I am looking for work, and welcome suggestions for posts.
Vasco Grilošø
Thanks, Joel. That makes sense.
Thanks for the post, Matthew!
Thereās a lot about ethics that isnāt obvious. Itās hard to know whether utilitarianism is the right ethical theory, whether desert is real, and which theory of well-being is right. But other ethical issues arenāt difficult.
In case anyone is wondering (like I was), desert in philosophy is āthe condition of being deserving of somethingā.
Thanks for your efforts, and sharing your struggles, Brad!
Thanks for doing this! I have only read this post, but it seems like the series could be a valuable reference.
Hi Joel,
Have you considered the impact of advanced AI on CEARCHās cost-effectiveness analyses? Do you think the discount rates, and projections of future burden in CEARCHās cost-effectiveness analyses are appropriate under Ege Erdilās median time of 20 years until full automation of remote work? I estimate advocacy for taxing SSBs would become 0.637 % and 3.88 % as cost-effective neglecting impact after 5 and 10 years, although I would not neglect impact even after the full automation of remote work, which I expect to take decades.
They have also done some internal CEAs of SSB taxation projects; I think their major concern (a frustration shared by us) relates to high uncertainty over the existing GBD estimates of the SSB burden (n.b. the estimates changed wildly from one iteration of the GBD to the next, and itās not clear to us how or why the methodology changed).
I wonder whether GiveWell has considered making a grant to decrease the uncertainty of the burden of SSBs.
Both. I do not have reasons to believe organisations are under or overspending on fundraising. Some organisations say they have a hard time finding people who are a good fit for fundraising (being ātalent-constrainedā), but I think this only means there are steep diminishing returns on spending more on fundraising by increasing the earnings of possible fundraising roles. It does not mean they are underspending on fundraising. In general, I think it is sensible to at least have a prior expectation that the various activities on which an impact-focussed organisation can spend more money on have similar marginal cost-effectiveness. Otherwise, they would be leaving impact on the table by not moving money from the least to the most cost-effective activities at the margin. At the same time, I expect to find inefficiencies after learning more.
Thanks for the good points, Caleb.
* not sure how predictive choice of org to work at is of choice of org to donate to, lots of people I know donate to the org they work at because they think itās the best, some donate to think they think are less impactful (at least on utilitarian grounds) than the place they work (e.g. see CEA giving season charity recs) - you seem to think that orgs people donate to are better than orgs they work at but Idk if thatās true
I am assuming people would donate to organisations which are more cost-effective than their own in expectation because donating to ones which are less cost-effective would decrease their impact. This still leaves open the possibility of people donating to their own organisation (or asking to earn less), but they selected this partly for personal fit reasons which do not apply to donations, so I would expect most unbiased people to think there are other organisations which are more cost-effective than their own.
* a bit confused about the net effects of joining an org on its capital, e.g. lots of hires unlock more funding via fundraising capacity, credibility, etc.
Roles unlocking funds should ideally be paid more until the point where increasing earnings by 1 $ only increases funds by 1 $.
Thanks for the post, Alfredo.
Using this framework, the suffering captured by a DLES would tentatively correspond to the Excruciating category,[6] described as follows
Cynthia Schuck-Paim from WFI said āExamples [of excruciating pain] would include severe burning in large areas of the body, dismemberment, or extreme tortureā.
Taking triage seriously (and assuming widespread sympathy for intense suffering) means, in my opinion, that any governmentās top health priority should be to bring the DLES burden down as much as possible.
Would you still believe this under expectational total hedonistic utilitarianism?
Hi,
What is LASU?
Hi Yonatan,
Would it be better to assume organisations are indifferent between having a person work for them, and receiving what they would pay the person? I think so. It corresponds to the organisationsā revealed preferences, and I believe these are more reliable than their stated preferences. Organisations wanting to maximise their own impact (at the expense of global impact) have an incentive to overestimate the money they would have to receive to be happy to let the person go because they know the person could then donate to many other organisations.
Hi Caleb,
Donating 10 % more of oneās gross earnings to an organisation 10 times as cost-effective as one one could join is 10 (= 0.1*10/ā0.1) times as impactful as working there if the alternative hire would be 10 % less impactful? If you agree, do you have any thoughts on what is implied by it, and the distribution of cost-effectiveness across the jobs of people replying to the EA Survey?
Thanks to @Michael St Jules šø for noting I may be significantly underestimating the effects on soil arthropods due to only considering mites and springtails. I agree. I estimate below the annual welfare of nonarboreal ants is similar to that of soil mites and springtails. However, I estimate the annual welfare of soil nematodes is 12.8 times that of soil mites and springtails, and therefore accounting for ants would not significantly change the effects on soil nematodes, mites, and springtails. In any case, accounting for more wild animals with negative lives would reinforce my conclusion that effects on them are much larger than effects on target beneficiaries.
From Rosenberg et al. (2023), āOur estimated global population of nonarboreal ants, considered as part of the soil and plant litter habitat, is 5 Ć 10^16 (uncertainty range, 1 Ć 10^16 to 9 Ć 10^16) individualsā. I assume the welfare per animal-year of a fully happy nonarboreal ant as a fraction of their welfare range is ā38 %, as Gemini guessed for soil springtails. In addition, I suppose the welfare range of a nonarboreal ant is 0.0133 (= 1.02*0.013), 1.02 times the value in RPās post for black soldier flies (BSF) of 0.013, considering RPās estimate for the probability of sentience of ants is 1.02 (= 0.3/ā0.294) times their estimate for BSF, and my guess that the welfare range conditional on sentience of nonarboreal ants is similar to that of BSF. So I infer nonarboreal ants have a welfare of ā0.00505 QALY/āanimal-year (= ā0.38*0.0133), and ā2.52*10^14 QALY/āyear (= ā0.00505*5*10^16), which is ā35.2 k (= ā2.52*10^14/ā(7.16*10^9)) times my estimate for humans, 1.47 (= ā2.52*10^14/ā(-1.72*10^14)) times my estimate for soil mites and springtails, but only 10.2 % (= ā2.52*10^14/ā(-2.46*10^15)) of my estimate for soil nematodes, mites, and springtails.
Thanks for the great post, Sofia! I think it is a good example of unpacking a job.
Thanks, Nick. Here are the estimates from the book.
The mean for:
Pigs is 85.4 % (= 0.44/ā0.515) of the estimate in RPās post.
Chickens is 1.20 (= 0.40/ā0.332) times the estimate in RPās post.
Carp is 2.70 (= 0.24/ā0.089) times the estimate in RPās post.
Octopuses is 1.27 (= 0.27/ā0.213) times the estimate in RPās post.
Shrimp is 8.71 (= 0.27/ā0.031) times the estimate in RPās post.
Black soldier flies (BSF) is 5.62 (= 0.073/ā0.013) times the estimate in RPās post.
There were significant updates for carp, shrimp, and BSF. @vicky_cox, @Neil_Dullaghanš¹ , and @Vince Mak šø, you may want to consider the welfare ranges from Bobās book instead of those from RPās post, and stop using the value from RPās post respecting silkworms to model the benefits of interventions targeting farmed arthropods. Bob said āWhat we [RP] stand behind now is really just what we published in the bookā, and this does not have estimates for the welfare range of larvae.
The higher values for shrimp and BSF also make me think one would obtain larger welfare ranges applying the bookās methodology to less complex organisms (although I know it is not supposed to be extended).
I am not sure whether RP stands by the bookās means or medians. Could you clarify this, Bob? In any case, they do not differ much.
Thanks, Bob.
1. Itās true that we donāt think you can take our methodology and extend it arbitrarily. We grant that itās very difficult to draw a precise boundary. However, itās 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.
Would you be able to provide a range for what is āextremely lowā, and briefly explain why the probability of sentience of black soldier flies (which are covered in the book) is not extremely low, but that of nematodes is extremely low? I do not know whether your āextremely lowā means lower than 10^-10 or 1 %, but I guess the probability of sentience of nematodes may well be over 10 %. From Andrews (2024):
Given the determinate development of their nervous systems, 30-some years ago it was taken as given that C. elegans are too simple to learn. However, once researchers turned to examine learning and memory in these tiny animals, they found an incredible amount of flexible behavior and sensitivity to experience. C. elegans have short-term and long-term memory, they can learn through habituation (Rankin et al., 1990), association (Wen et al., 1997), and imprinting (Remy & Hobert, 2005). They pass associative learning tasks using a variety of sensory modalities, including taste, smell, sensitivity to temperature, and sensitivity to oxygen (Ardiel & Rankin, 2010). They also integrate information from different sensory modalities, and respond differently to different levels of intoxicating substances, āsupport[ing] the view that worms can associate a physiological state with a specific experienceā (Rankin, 2004, p. R618). There is also behavioral evidence that C. elegans engage in motivational trade-offs. These worms will flexibly choose to head through a noxious environment to gain access to a nutritious substance when hungry enough (Ghosh et al., 2016)āthough Birch and colleagues are not convinced this behavior satisfies the marker of motivational trade-offs because it appears that one reflex is merely inhibiting another (Birch et al., 2021, p. 31).
C. elegans are a model organism for the study of nociceptors, and much of what we now know about the mechanisms of nociception comes from studies on this species (Smith & Lewin, 2009). Behavioral responses to noxious stimuli are modulated by opiates, as demonstrated by a study finding that administration of morphine has a dose-dependent effect on the latency of response to heat (Pryor et al., 2007). And, perhaps surprisingly, when the nerve ring that comprises the C. elegans brain was recently mapped, researchers found that different regions of the brain support different circuits that route sensory information to another location where they are integrated, leading to action (Brittin et al., 2021).
Even if we grant the authorās low confidence in nematodesā having marker five (motivational trade-offs), current science provides ample confidence that nematodes have markers one (nociceptors), two (integrated brain regions), four (responsiveness to analgesics), and seven (sophisticated associative learning). Given high confidence that nematodes have even three of these markers, the reportās methodology [Birch et al. (2021)] would have us conclude that there is āsubstantial evidenceā of sentience in nematodes.
What we stand behind now is really just what we published in the book
Thanks for clarifying you only stand by what is in the book! I did not know this. Have you communicated this to organisations which have been using the estimates you had shared on EA Forum? I believe the community would find helpful a post where you share what you stand by, and which (expected) welfare ranges you recommend using for comparing welfare across species.
4. Itās understandable that youāre skeptical of our specific welfare range estimates. We, of course, are also skeptical of them. Thatās why we have long encouraged people to focus on the order of magnitude estimates.
For 10 of 11 species, my preferred welfare range is less than 10 % of the estimate RP shared on the EA Forum (see last table in my post), so I think the vast majority of these is more than 1 order of magnitude too high.
For what itās worth, I think you are 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 systemāwhatever that system happens to beāand get out moral weights that we can use in expected value calculations for cause prioritization.
I had understood from your past comment that the methodology is not supposed to apply to all biological systems (and I had already understood it was not supposed to apply to non-biological systems). However, I think it makes sense to trust the methodology less if you believe applying it to silkworms or nematodes produces results which are way less trustworthy than those for black soldier flies (which are covered in the book), as these do not seem that different from the former.
If the values below refer to welfare ranges accounting for the probability of sentience, I am confident that you would get mainline welfare ranges of microorganisms sufficiently high for effects on them to dominate all else.
Here is the confirmation.
If you read the welfare range reportās footnotes, youāll find that the 6.8% probability of sentience estimate for c elegans is driven substantially by my interpretation there that āprobably not sentientā meant 10-35%, which was really just an off-the-cuff judgment. The other people whose views were included in that assessment gave under 1% or under 2% probabilities of sentience, and updating based on the proxies didnāt budge the priors much. On reflection, I think lower numbers are more appropriate than 6.8%, and I really would not anchor on that as āRPās own lightsā.
I would say the probability of sentience of nematodes is higher than 6.8 %. From Andrews (2024):
Given the determinate development of their nervous systems, 30-some years ago it was taken as given that C. elegans are too simple to learn. However, once researchers turned to examine learning and memory in these tiny animals, they found an incredible amount of flexible behavior and sensitivity to experience. C. elegans have short-term and long-term memory, they can learn through habituation (Rankin et al., 1990), association (Wen et al., 1997), and imprinting (Remy & Hobert, 2005). They pass associative learning tasks using a variety of sensory modalities, including taste, smell, sensitivity to temperature, and sensitivity to oxygen (Ardiel & Rankin, 2010). They also integrate information from different sensory modalities, and respond differently to different levels of intoxicating substances, āsupport[ing] the view that worms can associate a physiological state with a specific experienceā (Rankin, 2004, p. R618). There is also behavioral evidence that C. elegans engage in motivational trade-offs. These worms will flexibly choose to head through a noxious environment to gain access to a nutritious substance when hungry enough (Ghosh et al., 2016)āthough Birch and colleagues are not convinced this behavior satisfies the marker of motivational trade-offs because it appears that one reflex is merely inhibiting another (Birch et al., 2021, p. 31).
C. elegans are a model organism for the study of nociceptors, and much of what we now know about the mechanisms of nociception comes from studies on this species (Smith & Lewin, 2009). Behavioral responses to noxious stimuli are modulated by opiates, as demonstrated by a study finding that administration of morphine has a dose-dependent effect on the latency of response to heat (Pryor et al., 2007). And, perhaps surprisingly, when the nerve ring that comprises the C. elegans brain was recently mapped, researchers found that different regions of the brain support different circuits that route sensory information to another location where they are integrated, leading to action (Brittin et al., 2021).
Even if we grant the authorās low confidence in nematodesā having marker five (motivational trade-offs), current science provides ample confidence that nematodes have markers one (nociceptors), two (integrated brain regions), four (responsiveness to analgesics), and seven (sophisticated associative learning). Given high confidence that nematodes have even three of these markers, the reportās methodology [Birch et al. (2021)] would have us conclude that there is āsubstantial evidenceā of sentience in nematodes.
From your evaluation of ACEās Movement Grants:
Among the 24 Movement Grants ACE announced on 30 July 2025, the amount granted of 1ā3 (= 8ā24) fell between 49.7 k and 51 k$. Is there an informal cap, @eleanor mcaree?
From the same evaluation:
I wish there was a greater focus on cost-effectiveness analyses.