Obviously I’m the opposite of an expert here but here are my reasons, roughly from most important to least important
1. I think the best assessment we have of Animal sentience seems biased towards animals for at least 4 reasons as I outlined here. So I take RP numbers and downward multiply them by something like 10x − 1,000x depending on the animal. IMO the most important bias here was selecting a pro animal-welfare research team with zero animal welfare skeptics.
2. I’ve been generally unimpressed by responses to criticisms of animal sentience. I’ve rarely seen an animal welfare advocate make an even small concession. This makes me even more skeptical about the neutrality of thought processes and research done by animal welfare folks.
3. I don’t believe that response-to-stimuli is a very useful proxy for sentience, yet this is what sentience percentages are usually based on. Every organism responds to stimuli, including bacteria. They have to in order to survive, in increasingly complex ways as organism complexity increases. I doubt that strong noxious stimuli response = meaningful pain. I think response-to-stimuli is the go-to because we just don’t have a good alternative, but that doesn’t mean the “best” option is a good one.
I think that brains need a “special” set of conditions to be sentient (see point 4), and to feel pain in a meaningful way. Bees are wildly intelligent and have capabilities way beyond humans in some areas, but I don’t think probability of sentience increases roughly linearly with increased complexity of behaviour.
4. I think we should base our priors more on what we are more sure of, which is our understanding of human sentience. I don’t think humans are actually “sentient” until we are quite old (see Peter Singer’s old work). I have zero memories before age like 3 or 4? Whether a form of sentience is there in the womb or some time after birth, the human brain is way more complex than animals before sentience kicks in. Brain EEGs only show meaningful stimulation at 30 weeks of gestation, and likely it kicks in sometime after that. I think this is an underrated data point, because we have far more confidence in it than any of our pretty wild assertions about animals. To be clear I’m not a pure utilitarian and think humans AND animals have inherent worth outside of sentience and pain.
I’ll note that even despite all this I still think Animal Welfare work is hugely important, and even on debate week I voted that marginal money would be slightly better going there than to GHD. Even if my sentience estimate is 1,000x less than yours the work is still likely super important and cost effective.
Out of curiosity, why so low?
straight after a short spray, the bees vacated the roof. There might have been s lot more due later though but they looked not bad.
I meant why the low probability of bee sentience.
Obviously I’m the opposite of an expert here but here are my reasons, roughly from most important to least important
1. I think the best assessment we have of Animal sentience seems biased towards animals for at least 4 reasons as I outlined here. So I take RP numbers and downward multiply them by something like 10x − 1,000x depending on the animal. IMO the most important bias here was selecting a pro animal-welfare research team with zero animal welfare skeptics.
https://forum.effectivealtruism.org/posts/E9NnR9cJMM7m5G2r4/is-rp-s-moral-weights-project-too-animal-friendly-four
2. I’ve been generally unimpressed by responses to criticisms of animal sentience. I’ve rarely seen an animal welfare advocate make an even small concession. This makes me even more skeptical about the neutrality of thought processes and research done by animal welfare folks.
3. I don’t believe that response-to-stimuli is a very useful proxy for sentience, yet this is what sentience percentages are usually based on. Every organism responds to stimuli, including bacteria. They have to in order to survive, in increasingly complex ways as organism complexity increases. I doubt that strong noxious stimuli response = meaningful pain. I think response-to-stimuli is the go-to because we just don’t have a good alternative, but that doesn’t mean the “best” option is a good one.
I think that brains need a “special” set of conditions to be sentient (see point 4), and to feel pain in a meaningful way. Bees are wildly intelligent and have capabilities way beyond humans in some areas, but I don’t think probability of sentience increases roughly linearly with increased complexity of behaviour.
4. I think we should base our priors more on what we are more sure of, which is our understanding of human sentience. I don’t think humans are actually “sentient” until we are quite old (see Peter Singer’s old work). I have zero memories before age like 3 or 4? Whether a form of sentience is there in the womb or some time after birth, the human brain is way more complex than animals before sentience kicks in. Brain EEGs only show meaningful stimulation at 30 weeks of gestation, and likely it kicks in sometime after that. I think this is an underrated data point, because we have far more confidence in it than any of our pretty wild assertions about animals. To be clear I’m not a pure utilitarian and think humans AND animals have inherent worth outside of sentience and pain.
I’ll note that even despite all this I still think Animal Welfare work is hugely important, and even on debate week I voted that marginal money would be slightly better going there than to GHD. Even if my sentience estimate is 1,000x less than yours the work is still likely super important and cost effective.
Knowing Matt, I’d wager he’s asking why your p(bee sentience) is so low