Error

Another way to frame it is thinking about Marr’s three levels of analysis. The computational (what are we even trying to do?), the algorithmic (what algorithms/​heuristics should we run given we want to accomplish that?), and implementation (what, concretely should our next actions be to implement those algorithms in reality?). Cleanly separating which step you are working on prevents confusion.

I think this framing is a good one, but I don’t immediately agree with the conclusion you make about which level to prioritize.

Firstly, consider the benefits we expect from a change in someone’s view at each level. Do most people stand to improve their impact the most by choosing the best implementation within their cause area, or switching to an average implementation in a more pressing cause area? I don’t think this is obvious, but I lean to the latter.

Higher levels are more generalizable: cross-implementation comparisons are only relevant to people within that cause, whereas cross-cause comparisons are relevant to everyone who shares approximately the same values, so focusing on lower levels limits the size of the audience that can benefit from what you have to say.

Low-level comparisons tend to require domain-specific expertise, which we won’t be able to have across a wide range of domains.

I also think there’s just a much greater deficit of high-quality discussion of the higher levels. They’re virtually unexamined by most people. Speaking personally, my introduction to EA was approximately that I knew I was confused about the medium-level question, so I was directly looking for answers to that: I’m not sure a good discussion of the low-level question would have captured me as effectively.

Thank you. I commonly try to say something at a “high-level” (such as the difference between relative and absolute/​extreme poverty). Now, instead, I will mention something about distributing mosquito bed nets, steel roofs in Kenya (GiveDirectly) or developing clean meat. I anticipate some questions on that last one :)

I want to add something: It probably has been discussed before, but it occurs to me that when thinking about prioritisation in general it’s almost always better to think at the lowest level possible. That’s because the impact per dollar is only evaluable for specific interventions, and because causes that at first don’t appear particularly cost effective can hide particular interventions that are. And those particular interventions could be in principle even more cost effective than other interventions in causes that do appear cost effective overall. I think high-level cause prioritisation is mostly good for gaining a first superficial understanding of the promise of a particular class of altruistic interventions.