I am sceptical of suffering focused utilitarianism[2], and am worried that animal welfare interventions tend to lean strongly in favour of things that reduce the number of animals, on the assumption that their lives are net negative.
You could donate to organisations improving instead of decreasing the lives of animals. I estimated a past cost-effectiveness of Shrimp Welfare Projectâs Humane Slaughter Initiative (HSI) of 43.5 k times the marginal cost-effectiveness of GiveWellâs top charities.
Iâm sceptical of Rethinkâs moral weight numbers[1], and am more convinced of something closer to anchoring on neuron counts (and even more convinced by extreme uncertainty). This puts animal charities more like 10x ahead rather than 1000 or 1 million times. Iâm also sceptical of very small animals (insects) having a meaningful probability/âdegree of sentience.
I agree with the last sentence. Using Rethink Prioritiesâ welfare range for chickens based on neurons, I would conclude corporate campaigns for chicken welfare are 11.1 (= 1.51*10^3*0.00244/â0.332) times as cost-effective as GiveWellâs top charities.
Rethink Prioritiesâ median welfare range for shrimps of 0.031 is 31 k (= 0.031/â10^-6) times their welfare range based on neurons of 10^-6. For you to get to this super low welfare range, you would have to justify putting a very low weight in all the other 11 models considered by Rethink Priorities. In general, justifying a best guess so many orders of magnitude away from that coming out of the most in-depth research on the matter seems very hard.
2) I am sceptical of the intensity of pain and pleasure being logarithmically distributed (severe pain ~100x worse than moderate pain), and especially of this being biased in the negative direction. One reason for this is that I find the âfirst storyâ for interpreting Weberâs law in this post much more intuitive, i.e. that logarithmically distributed stimuli get compressed to a more linear range of experience
Assuming in my cost-effectiveness analysis of HSI that disabling and excruciating pain are as intense as hurtful pain (setting B2 and B3 of tab âTypes of painâ to 1), and maintaining the other assumptions, 1 day of e.g. âscalding and severe burningâ would be neutralised by 1 day of fully healthy life. I think this massively underestimates the badness of severe suffering. Yet, even then, I conclude the past cost-effectiveness of HSI is 2.17 times the marginal cost-effectiveness of GiveWellâs top charities.
I think something like worldview diversification is essentially a reasonable idea, for reasons of risk aversion and optimising under expected future information.
Farmed animals are neglected, so I do not think worldview diversidication would be at risk due to moving 100 M$ to animal welfare instead of global health and development. I calculated 99.9 % of the annual philanthropic spending is on humans.
In contrast, based on Rethink Prioritiesâ median welfare ranges, the annual disability of farmed animals is much larger than that of humans.
Thanks Vasco, I did vote for animal welfare, so on net I agree with most of your points. On some specific things:
You could donate to organisations improving instead of decreasing the lives of animals
This seems right, and is why I support chicken corporate campaigns which tend to increase welfare. Some reasons this is not quite satisfactory:
It feels a bit like a âhelping slaves to live happier livesâ intervention rather than âfreeing the slavesâ
Iâm overall uncertain about whether animals lives are generally net positive, rather than strongly thinking they are
Iâd still be worried about donations to these things generally growing the AW ecosystem as a side effect (e.g. due to fungibility of donations, training up people who then do work with more suffering-focused assumptions)
But these are just concerns and not deal breakers.
Rethink Prioritiesâ median welfare range for shrimps of 0.031 is 31 k (= 0.031/â10^-6) times their welfare range based on neurons of 10^-6. For you to get to this super low welfare range, you would have to justify putting a very low weight in all the other 11 models considered by Rethink Priorities.
I am sufficiently sceptical to put a low weight on the other 11 models (or at least withhold judgement until Iâve thought it through more). As I mentioned Iâm writing a post Iâm hoping to publish this week with at least one argument related to this.
The gist of that post will be: itâs double counting to consider the 11 other models as separate lines of evidence, and similarly double counting to consider all the individual proxies (e.g. âanxiety-like behaviourâ and âfear-like behaviourâ) as independent evidence within the models.
Many of the proxies (I claim most) collapse to the single factor of âdoes it behave as though it contains some kind of reinforcement learning system?â. This itself may be predictive of sentience, because this is true of humans, but I consider this to be more like one factor, rather than many independent lines of evidence that are counted strongly under many different models.
Because of this (a lot of the proxies looking like side effects of some kind of reinforcement learning system), I would expect we will continue to see these proxies as we look at smaller and smaller animals, and this wouldnât be a big update. I would expect that if you look at a nematode worm for instance, it might show:
âTaste-aversion behaviourâ: Moving away from a noxious stimulus, or learning that a particular location contains a noxious stimulus
âDepression-like behaviourâ: Giving up/âputting less energy into exploring after repeatedly failing
âAnxiety-like behaviourâ: Being put on edge or moving more quickly if you expose it to a stimulus which has previously preceded some kind of punishment
âCuriosity-like behaviourâ: Exploring things even when it has some clearly exploitable resource
It might not show all of these (maybe a nematode is in fact too small, I donât know much about them), but hopefully you get the point that these look like manifestations of the same underlying thing such that observing more of them becomes weak evidence once you have seen a few.
Even if you didnât accept that they were all exactly side effects of âa reinforcement learning type systemâ (which seems reasonable), still I believe this idea of there being common explanatory factors for different proxies which are not necessarily sentience related should be factored in.
(RPâs model does do some non-linear weighting of proxies at various points, but not exactly accounting for this thing⌠hopefully my longer post will address this).
On the side of neuron counts, I donât think this is particularly strong evidence either. But I see it as evidence on the side of a factor like âtheir brain looks structurally similar to a humanâsâ, vs the factor of âthey behave somewhat similarly to a humanâ for which the proxies are evidence.
To me neither of these lines of evidence (âbrain structural similarityâ and âbehavioural similarityâ) seems obviously deserving of more weight.
Farmed animals are neglected, so I do not think worldview diversidication would be at risk due to moving 100 M$ to animal welfare
I definitely agree with this, I would only be concerned if we moved almost all funding to animal welfare.
Iâd still be worried about donations to these things generally growing the AW ecosystem as a side effect (e.g. due to fungibility of donations, training up people who then do work with more suffering-focused assumptions)
Without more information, I would guess that funding work on improving rather than decreasing animal lives will at the margin incentivises people to follow the funding, and therefore skill up to work on improving rather than decreasing animal lives.
I am sufficiently sceptical to put a low weight on the other 11 models (or at least withhold judgement until Iâve thought it through more). As I mentioned Iâm writing a post Iâm hoping to publish this week with at least one argument related to this.
I am looking forward to the post. Thanks for sharing the gist and some details. You may want to share a draft with people from Rethink Priorities.
To me neither of these lines of evidence (âbrain structural similarityâ and âbehavioural similarityâ) seems obviously deserving of more weight.
Farmed animals are neglected, so I do not think worldview diversidication would be at risk due to moving 100 M$ to animal welfare instead of global health and development. I calculated 99.9 % of the annual philanthropic spending is on humans.
I think it would be more appropriate to use something like human welfare spending for low-income countries rather than counting ~all charitable activity as in a broad âhumanâ bucket. That is to maintain parity with the way youâve sliced off a particularly effective part of the animal-welfare pie (farmed animal welfare). E.g., some quick Google work suggests animal shelters brought in 3.5B in 2023 in just the US (although a fair portion of of that may be government contracts).
Companion animal shelters may be the animal-welfare equivalent of opera for human-focused charities (spending lots on relatively few individuals who are relatively privileged in a sense). While deciding not to give to farmed-animal charities because of dog shelter spending doesnât make much sense, I would submit that not giving to bednets because of opera spending poses much the same problem.
I donât think that changes your underlying point much at all, though!
Thanks Jason, I would say that giving to animal shelters might be more like giving to the cancer society, or even world vision, rather than opera but thatâs as fairly minor point.
Thanks for sharing your thoughts, Will!
You could donate to organisations improving instead of decreasing the lives of animals. I estimated a past cost-effectiveness of Shrimp Welfare Projectâs Humane Slaughter Initiative (HSI) of 43.5 k times the marginal cost-effectiveness of GiveWellâs top charities.
I agree with the last sentence. Using Rethink Prioritiesâ welfare range for chickens based on neurons, I would conclude corporate campaigns for chicken welfare are 11.1 (= 1.51*10^3*0.00244/â0.332) times as cost-effective as GiveWellâs top charities.
Rethink Prioritiesâ median welfare range for shrimps of 0.031 is 31 k (= 0.031/â10^-6) times their welfare range based on neurons of 10^-6. For you to get to this super low welfare range, you would have to justify putting a very low weight in all the other 11 models considered by Rethink Priorities. In general, justifying a best guess so many orders of magnitude away from that coming out of the most in-depth research on the matter seems very hard.
Assuming in my cost-effectiveness analysis of HSI that disabling and excruciating pain are as intense as hurtful pain (setting B2 and B3 of tab âTypes of painâ to 1), and maintaining the other assumptions, 1 day of e.g. âscalding and severe burningâ would be neutralised by 1 day of fully healthy life. I think this massively underestimates the badness of severe suffering. Yet, even then, I conclude the past cost-effectiveness of HSI is 2.17 times the marginal cost-effectiveness of GiveWellâs top charities.
Farmed animals are neglected, so I do not think worldview diversidication would be at risk due to moving 100 M$ to animal welfare instead of global health and development. I calculated 99.9 % of the annual philanthropic spending is on humans.
In contrast, based on Rethink Prioritiesâ median welfare ranges, the annual disability of farmed animals is much larger than that of humans.
I agree one should not put all resources into the best option, but we are very far from this (see 1st graph above).
Thanks Vasco, I did vote for animal welfare, so on net I agree with most of your points. On some specific things:
This seems right, and is why I support chicken corporate campaigns which tend to increase welfare. Some reasons this is not quite satisfactory:
It feels a bit like a âhelping slaves to live happier livesâ intervention rather than âfreeing the slavesâ
Iâm overall uncertain about whether animals lives are generally net positive, rather than strongly thinking they are
Iâd still be worried about donations to these things generally growing the AW ecosystem as a side effect (e.g. due to fungibility of donations, training up people who then do work with more suffering-focused assumptions)
But these are just concerns and not deal breakers.
I am sufficiently sceptical to put a low weight on the other 11 models (or at least withhold judgement until Iâve thought it through more). As I mentioned Iâm writing a post Iâm hoping to publish this week with at least one argument related to this.
The gist of that post will be: itâs double counting to consider the 11 other models as separate lines of evidence, and similarly double counting to consider all the individual proxies (e.g. âanxiety-like behaviourâ and âfear-like behaviourâ) as independent evidence within the models.
Many of the proxies (I claim most) collapse to the single factor of âdoes it behave as though it contains some kind of reinforcement learning system?â. This itself may be predictive of sentience, because this is true of humans, but I consider this to be more like one factor, rather than many independent lines of evidence that are counted strongly under many different models.
Because of this (a lot of the proxies looking like side effects of some kind of reinforcement learning system), I would expect we will continue to see these proxies as we look at smaller and smaller animals, and this wouldnât be a big update. I would expect that if you look at a nematode worm for instance, it might show:
âTaste-aversion behaviourâ: Moving away from a noxious stimulus, or learning that a particular location contains a noxious stimulus
âDepression-like behaviourâ: Giving up/âputting less energy into exploring after repeatedly failing
âAnxiety-like behaviourâ: Being put on edge or moving more quickly if you expose it to a stimulus which has previously preceded some kind of punishment
âCuriosity-like behaviourâ: Exploring things even when it has some clearly exploitable resource
It might not show all of these (maybe a nematode is in fact too small, I donât know much about them), but hopefully you get the point that these look like manifestations of the same underlying thing such that observing more of them becomes weak evidence once you have seen a few.
Even if you didnât accept that they were all exactly side effects of âa reinforcement learning type systemâ (which seems reasonable), still I believe this idea of there being common explanatory factors for different proxies which are not necessarily sentience related should be factored in.
(RPâs model does do some non-linear weighting of proxies at various points, but not exactly accounting for this thing⌠hopefully my longer post will address this).
On the side of neuron counts, I donât think this is particularly strong evidence either. But I see it as evidence on the side of a factor like âtheir brain looks structurally similar to a humanâsâ, vs the factor of âthey behave somewhat similarly to a humanâ for which the proxies are evidence.
To me neither of these lines of evidence (âbrain structural similarityâ and âbehavioural similarityâ) seems obviously deserving of more weight.
I definitely agree with this, I would only be concerned if we moved almost all funding to animal welfare.
Without more information, I would guess that funding work on improving rather than decreasing animal lives will at the margin incentivises people to follow the funding, and therefore skill up to work on improving rather than decreasing animal lives.
I am looking forward to the post. Thanks for sharing the gist and some details. You may want to share a draft with people from Rethink Priorities.
I find it hard to come up with other proxies.
I think it would be more appropriate to use something like human welfare spending for low-income countries rather than counting ~all charitable activity as in a broad âhumanâ bucket. That is to maintain parity with the way youâve sliced off a particularly effective part of the animal-welfare pie (farmed animal welfare). E.g., some quick Google work suggests animal shelters brought in 3.5B in 2023 in just the US (although a fair portion of of that may be government contracts).
Companion animal shelters may be the animal-welfare equivalent of opera for human-focused charities (spending lots on relatively few individuals who are relatively privileged in a sense). While deciding not to give to farmed-animal charities because of dog shelter spending doesnât make much sense, I would submit that not giving to bednets because of opera spending poses much the same problem.
I donât think that changes your underlying point much at all, though!
Thanks Jason, I would say that giving to animal shelters might be more like giving to the cancer society, or even world vision, rather than opera but thatâs as fairly minor point.