Currently occupying the lowest-status position in effective altruism: animal welfare desk researcher. Iāve done other stuff, like going through a charity incubator and failing to launch, operations for a funder, coaching for an EA residency, and EA grantmaking.
āWe have enormous opportunity to reduce suffering on behalf of sentient creatures [...], but even if we try our hardest, the future will still look very bleak.āāBrian Tomasik
Suffering-Focused Ethics: Defense and Implications
Good framing!
A framing I had in a draft post was whether, in the ITN framework, to focus mostly on (disclaimer: very rough):
Scale: this is the AI safety approach. āThis could be really big (and is also neglected), and therefore we should see whether itās tractableāeven if it has a minuscule chance of being tractable, itās worth pursuingā.
Tractability: this is more the OG-global-health approach: āWe can demonstrate that our intervention makes measurable progress on the issue. Other more neglected or larger-scale problems cannot do this.ā
Neglectedness: this is probably the least common proxy, but you see it in animal welfare sometimes: āNo one is working on this species, therefore someone should start a projectā.
I do think rigor-reach /ā tractability-scale is the main axis, and that neglectedness is more minor. I think this can lead to unresolvable disagreements on what interventions should be pursued.
On AI safety harms beyond historical induction, there was a recent short post on ways current AI safety efforts could be net-negative.
Under your own views, any portfolio of interventions can easily increase or decrease animal suffering?
I agree. This is why I find myself having to ājustifyā a decision procedure that just considers a few select (albeit uncertain) effects rather than say āIām maximizing expected valueā.
I think my decision procedure can make interventions that decrease our uncertainty very appealing. I currently work on animal welfare desk research, and am strongly considering getting into some form of primary research in the future. Caveats:
Sometimes, this is not an option when making a decision. E.g., when advising a friend who received two job offers in animal welfare, neither of which decrease uncertainty; or when recommending grants for a fund (where none of the grants decrease uncertainty).
I think moral weights and pain trade-offs might be so seriously underdetermined that I expect that most research I could do or support right now would not reduce uncertainty about them. I expect youāll have a counter in that, though, and I expect my views to evolve.
In comparison, assessing 2nd-order effects of certain interventions (such as economic effects, and land use effects) is tractable, although difficult.
I kind of rule out ālong-horizon thinkingā in my decision procedure on principle. Hence, if I expected pain tradeoff ratios to become a resolvable issue after 30 years of empirical research necessiating $50M of investment per year, I donāt think Iād want to seed it, because I think that the chance that this works is negligible (youāll likely object to this on decision-theoretic grounds, I assume).
This is more uncomfortable to admit (and Iām particularly uncertain on that point), but if I were to focus almost entirely on research that decreases uncertainty, appearing as someone who recommends no object-level actions may be costly to me in a way that could reduce the change that my research is successful in decreasing uncertainty, because it could receive less support.
We must settle for merely establishing to our own satisfaction that we can reasonably maintain our beliefs in the face of the skepticās contrary arguments.
Itās nice to have access to your cluster of reasons for disagreeing with cluelessness. However, Iād have been most interested in seeing a post where you make a compelling argument for how the DiGiovanni cluelessness argument pattern-matches radical skepticism, or other arguments that philosophers find intriguing on the surface but generally flawed.
I do think that a crux for me when reading your post here, and other writings of yours on cluelessness, is whether ācommon senseā and āimpartial welfareā can ever go together. Iām not skilled in philosophical debate, so Iāll summarize my view with one of my favorite EA Forum quotes, from @Max_Daniel:
I resonate with this but would lightly push back. As someone who oscillates between being hosted by relatives in a city with 2 other EAs in total, and big, expensive hubs, I feel like the COL difference isnāt that drastic if you are enthusiastic on frugal lifestyles (not saying people who feel differently are not frugalāthere may be other factors, like getting lucky in finding accommodation).
Iām currently in San Francisco for a few months in a coworking office, and my COL difference from a non-hub place in Europe is around $1.1K per month, more than half of which comes from the office membership. I sleep in a very cheap shared flat, and live on a mix of beans, protein powder, and leftovers from the office. And the productivity benefits from the office space are quite strong (beyond any sort of networking value), though I canāt confidently claim theyāre worth $1.1K a month.
I agree that the benefits of living in a hub can be overstated, having oscillated between both. If you have enough savings to attend a few conferences a year, and are proactive virtually, you can absolutely achieve 80ā20 networking outcomes (in most cases, it might be harder when your job is very network-based).
I think ādonāt live in a hubā is more important advice if you want to start a family or find a stable place to live, as in those cases, it can definitely lock in cost increases that are much larger.
(On the other hand, if you could actually do your job from anywhere and donāt benefit from networks much, I think people should maximize. European founder should envision the cheapest EU countries, while software engineers doing earn to give should try their best to work in Switzerland or California)
Assuming we end factory farming before 2100, which factor will have contributed the most?
Broader changes (ending agriculture altogether? finding extremely efficient ways to get nutrients that donāt involve traditional āfoodā? transition to a non-biological humanity?), but some of these involve some form of alternative proteins (in a broad sense). I assume <20% of these worlds involve āwe find a way to perfectly imitate meatā, though.
I had a fairly similar experience to yours (got into Charity Entrepreneurship right at the end of college), and can confirm the accuracy for curious commentersāand that I also wasnāt thinking sufficiently deeply about career fit at the time (I estimated a sub-2% chance of being accepted and wasnāt taking it very seriously, though I think on reflection, I am a decent fit for that kind of path).
Going in my pantheon of favorite posts of 2026. Detailed, transparent, action-guiding, and engaginly written.
A few years ago, spending on animal welfare reforms in EA was so limited that running rigorous, informative studies to reduce the movementās uncertainty might have cost more than the movementās spending on the campaigns themselves. But as you mention, now that the funding bottleneck might be easing, such studies would inform much larger amounts of spending, and the expected value of information looks much higher. It thus seems like a very good time to raise an issue that would have been much harder to make progress on in the past (though as Lewis Bollardās comment notes, the challenge is still daunting).
I found the post well-written and balanced, and the discussion in the comments mostly helpful. Strongly upvoted.
Thanks for engaging!
I would say I feel good when I hear that a given place has banned bullfighting, and find the practice horrifying. My claim is in the context of the animal welfare movement and how it should spend its resources.
Bullfighting causes intense suffering to a few thousand bulls, for a few hours each. This makes it a very small issue in a consequentialist, most likely less than 0.000001% of the suffering caused by the egg industry on a yearly basis. Then, the counterfactual impact of a single spectator at a bullfight on the outcome is probably low. It doesnāt seem like something altruistic resources are well-spent on.
Meanwhile, I donāt see attitude change as a clear a step toward a better world for animals. āA better world for animalsā is a hideously complicated thing, and currently has little to no correlation with peopleās attitudes. French people burnt cats alive for fun in the 16th century, but factory farming didnāt exist and human-caused suffering to animals was surely <1% of what it was today. A more pro-animal nation may contain more animal suffering (more pesticides, eating smaller and more numerous animals...). And I currently have no reason to believe that there is less total suffering in countries that have banned bullfighting entirely.
I keep admiring AIMās dedication to continuous iteration in the service of impact!
āEAG London this Saturday, 12ā1pm: Office hours with Aaron and the AIM team. Drop in to ask questions, kick ideas around, or just say hi. Find it on Swapcard.ā
Still not seeing this on Swapcard. Am I alone in this?
Low confidence flag, but from what Iāve heard, even GiveDirectly would be saturated quite early by that amount of money, and not be able to distribute it. If anyone has a figure on where that threshold is (or thinks Iām wrong), Iād be interested!
Nothing super constructive to say, but two years later, I still think about this post a lot and keep coming back to it to read some specific bullets. Itās somewhat moving, but also very content-dense, one of my favorite pieces of writing/āinformation on here.
I generally find pieces about how peopleās views and drives evolve helpful on many levels, and Iād welcome more of them, even from much less senior people. I try to keep track of this for myself: itās useful for gaining insights into, for example, what actually influences my priorities over time and how contingent those drives are.
I appreciate you raising this and Iām interested to see how people will answer. However, some weak counter-considerations come to mind:
Many grants go to individual independent researchers for relatively short periods of time, requiring a much higher amount of grantmaking time per FTE than other areas (say, global health) - I realize that this is perhaps not that strong of a consideration, given that many grantmakers mainly fund organizations
Grantmakers want the field to keep growing quickly, and I assume ācreating new granteesā is time-consuming. That means that the current ratio of grantmakers to AI safety FTEs is maybe not the best metric
Abraham has mentioned the new charity briefly on the Hive Slack.
āWhat exactly do you mean by moving goalposts?ā
Quick answer: itās a reference to this paragraph.
Reading a (free) book by Magnus Vinding two years ago is what got me into effective altruism, and I now work full-time on (hopefully) improving the lives of others. Iām very excited by this book, and have loved the chapter drafts that have been released early.
Fun fact: the draft chapter āhealthy habitsā from this new book convinced me to go from a night owl to a regular bedtime person, which is the single most transformative lifestyle change Iāve ever made. Iām awake for 2 hours more per day on average, and am less sleepy.
Thanks, Tom and Henning! Strongly upvoted.
I particularly appreciated the āSome Ramificationsā section. Their implications would be pretty demanding for EA Animal Welfare. Large infrastructure investments, probably tens of millions in research investment, to assess whether an intervention can get to initial scaling. All that, on timelines with which funders are usually uncomfortable.
An already ambitious ecosystem-level next step would be talent programs for specialist researchers to contribute to EA animal welfare research (this might not be the best way to fill the research gaps, though, and maybe more straightforward infrastructure for commissioning specialists would be more useful). Itās been quite neglected historically, and Iāve only heard of one or two organizations recently starting to work on this.