I am the Principal Research Director at Rethink Priorities. I lead our Surveys and Data Analysis department and our Worldview Investigation Team.
The Worldview Investigation Team previously completed the Moral Weight Project and CURVE Sequence / Cross-Cause Model. We're currently working on tools to help EAs decide how they should allocate resources within portfolios of different causes, and to how to use a moral parliament approach to allocate resources given metanormative uncertainty.
The Surveys and Data Analysis Team primarily works on private commissions for core EA movement and longtermist orgs, where we provide:
Formerly, I also managed our Wild Animal Welfare department and I've previously worked for Charity Science, and been a trustee at Charity Entrepreneurship and EA London.
My academic interests are in moral psychology and methodology at the intersection of psychology and philosophy.
Survey methodology and data analysis.
Sure, there can be differences in talent, but the question ought to be... how tangible is this difference and does it justify the cost of hiring?
One relevant data point is that in last year's Meta Coordination Forum Survey, respondents estimated the gap in value between their first and second most preferred candidate to be:
However, you can see from the distribution that in many cases valuations are significantly higher than the average.
And note that this is their valuation of the gap between the first and second after they've conducted extensive evaluations. The gap between their most preferred candidate after selection and the candidate they would get after not undertaking such an extensive evaluation process (perhaps their 10th or worse most preferred candidate) could be much higher.
As a more general point, it seems worth noting that the costs of sub-optimal hiring can be very high. I wouldn't be surprised if the costs to management and other staff time of even slightly sub-optimal hires would be >60 hours a year, with a long tail of much, much worse outcomes (including potentially just having to rehire), which is high even relative to quite intensive hiring rounds.
taking the Rethink Priorities 7 - 15% numbers at face value, when the arguments for those AFAICT don't even have particular models behind them. This is a pretty crazy sounding number that needs way better argumentation than "a poll of people said so", and here it's just asserted without much commentary at all.
I'm confused by this statement. The welfare range estimates aren't based on a "poll" and are based on numerous "particular models."
I strongly prefer, and recommend, arguing dispassionately and hyper-charitably.
But, it seems relevant that Lance's posts are responding to posts that also seem highly polemic[1] (Lance outlines why he views the posts as such at the start of his first post). It seems more appropriate (whether or not instrumentally advisable) to respond in kind if that's the established norm of the space (different from Forum norms).
For example, though not the main focus of the essay, the BB's first essay includes:
Cultural Relativism: Crazy, Illogical, and Accepted by no One Except Philosophically Illiterate Gender Studies Majors...
Problem: it’s obviously false...
People often accept cultural relativism because they’re vaguely confused and want to be tolerant.
It's at least possible that one can 'contain' debunking arguments, such that they don't extend across domains and self-undermine. We discuss this strategy in our chapter here.
Also in his original formulation "high status" environments are often simply nicer
Agreed. I think this is another important confound.
People's concern for relative status may seems clearer when we consider cases of 'moving up' into an area of people who are relatively wealthier, i.e. even if the environment were materially much nicer, I think most people will find it very salient if they are the only non-wealthy person there.
I don't think that Caplan's test is a good one, for a couple of reasons that commenters on his original post pointed out:
Thanks Ivan!
What stood out most was the idea that prioritization functions not only as analysis but also as a form of governance
I agree with this. And I think this framing makes clear why how we allocate the community's prioritisation is such an important question.
How much of the prioritization in EA is a design question about institutional learning, decision rights, and oversight?
From that angle, the current emphasis on within-cause work doesn’t just feel like a strategic imbalance; it may also reflect what’s easier to operationalize within existing organizational structures.
I also agree with this. As we allude to in the piece, institutional infrastructure for (and generally doing) within-cause prioritisation is generally easier: you can build on, or more easily develop, networks of domain-specialists and specialist institutions. And I think various factors push the community towards more siloed within-cause structures (e.g. network effects etc.).
So I think it's both the case that within-cause infrastructure is easier to set up and that, as you say, our current (heavily cause-specific) infrastructure makes within-cause prioritisation easier and cross-cause prioritisation harder (e.g. there are few institutions that are well-placed or have the remit to do cross-cause work).
I agree that we would need more structured systems (or more support for the existing systems) in order to do more cross-cause prioritisation. I don't want to communicate fatalism about this though: I think existing organizations and individuals could start doing significantly more cross-cause prioritisation if they decided it were valuable and that this would itself make it easier to build the relevant infrastructure.[1]
Though, of course, it would take further work for EA's actual allocations of resources to be influenced by this prioritisation work.
In principle this could be driven, in part, by scepticism about the absolute valuation of EA hires, rather than the relative valuation of hires.
Do you have a sense of what the percentage difference is between the typical first and second most preferred hire (i.e. the first most preferred hire is X% more impactful than the second), and what you think the absolute $ difference is?