1. I'm not sure if you understand what RCTs can and cannot do.
2. Counterfactual impact of a donation is easy to check if there is significant unfulfilled room for additional donations, as is the case for GiveDirectly.
3. Motivated reasoning is a hell of a drug.
Conversation
2. In 2017, OpenPhil was filling some, but not all, of the GiveWell-assessed funding gap, for reasons that are a bit unclear and involve game theory that makes assessing the true counterfactual impact hard. summarized the confusion at benjaminrosshoffman.com/givewell-and-p. 1/n
2
4
The ID'd opportunities have gotten a lot bigger since 2017, and Good Ventures now gives much larger grants. But the opportunities at GiveDirectly level of ROI are effectively infinite. The GiveWell bar is approximately 5-7X GiveDirectly, and there is a spectrum of opps between.
2
4
18
How much we give is primarily a question of timing (this year or a later year?) rather than a question of game theory.
2
7
By "game theory" I mean GiveWell/OpenPhil making other donors give money by splitting gaps, which is explicitly in blog.givewell.org/2015/11/25/goo.
If this is a question of timing, it's surprising that GiveWell gives this timing recommendation to OpenPhil, but not to other donors.
1
4
Pls see the note at the top of that post which clarifies it quickly became out of date.
Really, pacing is ultimately the donor's decision, inc. in the case of Good Ventures. Concretely, when I see others give to EA, I want to give faster bc I believe there will be more $s later.
1
14
Nothing makes me more conservative than thinking people believe Good Ventures money "covers" all of EA. That is the world in which we must be extremely judicious, bc the opportunities are vast.
2
18
By "conservative" you mean conservative about spending money?
The last sentence makes me unsure I'm reading this right. If people correctly believer that Good Ventures money covers all of EA, wouldn't that mean that there are fewer opportunities out there?
1
3
It makes me more conservative about GV granting money yes. I do *not* think it is a correct belief that Good Ventures money covers all of EA - it is not nearly enough money. If people assume it is, we must be more careful so that it can be allocated to the best opps across time.
1
19
That seems backwards - credibly exhausting your surplus wealth ought to motivate others to step in at least on whichever programs have visibly high-value results.
3
7
"Covers all of EA" doesn't make sense as a literal claim, it has to be some kind of posturing coming from a perspective that's trying to avoid moral liability, not do the most good.
It makes sense that you'd rather not subsidize people doing that & might therefore want to negotiate against them, but that involves misleading third parties taking cost per life saved & funding gap numbers seriously.
2
I agree spending down completely would "fix" the problem. But it's a gamble. In the meantime, spending 2X or 5X or 10X pace doesn't achieve the credibility required, and that's more practically where the decisions arise in practice. (based on currently identified opps)
3
To me, the more obvious question is: if people are genuinely misinformed into thinking your spending covers all of EA (as opposed to doing motivated reasoning to donate less), couldn't you solve that by informing them?
1
7
I have tried. I cannot convince my own friends with hours of convos. I will keep trying.
2
29
Show replies



