Comments
I was reading through 80,000 Hours's home page, and their "Most pressing problems" section was interesting to me. https://80000hours.org/problem-profiles/
#1 was AI-safety, #2 power concentration, etc...and #8 was global health.
I agree AI safety & factory farming are important...but more important than Global Health? AI safety MIGHT be a problem...whereas we KNOW Global Health is a problem.
Idk if they were intentionally ranking things or not, but they kinda were whether they meant to or not.
Does this actually reflect their thinking? That seems bewildering to me that they might think AI Safety is "more important" than something as imemdiate & tractable as Public Health.
I wrote this in a post, but it might be a better quick take: Would you describe charities within GCR as "super-effective?" They may or may not be good long-term bets, but clearly the outcomes are very uncertain & hard to trace. Whereas GHD + Animal Welfare are probably much more trackable.
I'm asking because I saw on GivingMultiplier.org that they were marketing GCR charities as "highly effective," and that gave me pause.
I don't think so. I'd say they have the potential to be "super-effective in expectation". To me, "super-effective" with no qualifiers implies high confidence.
Agreed. I was surprised that they'd describe something so speculative as expert-vetted, super effective, etc. If something hasn't undergone multiple RCTs, then I don't think it should be promoted as such a highly effective, sure thing.
I think the description is very justifiable, assuming that these charities are indeed good long-term bets.[1]
I agree with you that there is a difference between "a good long-term bet" and "a super-effective charity". However, for most people, it is a really bad idea to share a fully nuanced description of whether a charity is effective based on empirical data vs a high-risk high-reward bet that has a robust theory of change but is operating in a field where empirical data is impossible to collect. While this might be more precise, there are strong tradeoffs between precision and clarity. The full explanation would make most people stop engaging, while the simplified explanation will bring people into the movement.
For many in the EA space (including myself), it can feel dishonest to say simplified claims that are imprecise but still directionally correct, even though no dishonesty is occurring. This is ultimately a matter of context and audience. Describing a well-justified but speculative bet as "highly-effective" as a funder deciding who to support would be bad. As a community builder reaching out to an audience potentially unfamiliar with concepts like expected value, it's much more reasonable.
My general leaning is that most, but certainly not all, GCR charities are not good long-term bets. This poses more fundamental challenges to your point, but the solution would be to increase scrutiny in the GCR space, not fix the marketing. I'm setting this aside for the sake of this comment, as it appears to be a crux you do not share.
Well, the lack of consensus on GCR's efficacy is exactly the thing I'm pointing out. A highly untraceable intervention should not be called "super effective" ... I think that risks undermining their credibility. "Potentially a good long-term bet" is very, very different from "super-effective, expert-vetted."
I was just wondering if there was something I was missing about the proven efficacy of GCR interventions. It sounds like I'm not...in which case, I think it's a really bad idea for GivingMultiplier to pass those off like they have the proven efficacy of GHD interventions. It feels misleading, and I'm more "EA-friendly" than the average person they're probably targeting.
And I have no strong view on GCR interventions. Seems like they'd be hard to prove traceability/impact, but I haven't spend much time thinking about it.
I think there's multiple approaches they could take:
I can see an argument for 4 (i.e. start new donors with more straightforward work). Also, if GCR work is too speculative to justify pursuing, then 4 is clearly the best option.
However, if you believe that GCR work is sufficiently well-justified (as Giving Multiplier seems to), I do think 1 > 2 and 3.
2 and 3 decrease the effectiveness of the organization for the sake of precision. Their FAQ on the home page addresses how they defer to other actors, who make decisions based on the best available evidence. Anyone unsatisfied by that answer has links to Founders Pledge and can learn more about their methodology there.
Super-effective is a marketing label, not a scientific one. They define it reasonably, and I don't think it would come across as dishonest or misleading to someone without EA context.
Fair enough. I have no doubt that they picked the term for marketing purposes. But I'm sure you're right that they would defend it as "highest EV," as opposed to "rigorously studied causal chain."