Iām earning to give as a Quant Researcher at the Quantic group at Walleye Capital, a hedge fund. In my free time, I enjoy reading, discussing moral philosophy, and dancing bachata and salsa.
Iām also on LessWrong and have a Substack blog.
Ariel Simnegar šø
Karma: 2,805
MISHNA: Rabbi GWWC said in the name of Rabbi Singer: It is a mitzvah (good deed) to pledge 10%, but one is not required to take upon himself the chumra (stringency) of the Further Pledge.
GEMARA: Rava asks: One who takes the Further Pledge can be compared to the Nazirite, who is called a sinner, for he is depriving himself of what the Holy One, Blessed be He, has provided him. So how can Rabbi GWWC say that one who takes Further Pledge is a righteous man?
Abaye says in the name of Rabbi Singer: The mashal (parable) of the drowning child brings down that one is obligated to give up all of oneās possessions to save anotherās life. For this reason Rabbi GWWC says one who takes the Further Pledge is a righteous man. As Scripture teaches us, āone who saves a life is as though he has saved the world entireā.
Rava asks: But why then is 10% sufficient, if it is brought down that one must give up all of oneās posessions to save a life?
Abaye says: In the matter of the city of Sodom, the Lord says that āfor the sake of 10 righteous men, I would not destroy itā. By homiletic interpretation, if one donates even 10%, for his sake the world will be spared.
I agree that clinicians should use lidocaine or digoxin over potassium chloride (KCL) for the reason you gave.
I wrote that the injection is āoften of potassium chlorideā, not always.
-
Given that the fetus is receiving a lethal dose of potassium chloride, I donāt think adults tolerating a much smaller medicinal dose should tell us much about how painful a lethal dose would be?
I agree that the fetus isnāt being given potassium chloride intravenously, although I didnāt know that when I wrote the post (another commenter pointed it out). Iāll add a line in the post disclaiming that comparison.
Happy to hear we agree on fetal anesthesia :)
I also very much agree that thereās no conflict between this and the pro-choice position, and that increased abortion access would reduce fetal suffering in late-term abortions. (Although increasing abortion access has other, larger ethical problemsāfrom a total utilitarian perspective, there doesnāt seem to be much difference between preventing a fetus from living a full life and doing the same for an infant or adult.)
On comparing individual fetuses to individual farm animals, itās worth noting that a 13-week fetus has about half as many neurons as an adult cow. (Cows have 3 billion neurons, while 13-week fetuses have 3 billion brain cells. Since humans have a near 1:1 neuron-glia ratio, a 13-week fetusās neuron count should be about half as many as a cowās.) So on at least one metric, theyād be pretty comparable. Of course, Iām pretty sure this fact is swamped by the other facts about factory farming you gave.
I agree that this probably wouldnāt be competitive with animal welfare. However, if weāre holding it to the standard for suffering-reducing interventions for humans, it could plausibly be more competitive.
This description of labor induction abortion says:
The skin on your abdomen is numbed with a painkiller, and then a needle is used to inject a medication (digoxin or potassium chloride) through your abdomen into the fluid around the fetus or the fetus to stop the heartbeat.
That sounds like local anesthesia for the mother, which from what I understand is achieved through an injection which numbs the tissue in a specific area rather than through an IV drip. So I donāt think this protocol would have any anesthetic effect on the fetus, though Iām not a medical expert and could be wrong.
Based on this, I think the sentence āThe fetus is administered a lethal injection with no anesthesiaā is accurate.
Thanks for that info! I didnāt know that.
Thanks for this! I agree that apart from speciesism, there isnāt a good reason to prioritize GHD over animal welfare if targeting suffering reduction (or just directly helping others).
Would you mind expanding further on the goals of the āreliable global capacity growthā cause bucket? It seems to me that several traditionally longtermist /ā uncategorized cause areas could fit into this bucket, such as:
Under your categorization, would these be included in GHD?
It also seems that some traditionally GHD charities would fall into the āsuffering reductionā bucket, since their impact is focused on directly helping others:
Fistula Foundation
StrongMinds
Under your categorization, would these be included in animal welfare?
Also, would you recommend that GHD charity evaluators more explicitly change their optimization target from metrics which measure directly helping others /ā suffering reduction (QALYs, WELLBYs) to āglobal capacity growthā metrics? What might these metrics look like?
It might be that the strongest reason to prioritize GHD is because of flow-through effects, as youāve suggested. But I donāt think that those who prioritize GHD generally actually do so for that reason. They care about saving and improving peopleās lives in the near term, and the units they use (QALYs, income doublings, WELLBYs) and stories they tell (the drowning child) reflect that.
If GHD was trying to optimize for robustly increasing long-term human capacity, I think the GHD portfolio of interventions would look very different. It might include certain longtermist cause areas such as improving institutional decisionmaking. It would be surprising if the best interventions when optimizing for longterm flow-through effects were also the best when optimizing for immediate effects on individuals. If youāre optimizing for flow-through effects, I agree that itās non-obvious whether GHD or AW is better, but I think you probably shouldnāt be donating to either of those!
I think GHD donors choose GHD over AW simply because they care overwhelmingly more about humans than nonhuman animals. Thatās also why they usually ignore animal effects in their cost-effectiveness analyses, even though those effects would swamp the effects on humans for many GHD interventions. If they were trying to impartially help others in the near term, they would choose AW.
Hereās a classification of GHD/āAW which I think is more relevant to neartermistsā revealed preferences: The best impartial neartermist interventions are AW. The best neartermist interventions ignoring nonhuman animals are GHD. Under that classification, fetal welfare would be GHD.
I very much agree that itās a clear moral improvement unless thereās some strong countervailing consideration. I would guess the greatest practical difficulty would be the interventionās adjacency to politically contentious issues, which might make it intractable.
fwiw, I think a better comparison would be leading animal welfare interventions
I agree that there are many similarities between this proposal and animal welfare interventions. However, since I think the best animal welfare interventions are orders of magnitude more effective than GHD, Iād far rather GHD funding be diverted to this than animal welfare funding. I also just donāt think this intervention would be anywhere near the animal welfare cost-effectiveness bar, though it could conceivably pass the global health bar.
Thanks for putting this post together. It takes fortitude to commit so much to an altruistic project, and it takes integrity to make this decision and write up this explanation.
Insightful and well-argued post!
I found the hypothetical about NYT and CEA helpful for reasoning from first principles about acceptable journalistic practice. I came out of it empathizing more with Nonlinearās feelings before and during the publication of Ben Paceās article than I previously had.
Regarding Ben Paceās explicit seeking of negative information and unwillingness to delay posting, you updated me from thinking of these as simple mistakes to now considering them egregiously bad.
Great point that an article author canāt just state their disclaimers at the top and expect readers to rationally recalibrate themselves and ignore the vibes of the evidenceās presentation.
I found it hard to update throughout this story because the presentation of evidence from both parties was (understandably) biased. As you pointed out, āSharing Information About Nonlinearā presented sometimes true claims in a way which makes the reader unsympathetic to Nonlinear. Nonlinearās response presented compelling rebuttals in a way which was calculated to increase the readerās sympathy for Nonlinear. Both articles intentionally mix the evidence and the vibes in a way which makes it difficult to readers to separate the two. (I donāt blame Nonlinearās response for this as much, since it was tit for tat.)
Thanks again for putting so much time and effort into this, and Iām excited to see what you write next.
Hi! Iām assuming that by āthisā you mean the postās argument, āwild animalsā you mean wild animal welfare research, and āstray domestic animalsā you mean pet shelters. In that case, I think the postās argument might apply to wild animal welfare research, depending upon oneās model of the effects of that research. However, I think this postās argument is unlikely to apply to pet shelters.
Comparing area was intended :)
If itās unclear, I can add a note which says the circles should be compared by area.
Thanks so much David! :)
Agreed on avoiding harming insects!
Though itās commendable to try to help insects, putting a bug in the trash might be negative, because that increases insect populations, and insects might lead negative lives: https://āāwww.simonknutsson.com/āāhow-good-or-bad-is-the-life-of-an-insect
Avoiding silk, shellac, and carmine also helps reduce suffering for many insects: https://āāwww.wikihow.fitness/āāAvoid-Hurting-Insects
Thanks for the compliment :)
When I write āskepticism of formal philosophyā, I more precisely mean āskepticism that philosophical principles can capture all of whatās intuitively importantā. Hereās an example of skepticism of formal philosophy from Scott Alexanderās review of What We Owe The Future:
Iām not sure I want to play the philosophy game. Maybe MacAskill can come up with some clever proof that the commitments I list above imply I have to have my eyes pecked out by angry seagulls or something. If thatās true, I will just not do that, and switch to some other set of axioms. If I canāt find any system of axioms that doesnāt do something terrible when extended to infinity, I will just refuse to extend things to infinity...I realize this is āanti-intellectualā and ādefeating the entire point of philosophyā.
You make a good point regarding the relative niche-ness of animal welfare and AI x-risk. I agree that my postās analogy is crude and there are many reasons why peopleās dispositions might favor AI x-risk reduction over animal welfare.
Thanks Gage!
Thatās a good point I hadnāt considered! I donāt think thatās OPās crux, but it is a coherent explanation of their neartermist cause prioritization.
Absolutely! Most of whatās important in this essay is just a restatement of your inspiring CEA from months ago :)
This extra context makes the case much stronger.
Thanks for being charitable :)
On the percentile of a product of normal distributions, I wrote this Python script which shows that the 5th percentile of a product of normally distributed random variables will in general be a product of much higher percentiles (in this case, the 16th percentile):
import random MU = 100 SIGMA = 10 N_SAMPLES = 10 ** 6 TARGET_QUANTILE = 0.05 INDIVIDUAL_QUANTILE = 83.55146375 # From Google Sheets NORMINV(0.05,100,10) samples = [] for _ in range(N_SAMPLES): r1 = random.gauss(MU, SIGMA) r2 = random.gauss(MU, SIGMA) r3 = random.gauss(MU, SIGMA) sample = r1 * r2 * r3 samples.append(sample) samples.sort() # The sampled 5th percentile product product_quantile = samples[int(N_SAMPLES * TARGET_QUANTILE)] implied_individual_quantile = product_quantile ** (1/3) implied_individual_quantile # ~90, which is the *16th* percentile by the empirical ruleI apologize for overstating the degree to which this reversion occurs in my original reply (which claimed an individual percentile of 20+ to get a product percentile of 5), but I hope this Python snippet shows that my point stands.
I did explicitly say that my calculation wasnāt correct. And with the information on hand I canāt see how I couldāve done better.
This is completely fair, and Iām sorry if my previous reply seemed accusatory or like it was piling on. If I were you, Iād probably caveat your analysisās conclusion to something more like āUnder RPās 5th percentile weights, the cost-effectiveness of cage-free campaigns would probably be lower than that of the best global health interventionsā.
Hi Hamish! I appreciate your critique.
Others have enumerated many reservations with this critique, which I agree with. Here Iāll give several more.
why isnāt the ā1000xā calculation actually spelled out?
As youāve seen, given Rethinkās moral weights, many plausible choices for the remaining āmade-upā numbers give a cost-effectiveness multiple on the order of 1000x. Vasco Grilo conducted a similar analysis which found a multiple of 1.71k. I didnāt commit to a specific analysis for a few reasons:
I agree with your point that uncertainty is really high, and I donāt want to give a precise multiple which may understate the uncertainty.
Reasonable critiques can be made of pretty much any assumptions made which imply a specific multiple. Though these critiques are important for robust methodology, I wanted the post to focus specifically upon how difficult it seems to avoid the conclusion of prioritizing animal welfare in neartermism. I believe that given Rethinkās moral weights, a cost-effectiveness multiple on the order of 1000x will be found by most plausible choices for the additional assumptions.
(Although I got the 5th and 95th percentiles of the output by simply multiplying the 5th and 95th percentiles of the inputs. This is not correct, but Iām not sure thereās a better approach without more information about the input distributions.)
Sadly, I donāt think that approach is correct. The 5th percentile of a product of random variables is not the product of the 5th percentilesāin fact, in general, itās going to be a product of much higher percentiles (20+).
To see this, imagine if a bridge is held up 3 spokes which are independently hammered in, and each spoke has a 5% chance of breaking each year. For the bridge to fall, all 3 spokes need to break. Thatās not the same as the bridge having a 5% chance of falling each yearāthe chance is actually far lower (0.01%). For the bridge to have a 5% chance of falling each year, each spoke would need to have a 37% chance of breaking each year.
As you stated, knowledge of distributions is required to rigorously compute percentiles of this product, but it seems likely that the 5th percentile case would still have the multiple several times that of GiveWell top charities.
letās not forget second order effects
This is a good point, but the second order effects of global health interventions on animals are likely much larger in magnitude. I think some second-order effects of many animal welfare interventions (moral circle expansion) are also positive, and I have no idea how it all shakes out.
Funnily enough, that verse is often referenced to me by religious Jews when I talk about how many EAs donate >>20%.