Thanks for clarifying! Should your aversion to variance in possible outcomes be a reason for not recommending the Shrimp Welfare Project (SWP)? The probability of shrimps being sentient presented in Bob Fischer’s book about comparing welfare across species is 40 %. For a 75 % chance that SWP benefits shrimps conditional on their sentience, the probability of SWP benefiting shrimps is 30 % (= 0.40*0.75). However, the benefits to shrimp could be negligible even if they are sentient. Ambitious Impact’s (AIM’s) estimates of suffering-adjusted days (SADs), which you used to estimate the benefits of SWP, are proportional to the (expected) welfare range, the difference between the maximum and minimum welfare per unit time, and their welfare range of shrimps is 19.9 % (= 0.433*0.460) of that of humans for a probability of sentience of 43.3 %, and welfare range conditional on sentience of 46.0 % of that of humans (you can ask Vicky Cox for the sheet). However, I would say a much better best guess could be10^-6 of that of humans, the ratio between the number of neurons of shrimps and humans. For a probability of significant benefits conditional on their sentience lower than 1⁄3, the probability of SWP significantly benefiting shrimps would be lower than 10 % (= 0.3*1/3). I would also say the probability of SWP significantly benefiting shrimps is very uncertain, and you mentioned an aversion to unknown probabilities.
Thanks for clarifying! Should your aversion to variance in possible outcomes be a reason for not recommending the Shrimp Welfare Project (SWP)? The probability of shrimps being sentient presented in Bob Fischer’s book about comparing welfare across species is 40 %. For a 75 % chance that SWP benefits shrimps conditional on their sentience, the probability of SWP benefiting shrimps is 30 % (= 0.40*0.75). However, the benefits to shrimp could be negligible even if they are sentient. Ambitious Impact’s (AIM’s) estimates of suffering-adjusted days (SADs), which you used to estimate the benefits of SWP, are proportional to the (expected) welfare range, the difference between the maximum and minimum welfare per unit time, and their welfare range of shrimps is 19.9 % (= 0.433*0.460) of that of humans for a probability of sentience of 43.3 %, and welfare range conditional on sentience of 46.0 % of that of humans (you can ask Vicky Cox for the sheet). However, I would say a much better best guess could be 10^-6 of that of humans, the ratio between the number of neurons of shrimps and humans. For a probability of significant benefits conditional on their sentience lower than 1⁄3, the probability of SWP significantly benefiting shrimps would be lower than 10 % (= 0.3*1/3). I would also say the probability of SWP significantly benefiting shrimps is very uncertain, and you mentioned an aversion to unknown probabilities.