(obviously I have a bias here, but didn’t talk to anyone at WAI before leaving this comment, nor do I have any formal affiliation with them except via marriage).
I feel confused by most of this critique. To be clear, I think founding new charities is incredibly impactful, and there are great funding opportunities there, and that funding AIM charities at launch is a great opportunity for EA donors focused on GHD and animal welfare. But to address your concerns:
1. This critique just sounds like you wouldn’t accept any scientific results, and presumably this is also an issue for all AIM charities / animal welfare charities. If you’re worried about the replication crisis (obviously an issue, but much less so in bio/ecology than social sciences), then presumably every AIM-intervention based on scientific research should be viewed with a ton of skepticism too? Why should I think Shrimp Welfare Project, for example, has good reason for thinking their interventions are good except for through scientific research, which by your lights looks like should be dismissed? AIM’s work is backed by tons of science which presumably would also fail this test. It seems like if you want to make this critique, you’d need special reasons for thinking that WAI funded work is less likely to replicate than other animal welfare science work. I think the opposite is likely to be true, given that WAI seems highly focused on enforcing good scientific practice compared to the average funder?
I also think that this misunderstands what I take WAI’s theory of change to be — I don’t think the idea is that people will listen to scientists—it’s that there will be more scientific evidence so future interventions can have a bigger body of evidence behind them.
Finally, I think the timelines for WAW interventions are a lot shorter than you expect—I’ve been writing something up about this, but when WAI launched, I expected it to take decades to get to a place where we could be highly confident about the sign of WAW interventions. Now, I’d guess that we are under a decade away for some.
Finally, I’ll note that this critique has a common failure mode for discussing WAW — if you acknowledge that WAW is an issue, but are worried about the tractability due to flowthrough effects of interventions, than this should make you think basically all charitable interventions are less tractable, because they’ll all have unknown flowthrough effects on WAW. I think this is a bad line of reasoning, but at its core “WAW is a problem but intractable” seems like just as big a problem for other cause areas as WAW.
2. I don’t think this is accurate. I think it is true that AIM founded charities pay much less, and so do many startups. But WAI has been around several more years than the oldest AIM animal charities and is significantly more established. I think I’d expect most AIM charities that survive to having 20+ staff, etc to raise their salaries significantly, with AIM itself being an exception to this due to a particular ideological commitment.
Also, WAI’s salaries appear to be well below or competitive with many other EA animal charities, like THL or GFI.
But, I also think that this could just be a strategic disagreement — clearly outside the animal space, many EA charities believe people should be paid much more than AIM pays. I suspect many people do not apply to AIM roles due to the pay. That could be a worthwhile tradeoff, but paying lower salaries isn’t something that is all upside—it makes the talent pool smaller, and in many spaces, that could be directly bad for impact.
AIM incubated ~30 charities in the time WAI published 4-12 papers (for ~10 total citations) and attended some conferences.
This seems like it intentionally left off WAI’s main program and activity / the source of the vast majority of their impact? A little confused about why you didn’t include it and it makes me unclear how much to update on this comment or to take your critique overall. For reasons outlined below, I also think the AIM statistic is also misleading, because not all those charities were successful or impactful — very few of them were/are (which is still great, and I think on net, donating to the right new charity could be better than donating to WAI).
They seem several orders of magnitude less cost-effective than a field leading charity
Could you share the calculations you did for this or was this just a guess/made up figure? I’d be interested to see it since you seem confident and gave a pretty specific figure here, and this outcome is very different from other rigorous estimates I’ve seen in the space by independent parties.
***
Finally, as a high-level point, comparing founding a charity with an established charity is not as straightforward as looking at the dollars to each. If you think WAI does cost-effective work (which clearly many EA-aligned evaluators think, like EA Funds, OpenPhil, ACE), then you’re comparing a risky bet against a reliable outcome. Say AIM has a 1 in 5 success rate for founding impactful charities, but it’s hard to tell which of the new charities at founding will be successful. I think this would be incredibly successful and is way above where they currently are. But that means if each one needed $100k to start, you spend $500k overall to get the $100k in value from the good charity. The charity would need to generate 5x the value of WAI in this scenario to make funding a new charity from the pot over WAI the right choice.
Founding new charities is obviously a great way to have impact. But lots of new charities are bad, and by your lights donations to them will be wasted when the charity tries their intervention and it doesn’t work, or otherwise fails and shuts down. The comparison you should be making is more like: donate X to WAI with a known impact (Z progress on some issue), or donate X to a risky bet with a lot of upside, and a high likelihood (in my view >80% for AIM charities—still incredibly impressive and a testament to AIM) of going to $0. That’s much less obviously a clear bet one way or another in the abstract.
Time is precious, so I’m just going to commit to ending my involvement after this comment.
BLUF: I’m not trying to convey “AIM good, WAI bad”. Just that “WAI does not appear to be unbeatable and therefore other charities in this space might be better uses of funding” AIM was just an example of a charity which would be very hard.
This critique just seems weirdly anti-science and presumably an issue for all AIM charities / animal welfare charities. you’re worried about the replication crisis… then presumably every AIM-intervention based on scientific research should be viewed with a ton of skepticism too?
I didn’t claim there was anything wrong with the scientific method. I claimed there was something wrong with academia. Apologies if I am misinterpreting you, but you seem to imply that other charities are around equally as dependent on academia. I think that’s clearly false.
Quotes from the post
“Our primary goal is to support the growth of a self-sustaining interdisciplinary research community focused on reducing wild animal suffering.”
“Here are some highlights of our progress to date… Establishing that wild animal welfare science is a serious academic endeavor, including by… helping to get the NYU Wild Animal Welfare Program established.”
Regarding other charities based on scientific research, you can largely tell that they’re working within a year or two of them launching. The evaluation doesn’t require academics. It doesn’t depend on anyone taking published research seriously; the interventions help regardless of whether or not anyone takes the evaluation seriously.
The rest of your writing here seems to assume that I accept that wild animal welfare considerations dominate all others. I don’t come close to believing that.
I don’t think this is accurate. I think it is true that AIM founded charities pay much less, and so do many startups. But WAI has been around several more years than the oldest AIM animal charities and is significantly more established. I think I’d expect most AIM charities that survive to having 20+ staff, etc to raise their salaries significantly, with AIM itself being an exception to this due to a particular ideological commitment.
I don’t think that every AIM charity is field-leading. I’m trying to communicate that for Jason’s case to hold true, crudely paraphrased, that when fields funding is scarce we should rally around its incumbents), the incumbent charities need to be hard to beat. AIM is one example of a charity I think is field leading in it’s area. I’m trying to show that WAI is not nearly as hard to beat because the TOC seems a lot weaker and the cost is a lot higher.
This seems like it intentionally left off WAI’s main program and activity / the source of the vast majority of their impact? A little confused about why you didn’t include it and it makes me unclear how much to update on this comment or to take your critique overall
Based on the post, I haven’t read much further into their work, it seems that most their value cashes out in published research papers, either of them or their grantees. I included both in the statistics cited above, even preprints. AIM could equally complain that I’ve left off their research, for-profit incubation program, and researcher incubation program.
“WAI does not appear to be unbeatable and therefore other charities in this space might be better uses of funding”
I was basing my comment on you asserting WAI was several orders of magnitude less cost-effective then other charities in the space. I absolutely agree that WAI might be less cost-effective than other groups, but this claim is a lot less extreme than the first one you made. I’d still love to see your estimate if you have it, because I’d appreciate a more critical lens on WAI’s theory of change than I’ve seen before.
Regarding other charities based on scientific research, you can largely tell that they’re working within a year or two of them launching. The evaluation doesn’t require academics. It doesn’t depend on anyone taking published research seriously; the interventions help regardless of whether or not anyone takes the evaluation seriously.
...
I didn’t claim there was anything wrong with the scientific method. I claimed there was something wrong with academia. Apologies if I am misinterpreting you, but you seem to imply that other charities are around equally as dependent on academia. I think that’s clearly false.
I think this is interesting but probably incorrect — while other charities don’t have interventions that involve academia, any kind of claim about their effectiveness, and any ability they have to develop interventions is heavily reliant on it. We can’t tell at all if they are working without academics. Basically everything EA affiliated groups have done to help animals (cage-free campaigns, alternative proteins, welfare reforms, etc) have relied incredibly heavily on academia—they just happen to work on the other side of the academic research than WAI.
Taking Shrimp Welfare Project as an example (primarily because I suspect from a neartermist lens, SWP is far and away the most cost-effective animal charity, likely more cost-effective than WAI, and a great opportunity for donors right now too), they exclusively do welfare interventions. We literally would have no idea if they are helping animals without animal welfare science (like the science WAI funds). We might know that their interventions impacted a lot of animals, but we wouldn’t know the sign or degree of it at all. The only reason we have any idea that SWP, for example, is so impactful, is because of academic research. And SWP was only able to make determinations on what interventions to do based on that science. If you’re uncertain that science can be trusted as you express in your initial point, you’d have to throw this out all. The issues in science you point to are very very real. But that doesn’t mean we can’t make progress on animal issues based on scientific research, and that doing science well to support future interventions isn’t important.
Final comment—I’ve edited by initial comment to better reflect what I wanted to convey, but left the original text so the context of your reply remains for later readers.
Relatedly: People who have the skill set to found new EA orgs well, plus the ability and willingness to accept AIM-level pay, are not in unlimited supply. There’s a significant opportunity cost to putting the hypothetical WAW-startup founder in a WAW startup vs. (e.g.) an AIM startup. Conditioned on them being skilled enough to do well in a serious WAW funding downturn, they probably would have been top-tier talent elsewhere.
(obviously I have a bias here, but didn’t talk to anyone at WAI before leaving this comment, nor do I have any formal affiliation with them except via marriage).
I feel confused by most of this critique. To be clear, I think founding new charities is incredibly impactful, and there are great funding opportunities there, and that funding AIM charities at launch is a great opportunity for EA donors focused on GHD and animal welfare. But to address your concerns:
1. This critique just sounds like you wouldn’t accept any scientific results, and presumably this is also an issue for all AIM charities / animal welfare charities. If you’re worried about the replication crisis (obviously an issue, but much less so in bio/ecology than social sciences), then presumably every AIM-intervention based on scientific research should be viewed with a ton of skepticism too? Why should I think Shrimp Welfare Project, for example, has good reason for thinking their interventions are good except for through scientific research, which by your lights looks like should be dismissed? AIM’s work is backed by tons of science which presumably would also fail this test. It seems like if you want to make this critique, you’d need special reasons for thinking that WAI funded work is less likely to replicate than other animal welfare science work. I think the opposite is likely to be true, given that WAI seems highly focused on enforcing good scientific practice compared to the average funder?
I also think that this misunderstands what I take WAI’s theory of change to be — I don’t think the idea is that people will listen to scientists—it’s that there will be more scientific evidence so future interventions can have a bigger body of evidence behind them.
Finally, I think the timelines for WAW interventions are a lot shorter than you expect—I’ve been writing something up about this, but when WAI launched, I expected it to take decades to get to a place where we could be highly confident about the sign of WAW interventions. Now, I’d guess that we are under a decade away for some.
Finally, I’ll note that this critique has a common failure mode for discussing WAW — if you acknowledge that WAW is an issue, but are worried about the tractability due to flowthrough effects of interventions, than this should make you think basically all charitable interventions are less tractable, because they’ll all have unknown flowthrough effects on WAW. I think this is a bad line of reasoning, but at its core “WAW is a problem but intractable” seems like just as big a problem for other cause areas as WAW.
2. I don’t think this is accurate. I think it is true that AIM founded charities pay much less, and so do many startups. But WAI has been around several more years than the oldest AIM animal charities and is significantly more established. I think I’d expect most AIM charities that survive to having 20+ staff, etc to raise their salaries significantly, with AIM itself being an exception to this due to a particular ideological commitment.
Also, WAI’s salaries appear to be well below or competitive with many other EA animal charities, like THL or GFI.
But, I also think that this could just be a strategic disagreement — clearly outside the animal space, many EA charities believe people should be paid much more than AIM pays. I suspect many people do not apply to AIM roles due to the pay. That could be a worthwhile tradeoff, but paying lower salaries isn’t something that is all upside—it makes the talent pool smaller, and in many spaces, that could be directly bad for impact.
This seems like it intentionally left off WAI’s main program and activity / the source of the vast majority of their impact? A little confused about why you didn’t include it and it makes me unclear how much to update on this comment or to take your critique overall. For reasons outlined below, I also think the AIM statistic is also misleading, because not all those charities were successful or impactful — very few of them were/are (which is still great, and I think on net, donating to the right new charity could be better than donating to WAI).
Could you share the calculations you did for this or was this just a guess/made up figure? I’d be interested to see it since you seem confident and gave a pretty specific figure here, and this outcome is very different from other rigorous estimates I’ve seen in the space by independent parties.
***
Finally, as a high-level point, comparing founding a charity with an established charity is not as straightforward as looking at the dollars to each. If you think WAI does cost-effective work (which clearly many EA-aligned evaluators think, like EA Funds, OpenPhil, ACE), then you’re comparing a risky bet against a reliable outcome. Say AIM has a 1 in 5 success rate for founding impactful charities, but it’s hard to tell which of the new charities at founding will be successful. I think this would be incredibly successful and is way above where they currently are. But that means if each one needed $100k to start, you spend $500k overall to get the $100k in value from the good charity. The charity would need to generate 5x the value of WAI in this scenario to make funding a new charity from the pot over WAI the right choice.
Founding new charities is obviously a great way to have impact. But lots of new charities are bad, and by your lights donations to them will be wasted when the charity tries their intervention and it doesn’t work, or otherwise fails and shuts down. The comparison you should be making is more like: donate X to WAI with a known impact (Z progress on some issue), or donate X to a risky bet with a lot of upside, and a high likelihood (in my view >80% for AIM charities—still incredibly impressive and a testament to AIM) of going to $0. That’s much less obviously a clear bet one way or another in the abstract.
Time is precious, so I’m just going to commit to ending my involvement after this comment.
BLUF: I’m not trying to convey “AIM good, WAI bad”. Just that “WAI does not appear to be unbeatable and therefore other charities in this space might be better uses of funding” AIM was just an example of a charity which would be very hard.
I didn’t claim there was anything wrong with the scientific method. I claimed there was something wrong with academia. Apologies if I am misinterpreting you, but you seem to imply that other charities are around equally as dependent on academia. I think that’s clearly false.
Quotes from the post
“Our primary goal is to support the growth of a self-sustaining interdisciplinary research community focused on reducing wild animal suffering.”
“Here are some highlights of our progress to date… Establishing that wild animal welfare science is a serious academic endeavor, including by… helping to get the NYU Wild Animal Welfare Program established.”
If that doesn’t make it sufficiently clear, view their grants page; Almost all of their grants went to academics.
Regarding other charities based on scientific research, you can largely tell that they’re working within a year or two of them launching. The evaluation doesn’t require academics. It doesn’t depend on anyone taking published research seriously; the interventions help regardless of whether or not anyone takes the evaluation seriously.
The rest of your writing here seems to assume that I accept that wild animal welfare considerations dominate all others. I don’t come close to believing that.
I don’t think that every AIM charity is field-leading. I’m trying to communicate that for Jason’s case to hold true, crudely paraphrased, that when fields funding is scarce we should rally around its incumbents), the incumbent charities need to be hard to beat. AIM is one example of a charity I think is field leading in it’s area. I’m trying to show that WAI is not nearly as hard to beat because the TOC seems a lot weaker and the cost is a lot higher.
Based on the post, I haven’t read much further into their work, it seems that most their value cashes out in published research papers, either of them or their grantees. I included both in the statistics cited above, even preprints. AIM could equally complain that I’ve left off their research, for-profit incubation program, and researcher incubation program.
I was basing my comment on you asserting WAI was several orders of magnitude less cost-effective then other charities in the space. I absolutely agree that WAI might be less cost-effective than other groups, but this claim is a lot less extreme than the first one you made. I’d still love to see your estimate if you have it, because I’d appreciate a more critical lens on WAI’s theory of change than I’ve seen before.
I think this is interesting but probably incorrect — while other charities don’t have interventions that involve academia, any kind of claim about their effectiveness, and any ability they have to develop interventions is heavily reliant on it. We can’t tell at all if they are working without academics. Basically everything EA affiliated groups have done to help animals (cage-free campaigns, alternative proteins, welfare reforms, etc) have relied incredibly heavily on academia—they just happen to work on the other side of the academic research than WAI.
Taking Shrimp Welfare Project as an example (primarily because I suspect from a neartermist lens, SWP is far and away the most cost-effective animal charity, likely more cost-effective than WAI, and a great opportunity for donors right now too), they exclusively do welfare interventions. We literally would have no idea if they are helping animals without animal welfare science (like the science WAI funds). We might know that their interventions impacted a lot of animals, but we wouldn’t know the sign or degree of it at all. The only reason we have any idea that SWP, for example, is so impactful, is because of academic research. And SWP was only able to make determinations on what interventions to do based on that science. If you’re uncertain that science can be trusted as you express in your initial point, you’d have to throw this out all. The issues in science you point to are very very real. But that doesn’t mean we can’t make progress on animal issues based on scientific research, and that doing science well to support future interventions isn’t important.
Final comment—I’ve edited by initial comment to better reflect what I wanted to convey, but left the original text so the context of your reply remains for later readers.
Relatedly: People who have the skill set to found new EA orgs well, plus the ability and willingness to accept AIM-level pay, are not in unlimited supply. There’s a significant opportunity cost to putting the hypothetical WAW-startup founder in a WAW startup vs. (e.g.) an AIM startup. Conditioned on them being skilled enough to do well in a serious WAW funding downturn, they probably would have been top-tier talent elsewhere.