Effects on microorganisms are much larger than those on animals and plants under the assumptions of Rethink Priorities’ mainline welfare ranges?
By Vasco Grilo🔸 @ 2025-07-28T16:26 (+12)
Summary
- Under the assumptions of Rethink Priorities’ (RP’s) mainline welfare ranges, it seems clear that effects on microorganisms are much larger than those on nematodes, and that these are much larger than those on humans, farmed animals, and plants.
- My understanding is that Bob Fischer, who led RP’s moral weight project, thinks the methodology used to calculate RP’s mainline welfare ranges does not apply to silkworms, or organisms similarly or less complex than these, like microorganisms and nematodes. I wonder whether RP balked at estimates friendly to plants, nematodes, and microorganisms. They gathered lots of information about proxies for welfare for those organisms, but did not present it, estimate their welfare range, or arguably sufficiently explain how to interpret the mainline welfare ranges they presented for more complex organisms considering the estimation method is less applicable to less complex organisms.
- I think the welfare ranges of plants, nematodes, and microorganisms are much smaller than implied by the assumptions of RP’s mainline welfare ranges. However, I still expect effects on microorganisms to be much larger than those on nematodes, and these to be much larger than those on humans, farmed animals, and plants.
- I recommend being highly sceptical of RP’s mainline welfare ranges, which I believe are bound to be relatively close to 1. At the same time, I recommend being even more sceptical of non-systematic approaches to estimate welfare ranges. I think these are at higher risk of being disconnected from empirical evidence, and lacking transparency, thus being harder to improve upon.
- For organisms with neurons, I suggest using welfare ranges equal to the geometric mean between the number of neurons as a fraction of that of humans, and RP’s mainline welfare ranges. This ensures the welfare range is not solely determined by structural properties, but that these moderate the effect of behavioural proxies on it. My preferred welfare ranges are in the last table below.
- I speculate soil microorganisms have a negative welfare per organism-year, as is my best guess for soil nematodes, and therefore recommend decreasing their population. I expect the density of microorganisms to correlate with net primary production, and this to be lower for agricultural land than the counterfactual land use. So I recommend increasing agricultural land to increase the welfare of soil microorganisms. Nonetheless, I guess this is only 50 % likely to be beneficial, as I guess the probability of microorganisms having a negative welfare per organism-year is roughly 50 %, although below this.
- I estimated the cost-effectiveness of funding the High Impact Philanthropy Fund (HIPF) from the Centre for Exploratory Altruism Research (CEARCH) accounting for humans, and soil nematodes, mites, and springtails to be 16.5 times the past cost-effectiveness of the Shrimp Welfare Project’s (SWP’s) Humane Slaughter Initiative (HSI). Nonetheless, I think funding HIPF is way more cost-effective than this due to decreasing the population of microorganisms via increasing agricultural land. As a result, I recommend funding HIPF instead of SWP, or other organisations I had recommended targeting invertebrates.
- My top recommendation is decreasing the uncertainty about the welfare in QALY/organism-year of soil microorganisms and nematodes. In particular, it is crucial to know whether they have positive or negative lives. Nevertheless, I am not aware of any opportunities one can support now to make progress on that in a sufficiently targeted way which I consider more cost-effective than funding HIPF.
Introduction
I estimated effects on soil nematodes, mites, and springtails are much larger than those on target beneficiaries. So I decided to investigate the effects on microorganisms too. Here are my calculations.
Population
Based on Table S1 of Bar-on et al. (2018), I assume 10^13 plants, 10^21 nematodes, 10^27 protists, 10^29 archaea, and 10^30 bacteria. Plants, nematodes, and protists are eukaryotes. Archaea and bacteria are prokaryotes. From Flemming and Wuertz (2019), 25 % (= 3*10^29/(1.2*10^30)) of bacteria live in the soil.
I suppose 230 billion farmed shrimp as computed by RP for 2020. I have not considered animals besides nematodes, and farmed shrimp. I think effects on nematodes account for the vast majority of the effects on animals, and shrimp are the farmed animals existing in greater numbers, thus providing good context about farmed animals.
Welfare range
RP estimated their mainline welfare ranges based on 9 models giving a weight of 1/9 to each. They also calculated the welfare range conditional on sentience of plants, nematodes, protists, and prokaryotes for 2 of those 9 models, the cubic, and pleasure-and-pain-centric models. The sheet I just linked is no longer public, but here is a copy from July 2. I assume the welfare range conditional on sentience is 0 for all the other 7 models to get lower bounds for RP’s mainline welfare ranges of plants, nematodes, protists, and prokaryotes. In addition, I speculate the probability of sentience of plants, protists, and prokaryotes is 6.8 % that of nematodes, which I set to RP’s estimate of 6.8 %. This implies the ratio between the probability of sentience of plants/protists/prokaryotes and nematodes equals the ratio between the probability of sentience of nematodes and humans, who I assume are certain to be sentient. RP does not present a probability of sentience for shrimp, as I commented on February 12, so I supposed it to be equal to that of crayfish of 45.3 %.
The results are below. “E” stands for “*10^”. As I expected, farmed shrimp and nematodes have the 1st and 2nd highest lower bound for RP’s mainline welfare, and that of protists is higher than that of prokaryotes. The lower bounds for RP’s mainline welfare ranges of plants, nematodes, and microorganisms are relatively close to 1. This is because RP’s welfare ranges conditional on sentience of plants, nematodes, and prokaryotes for the cubic and pleasure-and-pain-centric models are relatively close to 1, and each of these models has a significant weight of 1/9.
| Organisms | Probability of sentience | Welfare range conditional on sentience for the cubic model | Welfare range conditional on sentience for the pleasure-and-pain-centric model | Lower bound for RP's mainline welfare range conditional on sentience | Lower bound for RP's mainline welfare range |
| Plants | 0.462% | 1.33E-04 | 0.0556 | 0.00619 | 2.86E-05 |
| Farmed shrimp | 45.3% | 0.0147 | 0.296 | 0.0345 | 0.0156 |
| Nematodes | 6.80% | 0.0456 | 0.389 | 0.0483 | 0.00328 |
| Protists | 0.462% | 0.00984 | 0.130 | 0.0155 | 7.18E-05 |
| Archaea | 0.462% | 7.75E-04 | 0.0741 | 0.00832 | 3.85E-05 |
| Bacteria | 0.462% | 7.75E-04 | 0.0741 | 0.00832 | 3.85E-05 |
Maximum annual welfare gain
I calculate the lower bound for the maximum annual welfare gain multiplying the population by the lower bound for RP’s mainline welfare range. I estimate that lower bound as a fraction of the maximum annual welfare gain for humans assuming this equals the human population in 2024 of 8.16 billion, as the welfare range of humans is 1 by definition.
The results are below. Under the assumptions of RP’s mainline welfare ranges, it seems clear that effects on microorganisms are much larger than those on nematodes (as I had commented), and that these are much larger than those on humans, farmed animals, and plants.
| Organisms | Lower bound for the maximum annual welfare gain | Lower bound for the maximum annual welfare gain as a fraction of the maximum annual welfare gain for humans |
| Plants | 2.86E+08 | 3.51% |
| Farmed shrimp | 3.60E+09 | 44.1% |
| Nematodes | 3.28E+18 | 4.02E+08 |
| Protists | 1.44E+23 | 1.76E+13 |
| Archaea | 3.85E+24 | 4.71E+14 |
| Bacteria | 3.85E+25 | 4.71E+15 |
My understanding is that Bob Fischer, who led RP’s moral weight project, thinks the methodology used to calculate RP’s mainline welfare ranges does not apply to silkworms, or organisms similarly or less complex than these, like microorganisms and nematodes. Laura Duffy, who wrote the report describing the estimation of RP’s mainline welfare ranges, said that, “When Bob was selecting the species, he was thinking of adult insects as the edge cases for the model (bees, BSF [black soldier flies]). He included juveniles [silkworms] to see what the model implies, not because he really thought the model should be extended to them”. I wonder whether RP balked at estimates friendly to plants, nematodes, and microorganisms. They gathered lots of information about proxies for welfare for those organisms, but did not present it, estimate their welfare range, or arguably sufficiently explain how to interpret the mainline welfare ranges they presented for more complex organisms considering the estimation method is less applicable to less complex organisms. I encouraged Bob Fischer and Laura Duffy to present information about proxies for welfare for nematodes on June 12, and asked for clarifications about RP’s welfare ranges conditional on sentience for the cubic and pleasure-and-pain-centric models on July 2, as I noted the very high values for plants, nematodes, and microorganisms, but they have not replied.
I think the welfare ranges of plants, nematodes, and microorganisms are much smaller than implied by the assumptions of RP’s mainline welfare ranges. However, I still expect effects on:
- Microorganisms to be much larger than those on nematodes:
- I guess the welfare range per unit mass decreases with mass, and bacteria individually have a way smaller mass than nematodes, so I think bacteria have a way larger welfare range per unit mass.
- Bacteria collectively have 3.50 k (= 70/0.02) times as much carbon as nematodes.
- Nematodes to be much larger than those on humans, farmed animals, and plants:
- Nematodes individually have a way smaller mass than humans, farmed animals, and plants, so I think nematodes have a way larger welfare range per unit mass.
- Nematodes collectively have 33.3 % (= 0.02/0.06), 20.0 % (= 0.02/0.1), and 0.00444 % (= 0.02/450) times as much carbon as humans, farmed animals, and plants, which I guess is enough for the above effect to dominate.
My recommendations
I recommend being highly sceptical of RP’s mainline welfare ranges, which I believe are bound to be relatively close to 1. A single behavioural proxy likely to be absent, meaning 12.5 % (= (0 + 0.25)/2) likely to be present, implies a welfare range conditional on sentience, and the rate of subjective experience of humans under the pleasure-and-pain-centric model of at least 0.00339 (= 0.125/37) regardless of the simplicity of the organism. The sheet I just linked has meanwhile been made private, but the model considers 37 equally important hedonic proxies whose effect is not moderated by structural properties like the number of neurons.
At the same time, I recommend being even more sceptical of non-systematic approaches to estimate welfare ranges. I think these are at higher risk of being disconnected from empirical evidence, and lacking transparency, thus being harder to improve upon.
For organisms with neurons, I suggest using welfare ranges equal to the geometric mean between the number of neurons as a fraction of that of humans, and RP’s mainline welfare ranges. This ensures the welfare range is not solely determined by structural properties, but that these moderate the effect of behavioural proxies on it. My preferred welfare ranges are below. I used the number of neurons from RP’s neuron count model. This sheet has meanwhile been made private, but there are similar values here.
| Animal | Number of neurons | Number of neurons as a fraction of that of humans | RP's mainline welfare range | My preferred welfare range | My preferred welfare range as a fraction of RP's mainline welfare range |
| Human | 8.60E+10 | 100% | 1.00 | 1.00 | 100% |
| Pig | 4.30E+08 | 0.500% | 0.515 | 0.0507 | 9.85% |
| Chicken | 2.20E+08 | 0.256% | 0.332 | 0.0291 | 8.78% |
| Octopus | 5.00E+08 | 0.581% | 0.213 | 0.0352 | 16.5% |
| Carp | 1.30E+07 | 0.0151% | 0.0890 | 0.00367 | 4.12% |
| Bee | 9.00E+05 | 0.00105% | 0.0710 | 8.62E-04 | 1.21% |
| Salmon | 1.30E+07 | 0.0151% | 0.0560 | 0.00291 | 5.20% |
| Crayfish | 8.00E+04 | 9.30E-07 | 0.0380 | 1.88E-04 | 0.495% |
| Shrimp | 8.00E+04 | 9.30E-07 | 0.0310 | 1.70E-04 | 0.548% |
| Crab | 1.00E+05 | 1.16E-06 | 0.0230 | 1.64E-04 | 0.711% |
| Black soldier fly | 3.50E+05 | 4.07E-06 | 0.0130 | 2.30E-04 | 1.77% |
| Silkworm | 8.60E+05 | 0.00100% | 0.00200 | 1.41E-04 | 7.07% |
I speculate soil microorganisms have a negative welfare per organism-year, as is my best guess for soil nematodes, and therefore recommend decreasing their population. I expect the density of microorganisms to correlate with net primary production, and this to be lower for agricultural land than the counterfactual land use. So I recommend increasing agricultural land to increase the welfare of soil microorganisms. Nonetheless, I guess this is only 50 % likely to be beneficial, as I guess the probability of microorganisms having a negative welfare per organism-year is roughly 50 %, although below this. Relatedly, Brian Tomasik has a post about the welfare of bacteria and plants.
I estimated the cost-effectiveness of funding HIPF from CEARCH accounting for humans, and soil nematodes, mites, and springtails to be 16.5 times the past cost-effectiveness of SWP’s HSI. Nonetheless, I think funding HIPF is way more cost-effective than this due to decreasing the population of microorganisms via increasing agricultural land. As a result, I recommend funding HIPF instead of SWP, or other organisations I had recommended targeting invertebrates.
My top recommendation is decreasing the uncertainty about the welfare in QALY/organism-year of soil microorganisms and nematodes. In particular, it is crucial to know whether they have positive or negative lives. Nevertheless, I am not aware of any opportunities one can support now to make progress on that in a sufficiently targeted way which I consider more cost-effective than funding HIPF. I asked people from the Animal Welfare Fund (AWF), RP, Wild Animal Initiative (WAI), and Welfare Footprint Institute (WFI) about this 2 months ago in the context of soil nematodes, mites, and springtails. Only Cynthia Schuck‑Paim from WFI replied, saying Wladimir Alonso from WFI is working on a project related to assessing differences in hedonic capacity, which I guess relates to this post.
Bob Fischer @ 2025-07-28T18:27 (+78)
Hi Vasco,
Several quick clarifications.
1. It's true that we don't think you can take our methodology and extend it arbitrarily. We grant that it’s very difficult to draw a precise boundary. However, it's standard to develop a model for a purpose and be wary about its application in a novel context. Very roughly, we take those novel contexts to be ones where the probability of sentience is extremely low. We acknowledge that we don’t have a precise cutoff for “extremely low,” as establishing such a cutoff would be a difficult research project in its own right. There are unavoidable judgment calls in this work.
2. RP has done lots of work on animal sentience. It is not all the Moral Weight Project. It is not all connected and integrated. And some of our earliest MWP ideas are ones we later abandoned. What we stand behind now is really just what we published in the book. It is not fair to ask us to tell a coherent story about disconnected projects with different purposes, as well as all stages of the same project, given that different teams worked on these projects and the evolving understanding of people on a team for a given project.
3. We don't think that the assumptions of our "mainline welfare ranges" imply anything about the welfare ranges of plants, nematodes, and microorganisms, as the models simply aren't intended to be used the way you are using them. That's why we aren’t replying to you about the welfare ranges of plants, nematodes, and microorganisms. We would need to do an independent project to form opinions on your questions. Right now, we don’t have the funding for that project.
4. It's understandable that you're skeptical of our specific welfare range estimates. We, of course, are also skeptical of those precise numbers. That's why we've long encouraged people to focus on the order of magnitude estimates. We also disagree that they "are bound to be relatively close to 1." A few orders of magnitude lower than 1 is not close to 1, at least by one reasonable interpretation of "relatively close." Laura has already discussed this elsewhere.
For what it's worth, I think you're approaching the Moral Weight Project as something it is not. You are treating it as a general methodology where we can enter some information about the abilities of a system—whatever that system happens to be—and get out moral weights that we can use in expected value calculations for cause prioritization. But we did not try to produce a maximally general methodology. We tried to produce something useful for updating on the kinds of questions that decision-makers actually face: "Do layer hens matter so much more than carp that, despite the differences in population sizes, you should prioritize layers?" "Can we rule out insects entirely?" “If your job requires you to apply a discount rate to the welfare of some animals relative to others, what kinds of numbers should you consider?” "Is there a good reason for thinking that, even if humans don't literally have lexical priority over animals, they have that kind of priority for practical purposes?" And so on. I do think that the MWP is useful for shedding some light on these questions for some actors. Beyond that, we should be cautious with the outputs. And mostly, we should try to do better, as we only meant to issue the first word.
Vasco Grilo🔸 @ 2025-08-27T16:17 (+2)
1. It's true that we don't think you can take our methodology and extend it arbitrarily. We grant that it’s very difficult to draw a precise boundary. However, it's standard to develop a model for a purpose and be wary about its application in a novel context. Very roughly, we take those novel contexts to be ones where the probability of sentience is extremely low. We acknowledge that we don’t have a precise cutoff for “extremely low,” as establishing such a cutoff would be a difficult research project in its own right. There are unavoidable judgment calls in this work.
I estimate effects on soil animals would still be much larger than those on the target beneficiaries for a welfare per animal-year of exactly 0 for animals with fewer neurons than those considered in your book, and "welfare range as a fraction of that of humans" = "number of neurons as a fraction of that of humans"^0.19, which explains very well the welfare ranges in Table 8.6 of your book (78.6 % of their variance for an exponent of 0.188). I calculate soil ants and termites have 2.91 and 1.16 times as many neurons as shrimp, so effects on them would still be relevant. I get the following increase in the welfare of soil ants and termites as a fraction of the increase in the welfare of the target beneficiaries for an exponent of 0.19 (the chicken welfare corporate campaigns would decrease animal welfare):
- For cage-free corporate campaigns, -20.4.
- For buying beef, 3.31 M.
- For broiler welfare corporate campaigns, -321.
- For GiveWell’s top charities, 83.6 k.
- For HIPF, 65.5 k.
justsaying @ 2025-08-13T18:48 (+3)
Hi Vasco, I am curious why you stop at nemotodes and don't extend your conclusions to single-called organisms such as bacteria.
Vasco Grilo🔸 @ 2025-08-13T19:50 (+2)
Hello. Could you clarify what would be extending my conclusions? I have analysed microorganisms, including bacteria, in the post, and took the same conclusion.
I speculate soil microorganisms have a negative welfare per organism-year, as is my best guess for soil nematodes, and therefore recommend decreasing their population. I expect the density of microorganisms to correlate with net primary production, and this to be lower for agricultural land than the counterfactual land use. So I recommend increasing agricultural land to increase the welfare of soil microorganisms.