Does P1 rely in any way on the idealized agent actually using EV specifically, as opposed to other theoretically possible aggregations of all or most of the possible consequences of A and B? It seems like no to me but I was curious since it is explicitly mentioned.
mal_grahamđ¸
Karma: 956
CountÂing anÂiÂmals: Stable popÂuÂlaÂtion size is not equivÂaÂlent to priÂorÂity level
Can you explain a little more what you mean by âcoarse-grainedâ in P3? Does that just mean âvery unlikely to include all the possible outcomes, so thereâs a ton of unknown unknowns we donât know how to assign probability toâ or something else?
I am not at all an expert in mosquitoes, but I think it would be hard to figure out more than the direct effects on mosquitoes, which are sex specific. A few assorted thoughts:
Scale: the Google Debug program website says when applied, they need to releases 100s of thousands to millions of mosquitoes each week, so the scale is reasonably high.
They release only male mosquitoes, but itâs not possible to rear only male mosquitoes (AFAIK), so they are killing all the females. The Debug projectâs Nature Biotechnology article says that they sort males from females as pupae (which may involve low welfare cost, I donât know when sentience emerges in insects and especially whether itâs possible to feel pain as a pupa), and again as adults (which, if insects are sentient, probably is painful for those females). They report that 95% of females are removed as pupae, so a much smaller amount are being killed as adults. About 2.5% of what make it through the first sort are females, so we can estimate that to achieve daily releases of ~ 75k males, ~2000 females are killed as adults daily. The paper does not seem to report how the females are killed for either pupae or adults, which is a big gap from a welfare perspective.
According to this, Wolbachia often shorten lifespans. But several strains also seem to enhance immune function and have other mutualisms. So in some cases it could have positive welfare effects. I asked Claude about the specific strain used in the Debug program, and it reported that most of the welfare effects land on females, who are already just being killed. Of course, we donât know that much about measuring mosquito welfare, but at least in terms of survival, ability to find a mate (which correlates to locomotor and sensory performance), etc., this strain isnât bad. And if they use something that really harms the released males they wonât outcompete the local males, which would drive up costs (because you have to release more mosquitoes to compensate).
The technique will result in a reduction in number of mosquitoes. Itâs not clear to me how other populations would respond to thatâIâm not sure what competitors there are for mosquito resources. But if you think mosquitoes have bad lives this could be a good thing (or neutral if you have person-affecting leanings). The FAQ on the Debug page claims that in urban areas, because the mosquitoes are invasive, it will just be a âreturn to normalââbut that ignores the fact that many other things have changed (reductions in native insect populations, for example). If itâs true that this specific species doesnât make up a meaningful part of any other animalâs diet the effect is likely to be small, but that sounds like something we donât really know (well-resolved urban food webs are rare). They might mean âno animal anybody in ecology especially cares about.â
For many bugs, high mortality during captive rearing is common. That doesnât seem to be the case for mosquitoesâsurvival rates in the 90s are reported in many captive colonies and mass die offs are unusual (according to this, thanks Claude). But the success rates for male mosquitoes in the Debug project is around 70%. The paper is fairly detailed on the rearing approach so someone with more mosquito expertise than I might be able to say more about welfare. Broadly, Iâd still expect itâs much better than wild mosquitoesâtheyâre fed, kept at comfortable temperatures, spared from predators, etc.
An interesting question long term would be how this affects pesticide use in urban areas. If most pesticide applications in cities aims at reducing mosquitoes, and we start using Wolbachia to control them instead, pesticide application might go down, resulting in recoveries of other insect populations and reduced sub-lethal harms. If you think bugs have good lives, yay. If you think they have awful ones, uh oh.
Iâd likely support doing thisâeffects on humans are good (which to be clear I weight very highly!), male mosquitoes seem broadly fine, and females are killed mostly as pupae, Iâm clueless about the welfare interpretation of population effects. That said, others might feel more concerned about population effects, which probably depends a lot on how you feel about insects generally and how much you think this could alter insecticide usage.
I spent less than an hour on this comment though, so obviously this is not the final word!
Hi Nickâjust regarding the team page issue, are you thinking of this page: https://âârethinkpriorities.org/ââour-research-areas/ââworldview-investigations/ââ ? It has the people listed in your screenshot from Claude.
As a reference, I got to this from CCF page --> support our work --> under the âour teamâ section. Notably, the link is from this text:
âThe Rethink Priorities Cross-Cause Fund sits on top of work done by our Worldview Investigations Team (WIT), and Interdisciplinary Research Team, groups established specifically to tackle the hard questions that most donors donât have time to engage with: how to compare welfare across species, how to reason under deep uncertainty, how to weigh present benefits against future ones, how to aggregate competing moral views into a single allocation.
The team brings together training in philosophy, economics, statistics, cognitive science, moral psychology, and decision theory . The fundâs allocations also draw on the in-house expertise of RPâs Global Health and Development, Animal Welfare, AI departments, so the cross-cause model is informed by researchers working directly in each area, not just secondary literature. â
So Iâm interpreting that paragraph as saying that WIT work goes into the report, but not necessarily that the WIT team did all the work (in particular, the interdisciplinary research team clearly was also involved, and the second paragraph suggests other teams contributed as well).
Nope, I just mean that on average, a bird who avoids a collision would live for a few more years based on typical lifespans. I meant âfully recovers from or avoids the collision.â Of course if they have some lingering injury or issue their life expectancy will be lower.
Itâs definitely correct that rats do not consume their entire daily liquid needs from contrapest in field conditions. I canât quickly find evidence for the 1.8L figureâthis study includes a DC trial, but might not be the same one (the citation links in the EA Forum post werenât working for me). Based on figure 9, 0 to 1200 mL of bait per month were consumed, depending on site and month. Thatâs not a good indicator of the effective dose, since treated adult rats might still consume some of it, but it could be a reasonable indicator of how much bait youâd need to buy to treat a given alleyâalthough for the study I link, thatâs complicated by the fact that I wouldnât take their reports of reductions particularly seriously. Issues include: camera traps have to be used quite carefully to be good estimates of population abundance, and pairing them with a bait station is not good for that; itâs hard to tell whether rats are adults or juveniles once theyâve left the burrow and especially using camera traps; thereâs no control sites, there were only two treatment sites, the treatment sites varied considerably in their features, etc.
Thanks for pointing out that the production mechanism might have an influence on land use. I think thatâs interesting and will reflect on it.
Unfortunately I think this post is otherwise not very relevant, mainly because no one uses Contrapestâthe liquid bait formulation doesnât work in any context where water is readily available because rats donât drink it, itâs extremely messy and hard for people to use. If you wanted to do this comparison more seriously, Iâd encourage you to focus on Evolve, which has a more similar profile production-wise (itâs a solid bait, like most rodenticides) and conservationists have proven more willing to try it (whether it works outside the lab is still unclear, the company has not made its test data particularly accessible for review and no independent test runs have been published yet).
To clarify, the reason I viewed (at time of writing of the previous post) that rodenticide replacement on islands might be approximately ecologically inert is that conservationists use products until rats are entirely eradicated and then stop. This means any ecological impact differences between product 1 for eradication and product 2 for eradication are essentially transient. You have this short duration of time during the eradication, and then a long period after where conditions on the island are basically the same. On reflection, youâre probably right that this is not sufficient to mean they are ecologically inert in the strict sense.[1] Overall I suppose I was imagining that people might have degrees of commitment to the ecologically inert concept, but I could have written more carefully.
A few other notes, not comprehensive:
I strongly doubt that comparing âcost to prevent one birth/âcause one deathâ is the right comparison. The population dynamics of suppressing fertility are different than those from killing adults, so the timeline to eradication (in the island context) and the application volumes would likely be different, among other things.
The dosages are difficult to determine. Although the paper you cite reports bait consumption in ideal laboratory conditionsââthe median rat (ranked by consumption) consumed an average of 74 ml of bait per day; individual means ranged from 66 to 83 ml per dayââall the male rats in this study gained weight. These are wild rats who have been taken into captivity and given ad libitum access to food, water, and Contrapestâthese are just not realistic conditions to estimate field consumptions. Senestech itself recommends 400mL tank as a 1 month supply (for an unclear area based on my quick review of the website).
Other complications: The dosing requirements also seem to be quite different between species of rats. Published studies are almost all co-authored by Senestech staff, and therefore somewhat less confidence-inducing than independent validations would be.
If weâre talking about use in settings other than islands, I want to reiterate that those are definitely not ecologically inert even to some approximation, and there are tons of things youâd have to look at besides production method if you wanted to estimate total animal effects. Age structure changes in rat populations, food web dynamics, most rodent fertility control agents appear to have reasonably strong insecticidal effects while brodifacoum doesnât affect most invertebrates weâve studied, etc.
Since Iâm not sure Evolve works either, it may be that any fertility control product used in the future has an entirely different formulation, making speculation here quite difficult.
I think Iâm just generally more comfortable than you balancing my uncertainties with a combination of direct work on arthropods, and work on things that have (1) high non-arthropod upsides and (2) where Iâm clueless about the effects on arthropods. Iâm not sure Iâm perfectly actioning my own values (given my moral uncertainty, Iâm not sure how I could be) but I would burn out if I only worked on insects all the time; itâs too depressing.
- ^
although I do wonder if the limited duration of these eradications and amount of use for island conservation relative to overall production is high enough to make a meaningful impact on land use? I donât think these things are continuous so a small enough demand shift signal might not have any impact? Just speculating...
Re: your footnote: I think this depends heavily on how severe we are talking. I donât have a strong opinion, because I really think no one has looked at it, about how much more severe things can get from disease than from something like keel bone fractures. A priori it doesnât seem unreasonable to assume that the artificial conditions of factory farming enable a chicken to live in pain much longer, and therefore have higher overall suffering, than we would ever see in the wildâbut Iâm not that confident in that idea, so it would be good to look at more diseases. The point being that a severe enough disease could still be worth working on in dalys/âdollar terms even if it doesnât affect that many individuals, and that would also make it more ecologically inert in many cases (since changing the circumstances of very large numbers of animals seems riskier).
WAI facilitated a grant from Coefficient (then OP) years ago to look at disease severity; they came out with a few papers recently here and here. As is perhaps unsurprising, but disappointing, much of the research on disease in wildlife doesnât provide enough info to do a good job estimating the welfare burden. But the high scoring bacterial zoonoses in the first paper could be a good place to start a research project attempting to better assess the severity and numerosity compared to FAW conditions (as a cost effectiveness bar).
Iâve been thinking about this approach since last year, and havenât had time to prioritize it to do detailed work on a framework, but I have some initial thoughts. I think youâre right that, if youâre comfortable with that sort of cluelessness, this kind of thing is relatively safe to do (although as @Michael St Jules đ¸ notes, Iâd also want to do some proper population modeling in a highly studied ecosystem to get some grounding for the idea).
But I think you can actually do better than focusing on only the very worst diseases depending on population parameters. For example, in populations that are top-down regulated (i.e., the population size is held under the carrying capacity of the resource by an external factor), you would not expect increases in starvation as a result of removing a disease (caveat: if that disease *is* the top down regulator, than you would have a problemâwhich unfortunately is the case in many CWD contexts). So then the disease doesnât need to be worse than both starvation and predation, say, but rather just worse than predation. The population size would equilibriate somewhere a bit higher, but the top-down regulation creates a buffer between population size and resource carrying capacity, and at high enough predation pressures you might reasonably expect almost no population increase.So I think in an ideal case, youâd identify (1) a high suffering disease that (2) affects a population primarily controlled by intense predation pressure in (3) a predator that mainly eats the target population (so the increases in predator population sizes donât affect other animals, who arenât having a disease treated and for whom this would just represent an increase in suffering).
Of course, if you have a population with high predation pressure, the target population probably dies very quickly after getting the disease, so the suffering caused by the disease might not be very long in duration. But if its a really awful disease that could still be a lot of suffering.
I donât think anyoneâs done a scan of the literature for diseases with these properties, and I doubt youâd easily find a perfect caseâmost populations are a mixture of top-down and bottom-up regulation. But I also think that probably my few hours of playing around with these ideas on the side of my other work are not likely to be the final word on the question :) so Iâm optimistic someone spending a lot more time with this could identify other âecological profilesâ of diseases that make them âsaferâ in indirect effects terms to work on than others (I think thereâs some things to say about bottom-up regulated populations as well, for exampleâprobably there you would want a disease that is a lot worse than starvation).
@Eli Roseđ¸ I think Anthony is referring to a call he and I had :)
@Anthony DiGiovanniI think I meant more like there was a justification of the basic intuition bracketing is trying to capture as being similar to how someone might make decisions in their life, where we may also be clueless about many of the effects of moving home or taking a new job, but still move forward. But I could be misremembering!Just read your comment more carefully and I think youâre right that this conversation is what I was thinking of.
This is not an unreasonable take, but just in the interest of having an accurate public record, Iâm actually the strategy director for WAI (although I was the executive director previously). Also, none of us at Arthropoda are technically animal welfare scientists. Our training is all in different things (for example, my PhD is in engineering mechanics and Bobâs a philosopher who published a lot of skeptical pieces on insects).
Basically, I think we came to Arthropoda because the work we did before that changed our minds. More importantly, I donât think the majority of Arthropodaâs work will be about checking for sentience? Rather, weâre taking a precautionary framework about insects being sentient and asking how to improve their welfare if they are. In this context our views on sentience seem less likely to cause a COIâalthough I also expect all our research to be publicly available for people to red-team as needed :)
Finally, fully agree on the extreme personnel overlap. I would love to not be co-running a bug granting charity as a volunteer in addition to my two other jobs! But the resource constraints and unusualness of this space are unfortunately not particularly conducive to finding a ton of people willing to take on leadership roles.
All very interesting, and yes letâs talk more later!
One quick thing: Sorry my comment was unclearâwhen I said âprecise probabilitiesâ I meant the overall approach, which amounts to trying to quantify everything about an intervention when deciding its cost effectiveness (perhaps the post was also unclear).I think most people in EA/âAW spaces use the general term âprecise probabilitiesâ the same way youâre describing, but perhaps there is on average a tendency toward the more scientific style of needing more specific evidence for those numbers. That wasnât necessarily true of early actors in the WAW space and I think it had some mildly unfortunate consequences.
But this makes me realize I should not have named the approach that way in the original post, and should have called it something like the âquantify as much as possibleâ approach. I think that approach requires using precise probabilitiesâsince if you allow imprecise ones you end up with a lot of things being indeterminateâbut thereâs more to it than just endorsing precise probabilities over imprecise ones (at least as Iâve seen it appear in WAW).
Thanks Eli!
I sort of wonder if some people in the AI communityâany maybe you, from what youâve said here? -- are using precise probabilities to get to the conclusion that you want to work primarily on AI stuff, and then spotlighting to that cause area when youâre analyzing at the level of interventions.
I think someone using precise probabilities all the way down is building a lot more explicit models every time they consider a specific intervention. Like if youâre contemplating running a fellowship program for AI interested people, and you have animals in your moral circle, youâre going to have to build this botec that includes the probability an X% of the people you bring into the fellowship are not going to care about animals and likely, if they get a policy role, to pass policies that are really bad for them. And all sorts of things like that. So your output would be a bunch of hypotheses about exactly how these fellows are going to benefit AI policy, and some precise probabilities about how those policy benefits are going to help people, and possibly animals to what degree, etc.I sort of suspect that only a handful of people are trying to do this, and I get why! I made a reasonably straightforward botec for calculating the benefits to birds of bird-safe glass, that accounted for backfire to birds, and it took a lot of research effort. If you asked me how bird-safe glass policy is going to affect AI risk after all that, I might throw my computer at you. But I think the precise probabilities approach would imply that I should.
Re:It might be interesting to move out of high-level reason zone entirely and just look at the interventions, e.g. directly compare the robustness of installing bird-safe glass in a building vs. something like developing new technical techniques to help us avoid losing control of AIs.
Iâm definitely interested in robustness comparisons but not always sure how they would work, especially given uncertainty about what robustness means. I suspect some of these things will hinge on how optimistic you are about the value of life. I think the animal community attracts a lot more folks who are skeptical about humans being good stewards of the world, and so are less convinced that a rogue AI would be worse in expectation (and even folks who are skeptical that extinction would be bad). So I worry AI folks would view âpreserving the value of the futureâ as extremely obviously positive by default, and that (at least some) animal folks wouldnât, and that would end up being the crux about whether these interventions are in fact robust. But perhaps you could still have interesting discussions among folks who are aligned on certain premises.Re:
What would the justification standards in wild animal welfare say about uncertainty-laden decisions that involve neither AI nor animals: e.g. as a government, deciding which policies to enact, or as a US citizen, deciding who to vote for President?
Yeah, I think this is a feeling that the folks working on bracketing are trying to capture: that in quotidian decision-making contexts, we generally use the factors we arenât clueless about (@Anthony DiGiovanniâI think I recall a bracketing piece explicitly making a comparison to day-to-day decision making, but now canât find it⌠so correct me if Iâm wrong!). So Iâm interested to see how that progresses.
I suspect though, that people generally just donât think about justification that much. In the case of WAW-tractability-skeptics, Iâd guess some large percentage are likely more driven by the (not unreasonable at first glance) intuition that messing around in nature is risky. The problem of course is that all of life is just messing around in nature, so thereâs no avoiding it.
Yeah, I could have made that more clearâI am more focused on the sociology of justification. I supposed if youâre talking pure epistemics, it depends whether youâre constructivist about epistemological truth. If you are, then youâd probably have a similar positionâthat different communities can reasonably end up with justification standards, and no one community have more claim to truth than the other.
I suspect, though, that most EAs are not constructivists about epistemology, and so vaguely think that some communities have better justification standards than others. If thatâs right, then the point is more sociological: that some communities are more rigorous about this stuff than others, or even that they might use the same justification standards but differ in some other way (like not caring about animals) that means the process looks a little different. So the critic Iâm modeling in the post is saying something like: âsure, some people do justification better than others, but these are different communities so it makes sense that some communities care more about getting this right than others do.â
I guess another angle could be from meta-epistemic uncertainty. Like if we think there is a truth about what kinds of justification practices are better than others, but weâre deeply uncertain about what it is, it may then still seem quite reasonable that different groups are trying different things, especially if they arenât trying to participate in the same justificatory community.
Not entirely sure Iâve gotten all the philosophical terms technically right here, but hopefully the point Iâm trying to make is clear enough!
Hi Vasco! As weâve discussed in other threads/âemails/âetc, we have different meta-ethical views and different views about consciousness. So Iâm not surprised weâve landed in somewhat different places on this issue :)
Bob and I make most of the strategic and granting decisions for Arthropoda, and we have slightly different views, so I donât know exactly where we will land (heâll reply in a second with his thoughts). But broadly, we both agree that we donât think soil nematodes and some other soil invertebrates have enough likelihood of being sentient to be a high priority, nor do we think that (for those that are sentient) we have a good enough understanding of what would help them to make action-oriented grants (which is Arthropodaâs focus) â in part because we donât endorse precise-probabilities approaches to handling uncertainty, and so want to make grants that are aimed towards actions that appear robustly positive under a range of possible probability assignments/âways of handling uncertainty.
That said, our confidence in our own position is not high. So, weâd be willing to fund things to challenge our own views: If we had sufficient funding from folks interested in the question, Arthropoda would fund a grant round specifically on soil invertebrate sentience and relevant natural history studies (especially in ways that attempt to capture the likely enormous range of differences between species in this group). Currently, much of our grant-making funds are restricted (at least informally) to farmed insects and shrimp, so itâs not an option.
As a result, I expect that Arthropoda is probably still one of the better bets for soil invertebrate interested donors. As a correction to your comment, Arthropoda is not restricted in focus as a matter of principle, but just has happened for contingent reasons to focus on farmed animals in its first rounds. We collaborate with Wild Animal Initiative (Iâm the strategy director at WAI) to reduce duplication of effort, and have a slightly better public profile for running soil invertebrate studies, so we expect it will generally be Arthropoda rather than WAI who would be more likely to run this kind of program. I donât want to speak for CWAW, so Iâll let them reply if they have interests in this area; but from my own conversations I doubt they would be in a good position to make soil invertebrates a priority in the next couple of years. Finally, you havenât mentioned them, but Rethink Priorities may also be open to some work in this area (Iâm not sure though).
Arthropoda treasurer hereâpretty much option 2. We are hoping to increase our expenditure next year to run an extra grants round, add a contractor to help manage some things (currently weâre almost entirely volunteer), add a bit to our strategic reserve (to carry us through donation fluctuations without needing to pause grant-making), and a few other small bits and pieces. A good chunk of this expansion can be covered by our reserves + some existing donor commitments, and 55k is about whatâs left.
We have actually a much higher room for more funding in theory, up to several million to run a couple of targeted programs we have in mind. These activities would require hiring someone to run them as a program manager as well as a lot more in grants. But weâre not really expecting EA Forum readers to fill that gap unless they happen to run a large foundation :)
haha I can confirm I did not karma knock you and I was kind of surprised you had gotten so downvoted! I actually upvoted when I saw that to counteract.
One random thought Iâll add is that since you are most experienced (afaict?) in ghd, Iâd expect your arguments to be at their best in that context, so you getting upvoted on GHD and downvoted on AW is at least consistent with having more expertise in one than the other, so not necessarily evidence that AW folks are more sensitive. Although Iâm not ruling that out!The other thing Iâm not sure I understand is how much weight a single individualâs downvote can haveâis there any chance that a few AW people have a ton of karma here, so that just a few people downvoting can take you negative in a way that wouldnât happen as much in GHD?
Thanks! I think I might end up writing a separate post on palatability issues, to be honest :)
On the intervention front, the movement of WAW folks is turning now to interventions in at least some cases (in WAIâs case, rodenticide fertility control is something theyâre trying to fundraise for, and at NYU/âArthropoda Iâm working on or fundraising for work on humane insecticides and bird window collisions). I just meant that perhaps one reason we donât have more of them is that thereâs been a big focus on field-building for the last five years.For field-building purposes, thereâs still been some focus on interventions for the reasons you mention, but with additional constraintsânot just cost-effective to pursue but also attractive to scientists to work on and serves to clarify what WAW is, etc., to maximize the field-building outcomes if we can.
Sorry Iâm probably missing something, but Iâm not understanding why real world examples from EA would be particularly relevant given how young a movement it is. I think someone could grant that we have the ability to be justified in assigning probabilities to things that are likely to happen soon, and agree that the risk of things weâre totally unaware of happening in the next ~ 10-50 years might be (at least in some circumstances) sufficiently small to not have unawareness problems.
But once you start trying to be an impartial altruist about far future beings, that seems to me where you really canât get away from unawareness problems. And so I guess if you wanted to convince me I was wrong about that, we should be looking at things that people thought 1000 years ago, and how things they caused today were bad even though they were trying to do good for reasons they werenât only poorly calibrated on but in fact totally unaware ofâand it just seems likely to me there would be tons of examples of that?
Maybe the development of gunpowder stands out here as something being pursued in the hopes of achieving eternal life (ostensibly an altruistic motivation) and presumably the possibility of guns was not on peopleâs radar. I guess it would eventually have been figured out anyway, but how much harm did having gunpowder X years earlier cause?
Maybe an objection here is that an âidealâ agent would have of course considered the possibility of any chemical work being misused, but IDKâthey werenât even trying to make something explosive. I donât see how even a perfectly rational being could have predicted all the harms gunpowder would cause given that they were aiming to do alchemy. What probability could they have possibly been justified, given their epistemic position, in assigning to âsuper bad outcomes from pursuing eternal life chemistryâ given that they probably could not have imagined the scale of modern warfare?
I do get a little mixed up on this between âpeople are not ideal and so regularly make large mistakes that look like cluelessnessâ vs âeven an ideal agent could not be justified in their probability assignments given what is theoretically knowableâ so maybe Iâm misunderstanding something.