Thanks for the reply.

Firstly, it should be noted that the overall ratio used for the 2025 SADs was 1000x not 7x.

Right. There was a weight of 45 % on a ratio of 7.06, and of 55 % on one of 62.8 k (= 3.44*10^6/54.8), 8.90 k (= 62.8*10^3/7.06) times as much. My explanation for the large difference is that very little can be inferred about the intensity of excruciating pain, as defined by the Welfare Footprint Institute (WFI), from the academic studies AIM analysed to derive the pain intensities linked to the lower ratio.

Just as an example this study on 37 women who have just gone through labour, roughly one third of them would prefer a 9/10 pain for 2 hours than a 1/10 pain for 18 hours!

The study is not relevant for assessing excruciating pain? Excruciating pain is "not normally tolerated even if only for a few seconds". Here is the clarification of what this means from Cynthia Schuck, WFI's scientific director.

By not tolerated we mean that Excruciating pain can't be sustained for long (e.g., hours, as opposed to minutes) without neurological shutdown. It will tend to overrides an organism's ability to function or respond coherently, with even powerful opioids providing minimal relief, or desperate attempts to escape pain even at the risk of death . Examples would include severe burning in large areas of the body, dismemberment, or extreme torture.

I doubt the women in the study would be indifferent between 2 h of "severe burning in large areas of the body, dismemberment, or extreme torture" and 18 h of a "1/10 pain". I believe excruciating pain is way more intense than their "9/10 pain" assuming they are indifferent between 2 h of "9/10 pain" and 18 h of "1/10 pain".

From the study, "In the questionnaire, the intensity of pain was evaluated using an NRS 0–10, with 0=no pain and 10=worst pain imaginable". However, this does not imply the women's "10/10 pain" was excruciating. I guess their "10/10 pain" was disabling as defined by WFI.

Finally, I defend putting at least some weight on counter-intuitive results of academic research. I especially defend this in the case where you are analysing and pooling the results of many papers and expect some results to be bias upwards and some results to be bias downwards. The new SAD spreadsheet links to 15 different studies / pieces of evidence on this topic. Of those 15 some of which show counterintuitively low and some counterintuitively high relative preferences for different levels of pain.

I think point estimates like AIM's SADs derived from aggregating such different results have very little robustness. The takeaway for me is that we have basically no idea about which pain intensities to use when they differ so much. I believe this calls for a more robust estimation method and input research, not for aggregating more widely different results, although I still expect large uncertainty will remain (just not so large).

@vicky_cox, has AIM has considered commissioning surveys asking random people, people who regularly experience disabling pain, and people who have experienced excruciating pain about how they trade-off WFI's pain and pleasure categories. I believe Rethink Priorities' (RP's) surveys and data analysis team would be a good fit to run such surveys. @Vince Mak 🔸, has ACE considered commissioning such surveys?

Thanks, Wladimir. That makes sense. I look forward to your future work on this. Let me know if funding ever becomes bottleneck, in which case I may want to help with a few k$.

Here is the crosspost on the EA Forum. Rob preferred I shared it myself.

The critical question is whether shrimp or insects can support the kinds of negative states that make suffering severe, rather than merely possible.

I think suffering matters proportionally to its intensity. So I would not neglect mild suffering in principle, although it may not matter much in practice due to contributing little to total expected suffering.

In any case, I would agree the total expected welfare of farmed invertebrates may be tiny compared with that of humans due invertebrates' experiences having a very low intensity. For expected individual welfare per fully-healthy-animal-year proportional to "individual number of neurons"^"exponent", and "exponent" from 0.5 to 1.5, which I believe covers reasonable best guesses, I estimate that the expected total welfare of farmed shrimps ranges from -0.282 to -2.82*10^-7 times that of humans, and that of farmed black soldier fly (BSF) larvae and mealworms from -4.80*10^-4 to -6.23*10^-11 times that of humans. In addition, I calculate the Shrimp Welfare Project’s (SWP’s) Humane Slaughter Initiative (HSI) has increased the welfare of shrimps 0.00167 (= 2.06*10^5/0.0123) to 1.67 k (= 20.6/0.0123) times as cost-effectively as GiveWell's top charities increase the welfare of humans.

And even granting the usual EA filters—tractability, neglectedness, feasibility, and evidential robustness—the scale gradient from shrimp to insects (via agriculture-related deaths) is so steep that these filters don’t, by themselves, explain why the precautionary logic should settle on shrimp. All else equal, once you shift to a target that is thousands of times larger, an intervention could be far less effective [in terms of robustly increasing welfare in expectation] and still compete on expected impact.

I very much agree. Moreover, I do not even know whether electrically stunning farmed shrimps increases or decreases welfare due to effects on soil animals and microorganisms.

Are you thinking about humans as an aligned collective in the 1st paragraph of your comment? I agree all humans coordinating their actions together would have more power than other groups of organisms with their actual levels of coordination. However, such level of coordination among humans is not realistic. All 10^30 bacteria (see Table S1 of Bar-On et al. (2018)) coordinating their actions together would arguably also have more power than all humans with their actual level of coordination.

I agree it is good that no human has power over all humans. However, I still think one being dominating all others has a probability lower than 0.001 % over the next 10 years. I am open to bets against short AI timelines, or what they supposedly imply, up to 10 k$. Do you see any that we could make that is good for both of us under our own views?

Hi Guy. Elon Musk was not the only person responsible for the recent large cuts in foreign aid from the United States (US). In addition, I believe outcomes like human extinction are way less likely. I agree it makes sense to worry about concentration of power, but not about extreme outcomes like human extinction.

Thanks for sharing, James.

Thanks for the relevant post, Wladimir and Cynthia. I strongly upvoted it. Do you have any practical ideas about how to apply the Sentience Bargain framework to compare welfare across species? I would be curious to know your thoughts on Rethink Priorities' (RP's) research agenda on valuing impacts across species.

Thanks for the great post, Lukas. I strongly upvoted it. I also agree with your concluding thoughts and implications.