn by trying. And this learning is not wasted. Either you pave the way for others in the community, highlighting a new area where impact can be made. Or, if it turns out it’s hard for reasons, then you’ve learnt why, and can pass that on to others who might try.
Needless to say this impact finding strategy scales better than one where everyone is exclusively focused on EA org jobs (although you need some of that too). On a movement scale, I’d make a bet that we’re too far in the direction of thinking that EA orgs is a better path to impact and have significantly under-explored ways of making impact in non-EA orgs, and there are social reasons why we’d naturally bias in that direction. Alternatively, like Sarah said elsewhere, it’s just less visible.
I just realised I haven’t asked—why are high impact non-EA org jobs are hard to find, in your view?
I agree that if you choose at random from EA org and non-EA org jobs, you are more likely to have more impact at an EA org job. And I agree that there is work involved in finding a high impact non-EA job.
However, I don’t think the work involved in finding a high impact non-EA org job is hard because there are so few such opportunities out there, but because finding them requires more imagination/creativity than just going for a job at an EA org does. Maybe you could start a new AI safety team at Facebook or Amazon by joining, building the internal capital, and then proposing it. Maybe you can’t because reasons. Either way, you learn by trying. And this learning is not wasted. Either you pave the way for others in the community, highlighting a new area where impact can be made. Or, if it turns out it’s hard for reasons, then you’ve learnt why, and can pass that on to others who might try.
Needless to say this impact finding strategy scales better than one where everyone is exclusively focused on EA org jobs (although you need some of that too). On a movement scale, I’d make a bet that we’re too far in the direction of thinking that EA orgs is a better path to impact and have significantly under-explored ways of making impact in non-EA orgs, and there are social reasons why we’d naturally bias in that direction. Alternatively, like Sarah said elsewhere, it’s just less visible.
I just realised I haven’t asked—why are high impact non-EA org jobs are hard to find, in your view?