🏅 Top traders

Alright. I don't expect to change my mind within two weeks, but at the same time I'm not fully comfortable resolving to NO. What do people think of a resolution at 10%?

@NicoDelon seems more than fair

I agree that the kinds of interventions the paper critiques are implausible. I still think there's a lot of gray area around what counts as anthropogenic suffering, and also many plausible interventions that could effect (though certainly not eradicate) wild animal suffering.

What about policy decisions about the reintroduction of predators? Seems like an EA-like NGO could potentially have some effect on which laws get passed and how they are implemented. This could in turn could affect the population and quality of life of wild prey animals across entire countries.

@JoelHancock do you mean that reintroducing predators could eliminate WAS or that not reintroducing them could? If the latter, then no, I don't think we have anything close to a blueprint for how reducing predators population would have a net positive effect on suffering, much less address the large-scale problem.

@NicoDelon Thanks for your reply.

Reintroducing predators definitely wont eliminate WAS.

I suppose it comes down to the semantics of 'tractable' - does have a problem have to solvable in its entirety to count (Which WAS is certainly not, with foreseeable levels of technology), or is merely enough that EA could in theory, by addressing the issue, have an effect per units of money/effort similar in magnitude to standard causes (in which case it's certainly still debatable, but I'd say has odds >20% and certainly can't be ruled out without more research.)

As to the specific question "would introducing predators increase or decrease the total amount of suffering?", I am actually not sure of the answer, though I lean decrease. I just have a hard time believing it would be anywhere close to zero.

@NicoDelon did you read this interview we did with Jordan Hampton? Not, I don't think, anything new to you, but curious if you caught it and/or if you know Jordan. https://asteriskmag.com/issues/02/animal-welfare-in-the-anthropocene

Animal Welfare in the Anthropocene—Asterisk

Wild animals outnumber farmed animals by orders of magnitude. Are we obligated to help them? And even if we wanted to, do we know how?

@JCE No but I look forward to reading it. Thanks!

Requesting yet more clarification on "tractable", which could have several meanings here:

1. Tractable = seems like humanity has a good shot at solving or kinda/mostly-solving this issue. (In this sense, climate change is similarly tractable today -- it seems like by 2100 there's a good chance that humanity will have climate change under control. For animal welfare, it seems like a true solution to the problem would either require godlike mastery of the natural world probably involving superintelligent AI, or else some kind of clever genetic hack like the FAAH-out project: https://www.faroutinitiative.com/)

2. Tractable = seems like we have thought up some interventions which are super-duper worthwhile on a marginal, bang-for-buck basis, and EA organizations are excited about funding these interventions, but we're just taking marginal bites out of the problem and we aren't even really trying to solve the entire problem. (In this sense, neartermist global health and development causes are tractable today -- buying bednets to prevent malaria is super worthwhile in terms of saving lives, but at the same time we're not trying to completely solve poverty or anything. Or consider a scenario where some researchers coming up with an awesome plan to eliminate a certain type of painful deer parasite using some kind of vaccine or gene drive at very low cost. OpenPhil loves the idea and funds it, and many deer consequently suffer less. There is hope to deploy similar gene drives to eliminate other parasites in other animals, but this obviously isn't a solution to the entire problem of wild animal suffering.)

3. Tractable = seems like we have thought up some interventions which are somewhat worthwhile on a marginal, bang-for-buck basis... not so effective that EA organizations are excited about funding them, but tractable enough that (if the government cared about wild animal welfare) it would be on par with other government spending on environmental / social / etc causes. (Basically, what if some researchers come up with the idea mentioned in #2, and it doesn't quite meet EA's exactingly high standard of effectiveness, but it still seems "tractable" in the normal sense of the word -- the sense in which PEPFAR is a tractable way to reduce HIV deaths, or renewable energy subsidies are a tractable way to fight climate change, or whatever.)

@JacksonWagner Thanks for your questions and I’m sorry I missed them. For the purposes of this market I’m thinking of WAS in the first sense as a large-scale, general issue. I already believe that some particular issues will become tractable in the near future, such that EA orgs, among others, will have good reason to fund related interventions that can have notable marginal impacts. I don’t think we’re anywhere near even coming to terms with what it would take for the general issue to be tractable in that sense. That we can imagine a super intelligence someday devising solutions to that general problem is, to me, irrelevant. We have no idea what such an intelligence would actually be like and how to get there.

@NicoDelon And FWIW if you want more calibration, I don’t think climate change is tractable in this century in the sense that matters here. It’s something we can mitigate; not something we’re going to get rid of.

I’ve decided to sell my positions in all my subjective markets. Not because I’ve changed my mind one way or another. The positions I’ve held until now roughly reflect my current credence.

Why doesn't "capture/kill/extinct all the wild animals" work as a solution? Obviously those all sound terrible, but we're in utilitarian ethics land here, so who knows

@jonsimon Nicolas thinks this would cause human extinction, and also that wild animals may have net positive lives.

@MartinRandall

wild animals may have net positive lives.

So what's the problem then? Why not leave them alone and consider the problem solved?

@jonsimon Well for example humans may have net positive lives, but there is still a problem of human suffering.

@MartinRandall I see. Yeah sounds pretty impossible, if we're talking about any remotely-plausible solutions. (Something like "capture and wirehead all animals in the world" would presumably work, but is silly)

@jonsimon @MartinRandall To be more precise, I’m not confident they have net negative lives. That doesn’t mean there’s not a lot of suffering to be concerned about that I (as of now) think is intractable.

Re extinction. Yes, that would count if (1) there were no catastrophic downstream effects and (2) it could actually be done. But I think there would be such effects and despite our best efforts we’re not remotely close to being able to wipe out wild animals.

Would more humane slaughter of human caught wild fish resolve this market yes?

@MartinRandall

@NicoDelon ?

@MartinRandall Sorry I missed that one. No since that would be anthropogenic suffering. Will clarify.

When we accidentally turn the world into paperclips it will probably reduce animal suffering.

@MartinRandall If wild animals do not contribute to energy or materials needed to produce paper clips that’s certainly possible.

@NicoDelon Would you mind saying more about what you mean by "can become tractable", then? It seems clear that sufficiently advanced technology could one day make it tractable even if it's not tractable now.

@StevenK If I believed such technology was forthcoming—say, within the next century—then that would surely shift my credence. As of now, I don't see what concretely could make the large-scale issue tractable (in the sense used by EAs for cause prioritization), including what such technology would look like. Do you have something concrete in mind?

@NicoDelon I think the extinction of wild animals would reduce wild animal suffering over a sufficiently long time horizon, and that seems all-too-tractable.

@MartinRandall The extinction of all wild animals would likely mean the extinction of all life and civilization.

@NicoDelon I think Martin is talking about whether it could be done and you're talking about whether it should be done. How about much more advanced (maybe superintelligent) modeling of ecosystems? The linked abstract already says that "we need to develop models that predict the effects of interventions on biodiversity, ecosystem functioning, and animals’ well-being". Does that not count as a way in which it can become tractable?

@StevenK when EAs evaluate the tractability of an issue, they typically don't count the extinction of life and civilization as a viable pathway. So, no, eradicating all wild animals is not a way to make wild animal suffering tractable, unless we can make it so it doesn't have the catastrophic outcome I have in mind. But if you could convince me we can, then you might change my mind! (I happen to think there are reasons not to eradicate wild animals but I'm not counting them against tractability here.)

@StevenK I updated the description. I'm not interested in speculation about logical possibility. I'm interested in the actual tractability of the issue based on our current best understanding of the relevant science and technology.

@NicoDelon This was meant as a response to Martin, right? I'm not talking about extinction, I'm just saying that sufficiently good modeling could prevent the kinds of unpredictable backfiring of interventions that it looks like you worry about in the paper. I don't mean that as just a logical possibility; I think if we survive then we'll probably be able to invent most of the technologies we'd need in this century because of AI contributing to R&D. But it sounds like you disagree, so the answer might depend a lot on how far you think we'll get in those 100 years.

@NicoDelon Do you think that wild animals currently have (or historically had) net positive lives? If you do then I agree that makes the problem less tractable in the sense you mean.

If not, then humanity has already significantly reduced the number of wild animals. I'm reminded of https://xkcd.com/1338/ . Wild mammals are rare and becoming rarer.

I suppose the few remaining wild mammals could be critical to our survival, but this honestly seems unlikely to me, I think we're more dependent on insects and fish and such.

Relatedly the leading plan for reducing farm animal suffering is to dramatically reduce the number of farmed animals via reduced demand and increased costs. So there's a precedent for reducing suffering by reducing numbers.

I suppose I'm also assuming that a problem is tractable if harm can be reduced, even if eliminating the harm is impossible.

@NicoDelon I apologize for not noticing that you're an author on the linked paper, so I will defer to your expertise.

@MartinRandall No apologies needed and please don't defer! I'm enjoying the discussion.

I don't know if wild animals have (had) net positive or negative lives but I tend to have a more optimistic take than most folks in the area such as Oscar Horta, Brian Tomasik, Kyle Johannsen, or Will MacAskill.

Yeah, my concern is smaller animals, whose lives are more likely to be bad but whose existence is more crucial to ecosystems.

I agree that reducing numbers, all else equal, is a way to reduce suffering. If that happens to be a viable pathway, I'll consider it. I'm, as of now, skeptical we can make a dent on a large scale given the scope of the issue and the number of animals we're talking about.

@StevenK I guess it was in response to you and Martin. Sure that's conceivable, but it's trivial to say something like: if (say) AGI is such that it can solve even the most complex issues then it can also solve that issue. This makes any potential issue trivially tractable in some sense if the probability of such AGI is non-negligible. But surely that's not how EAs assess tractability in the relatively short term (on a 100 years horizon, as I stipulated).

@NicoDelon Future AGI doesn't make problems tractable now, but I don't see how it wouldn't make them potentially tractable later according to how EAs assess tractability. Regardless, I'm happy to predict on this market on the basis of something like "will Nicolas believe that wild animal suffering may become tractable even in the absence of AGI or other developments that would make many formerly intractable problems tractable".