Sometimes working on animal issues feels like an uphill battle, with alternative protein losing its trendy status with VCs, corporate campaigns hitting blocks in enforcement and veganism being stuck at the same percentage it's been for decades.
"Overall this made me a bit cautious about recommending it as a place to spend time even when someone was sold on ending factory farming."
I would like the animal movement to have the goal of increasing animal welfare, and I think there is a tension between this and ending factory-farming. I believe decreasing factory-farming is harmful to animals due to increasing the suffering of wild animals way more than it decreases the suffering of farmed animals (https://measuredlife.substack.com/p/three-theories-of-change-for-the/comment/127815671). I would be curious to know your thoughts on this.
Hey Vasco, At a certain level of robustness, I do not take CEAs as sufficient evidence to update on, and these estimates do not pass that bar. This post (https://blog.givewell.org/2011/08/18/why-we-cant-take-expected-value-estimates-literally-even-when-theyre-unbiased/ ) is the best articulation of how I think about evidence. I also think other folks, both publicly and privately, have made this argument to you, so I do not plan on going into it in comment threads here, although I appreciate your enthusiasm for helping animals.
I also think about evidence as described in that post, but I still believe the effects on wild animals of interventions targeting farmed animals cannot be neglected. The post explains one should update less given uncertain evidence, but also update more given an uncertain prior. The evidence about the effects on wild animals is very uncertain, but so is the respective prior. In order to neglect the effects on wild animals, one would need a super strong prior that the effects on wild animals are negligible. How do you justify this?
I suspect you and others are misinterpreting the post due to conflating the prior about the effects on humans with the prior about the effects on all animals. The prior about the effects on humans is much stronger than the prior about the effects on all animals because there is much more prior evidence about the effects on humans. Assuming impartiality, one should care about the effects on all animals, not just humans, and therefore rely on the prior about the effects on all animals.
No matter how uncertain the prior, if the evidence is extremely uncertain, it doesn't make sense to act on if doing so comes at the expense of far more certain first order effects (e.g. effects on farmed animals). Especially considering that actually achieving meaningful stuff in the world takes significant time invested moving in a consistent direction, when your views on (e.g.) what's good for wild animals flip flop so often as new bits of weak evidence come in, it's worth waiting to be more certain before starting the slow process of changing the direction of the ship
Are you disagreeing with my mathematical point that having a more uncertain prior implies updating more on new evidence? You can think about this as inverse-variance weighting in meta-analyses (https://en.wikipedia.org/wiki/Inverse-variance_weighting), where effect sizes respecting greater variance are weighted less.
I am not proposing any radical actions to help soil nematodes, mites, and springtails. I recommended advocating for donating to the most cost-effective ways of saving human lives like GiveWell's top charities, and decreasing the uncertainty about the effects on those soil animals (https://forum.effectivealtruism.org/posts/Rjutj7Jd2v2KHvDyA/cost-effectiveness-accounting-for-soil-nematodes-mites-and). I think both of these are better than advocating for supporting farmed animals from my perspective of maximising expected welfare, and from the perspectives you mentioned too. I would say it is easier to sustain efforts to save human lives, and that their 1st order effects are more certain than effects on farmed animals. Decreasing uncertainty also mitigates the problem you referred to of changing top interventions too often. In any case, the cost of changing interventions is pretty minor if you consider my arguments are only expected to change a very small fraction of funding.
Thanks for the post, Joey!
"Overall this made me a bit cautious about recommending it as a place to spend time even when someone was sold on ending factory farming."
I would like the animal movement to have the goal of increasing animal welfare, and I think there is a tension between this and ending factory-farming. I believe decreasing factory-farming is harmful to animals due to increasing the suffering of wild animals way more than it decreases the suffering of farmed animals (https://measuredlife.substack.com/p/three-theories-of-change-for-the/comment/127815671). I would be curious to know your thoughts on this.
Hey Vasco, At a certain level of robustness, I do not take CEAs as sufficient evidence to update on, and these estimates do not pass that bar. This post (https://blog.givewell.org/2011/08/18/why-we-cant-take-expected-value-estimates-literally-even-when-theyre-unbiased/ ) is the best articulation of how I think about evidence. I also think other folks, both publicly and privately, have made this argument to you, so I do not plan on going into it in comment threads here, although I appreciate your enthusiasm for helping animals.
Thanks for the reply, Joey!
I also think about evidence as described in that post, but I still believe the effects on wild animals of interventions targeting farmed animals cannot be neglected. The post explains one should update less given uncertain evidence, but also update more given an uncertain prior. The evidence about the effects on wild animals is very uncertain, but so is the respective prior. In order to neglect the effects on wild animals, one would need a super strong prior that the effects on wild animals are negligible. How do you justify this?
I suspect you and others are misinterpreting the post due to conflating the prior about the effects on humans with the prior about the effects on all animals. The prior about the effects on humans is much stronger than the prior about the effects on all animals because there is much more prior evidence about the effects on humans. Assuming impartiality, one should care about the effects on all animals, not just humans, and therefore rely on the prior about the effects on all animals.
No matter how uncertain the prior, if the evidence is extremely uncertain, it doesn't make sense to act on if doing so comes at the expense of far more certain first order effects (e.g. effects on farmed animals). Especially considering that actually achieving meaningful stuff in the world takes significant time invested moving in a consistent direction, when your views on (e.g.) what's good for wild animals flip flop so often as new bits of weak evidence come in, it's worth waiting to be more certain before starting the slow process of changing the direction of the ship
Thanks for the comment, Aidan!
Are you disagreeing with my mathematical point that having a more uncertain prior implies updating more on new evidence? You can think about this as inverse-variance weighting in meta-analyses (https://en.wikipedia.org/wiki/Inverse-variance_weighting), where effect sizes respecting greater variance are weighted less.
I am not proposing any radical actions to help soil nematodes, mites, and springtails. I recommended advocating for donating to the most cost-effective ways of saving human lives like GiveWell's top charities, and decreasing the uncertainty about the effects on those soil animals (https://forum.effectivealtruism.org/posts/Rjutj7Jd2v2KHvDyA/cost-effectiveness-accounting-for-soil-nematodes-mites-and). I think both of these are better than advocating for supporting farmed animals from my perspective of maximising expected welfare, and from the perspectives you mentioned too. I would say it is easier to sustain efforts to save human lives, and that their 1st order effects are more certain than effects on farmed animals. Decreasing uncertainty also mitigates the problem you referred to of changing top interventions too often. In any case, the cost of changing interventions is pretty minor if you consider my arguments are only expected to change a very small fraction of funding.