Yeah, thanks. I flip back and forth too. Currently thinking a little bit more extreme in the “useful later” path than you, in that I think I’m gonna make a bet on a skillset that isn’t correlated with the rest of the movement.
It is my anecdotal sense, modulo that the type of people you meet at conferences are selected a particular way, is that EA is overexposed to some impatient theories of change (mostly to do with passive impact, support roles or community building). Furthermore, I think there’s a pressure in this direction because it is easier than hard math/cs, alignment research takes a lot of emotional resilience, facing your limitations, failing often, etc. I worry that we’re deeply underexposed to “Alice is altruistically motivated, can think for herself about population ethics and threatmodels, and conveniently is an expert in super niche xyz with a related out-of-movement friendgroup who can help her with some project that wasn’t on OpenPhil’s radar 5 or even 2 years ago”.
Yeah, EA is probably underrating uncorrelated skillsets and breadth of in-movement expertise.
Yeah, thanks. I flip back and forth too. Currently thinking a little bit more extreme in the “useful later” path than you, in that I think I’m gonna make a bet on a skillset that isn’t correlated with the rest of the movement.
It is my anecdotal sense, modulo that the type of people you meet at conferences are selected a particular way, is that EA is overexposed to some impatient theories of change (mostly to do with passive impact, support roles or community building). Furthermore, I think there’s a pressure in this direction because it is easier than hard math/cs, alignment research takes a lot of emotional resilience, facing your limitations, failing often, etc. I worry that we’re deeply underexposed to “Alice is altruistically motivated, can think for herself about population ethics and threatmodels, and conveniently is an expert in super niche xyz with a related out-of-movement friendgroup who can help her with some project that wasn’t on OpenPhil’s radar 5 or even 2 years ago”.
Yeah, EA is probably underrating uncorrelated skillsets and breadth of in-movement expertise.