Note: This post is a discussion of charity cost-effectiveness with specific focus on causes that interest me. However, I hope that other readers will find the discussion valuable, which is the reason I wanted to make the conversation public.

For a while, my general plan has been to make money in the corporate world that I can use to fund projects to reduce suffering. I'm still very much uncertain about exactly which causes are most cost-effective, so I intended to invest my earnings until I was older and wiser. However, for several reasons I now think it would be financially advantageous for me to begin donating sooner. I haven't yet determined the exact amount, but it would probably be at least $15,000 per year. In addition, it must be given to a US-based public 501(c)(3) charity (not a private foundation).

I'm very interested in organizations whose work would increase the number of people that care about animal welfare -- especially in the utilitarian sense of wanting to reduce animal suffering as much as possible, including that of animals in the wild. (It would be a shame, for instance, to increase the number of people who care about "animal rights" because of an ideology that humans should not be interfering with the natural world.) More generally, I would like to find cost-effective ways to make people care more about various types of suffering (or possible suffering) that usually go ignored: e.g., on the part of insects, simulations, etc. Being an irrational human, I incline toward more immediate and tangible benefits (e.g., preventing chicken suffering during slaughter) even if they have lower expected value, but I may also take Eliezer Yudkowsky's advice and split my donation into two or more pieces, one for feel-good value and one for maximum expected benefit.

The purpose of this post is to help me figure out where exactly to give. At the moment, I would most like to donate toward promoting concern for the suffering of wild animals, among academics and/or the general public, as described here. However, I'm not aware of any existing organizations (much less 501(c)(3)'s in the US) engaged in this work -- do readers have any ideas?

Perhaps there's a cadre of animal-welfare supporters who would like to help found such a group? If so, I could provide seed funding once it was registered as a charity. It needn't be a particularly large undertaking -- perhaps just a legal structure set up to funnel at least $15,000 toward high-quality academic work or public outreach. Alternatively, perhaps there's an existing animal-welfare organization on which I could piggyback such work? (The key, of course, is to find projects that wouldn't be undertaken without that funding. There's no benefit in, say, providing money to a researcher who could almost as easily pay for the project through grants from the government or private foundations.)

If I can't find or jump-start an organization to work on wild-animal suffering, there are other possibilities I might consider. Below I've listed just a few organizations that came to mind, but there are plenty of others (please feel free to suggest more!).

I encourage representatives of organizations that would like funding to answer this question: If you got an extra donation of, say, $15,000 next year, what specifically could you do with it? What additional project(s) would be supported that wouldn't otherwise have been possible? How exactly would those projects help to reduce the expected amount of suffering in the multiverse?

My hope is to make this process as transparent and public as I can, so that, among other things, readers will benefit from the exchange (similar to GiveWell's Clear Fund), so I encourage replies on the forum. However, I realize that there may be some things better discussed in private; if so, I encourage readers to write to me at the email address given on the front page of my website. Let me know if any portions of your emails can be posted on this page.

Here are some example charities:

1. Vegan Outreach

This organization is a small but genuine group of activists who achieve some impressive results. In terms of direct impact on animal suffering in factory farms, I've estimated that each dollar donated prevents between 100 days and 51 years of farmed-animal life-years and between 2 and 358 slaughters. Moreover, the organization has a practical, utilitarian mindset of the type very much needed within the animal-welfare movement.

But perhaps the most important impact of Vegan Outreach's work is to expand the base of people who are aware of animal suffering and want to do something about it. While I often find that such concern can be misdirected toward low-priority issues, I hope that at least a few of the animal activists that Vegan Outreach creates end up working on more important efforts. In the long run, I hope that such people will think about the immense amount of suffering that occurs in nature to the point that -- eventually, if and when it becomes possible to address the problem without causing more harm than good -- there is support for action on the issue. At the very least, I hope such people would think more circumspectly about actions that could vastly increase the amount of wild-animal suffering.

I have two main worries with donating to Vegan Outreach, both related to wild animals.

The first is the one raised here: Namely, does vegetarianism increase the number of animals suffering in the wild? If such animals have sufficiently low welfare, then might vegetarianism actually cause more animal pain than it averts?

Second, I fear that some of the animal-rights activists created by Vegan Outreach's brochures may, as suggested previously, adopt the view that humans ought to avoid interfering with animals entirely, which could lead them to oppose efforts to reduce animal suffering in the wild. My guess is that the fraction of such people would be relatively small -- especially among the newly created activists, since many of the people holding that type of extreme non-interference view are likely already vegans. And this is probably much less a concern with Vegan Outreach than other pro-veg organizations because Vegan Outreach's brochures tend to have a consequentialist slant.

2. Humane slaughter

A second approach to reducing suffering on factory farms is to improve welfare conditions, through legislation and campaigns encouraging corporations to commit to better standards. I think of this as the conservative, low-risk but low-expected-value option. Unlike Vegan Outreach, I see no real potential adverse consequences here, since the number of farmed animals -- and hence the environmental implications -- probably wouldn't change much in either direction. (Indeed, if people see meat as more humane, they may be less averse to buying it. On the other hand, the public outreach associated with these campaigns would involve depictions of the cruelty of factory farming, which might lead some people to go vegetarian.) On the other hand, this option is wanting in precisely the domain where Vegan Outreach is strongest: Changing society's long-term attitudes toward animal suffering. While the impact here is still probably positive (for instance, because changed laws or corporate policies will tend to normalize the notion that animals deserve humane treatment, similarly to the way in which anti-discrimination laws have helped to normalize concern for people of all races), I would guess it's considerably smaller than the direct, person-to-person influence that leafleting has.

How do the direct factory-farming impacts of this approach compare? Here's a crude attempt at an analysis. One example of a welfare campaign that would impact a large number of animals is the Humane Socity's effort to include birds in the Humane Methods of Slaughter Act (HMSA). The number of chickens slaughtered in 2008 was roughly 9 billion, and will likely remain at a similar level for at least, say, 20 years before in vitro meat is developed and adopted. A change to the law would then conservatively impact on the order of 200 billion animals. If we suppose that the law would -- due to poor enforcement, etc. -- reduce the painfulness of slaughter by only, say, 1/4, the impact would be equivalent to preventing roughly 50 billion instances of slaughter.

Estimating probabilities of success as a function of money spent is difficult, but here's a stab. Suppose it would take $10 million to increase the probability of success by 50%. (As an illustration, the Yes on 2 campaign in California raised $10.6 million, though this figure obviously ignores the enormous contributions of volunteers to the effort.) The expected number of equivalent slaughters prevented per dollar would then be (50 billion) * (0.5) / (10 million) = 2,500, which is higher than even the upper bound for Vegan Outreach. This remains true even if we increase the spending amount as high as $65 million. Of course, vegetarianism does much more than prevent slaughter -- most of the benefit comes from preventing ordinary life on a factory farm -- so the comparison isn't completely fair. I'm also assuming that the change in probability of success is linear in the amount donated, which isn't true in practice. Finally, Vegan Outreach's work is itself connected with political efforts like covering birds under HMSA, because some fraction of the people that Vegan Outreach reaches will go on to lobby for improved animal welfare.

Of course, in addition to covering birds under HMSA, there are lots of other farm-animal-welfare efforts, such as HSUS's cage-free-egg campaign, or maybe promotion of controlled-atmosphere killing. Any recommendations on one that's particularly cost-effective?

3. Singularity Institute for Artificial Intelligence (SIAI)

I've written elsewhere with a few comments about friendly artificial intelligence. I support the idea in theory, but it's less clear to me exactly what would be accomplished by a donation to SIAI: What specific types of research projects would be funded? And is there reason to think that research wouldn't take place otherwise? Would there be any way to direct my donation toward particular topics that most interest me: e.g., infinite decision theory or examination of dystopic future scenarios?

As I mentioned, I welcome suggestions on other charities to consider. Thanks in advance for the comments!

Despite the lack of public conversation on this forum, a lot of discussion of my question has been going on in private email. While no formal commitment has yet been made, I'm leaning toward the third of the options suggested above: research by SIAI. So that the details don't get lost in a comment, and in order to open up the topic for more general discussion, I've started a new thread on the subject here.

I'd missed this thread, sorry Alan. I've got a couple of thoughts on the subject, but need to sign off now. I'll try and reply shortly.

Personally, I would donate to New Harvest. In vitro meat will end many of these problems, it's simply a matter of the money to get there. Your $15,000 won't produce in vitro meat tomorrow, but it would help greatly I'm sure and could help get the attention of a philanthropist out there who would fully fund in vitro meat. It would probably feel good to know that you contributed to such an important cause at the beginning and helped bring about what I would classify as a revolution.

If you want immediate impact, however, I would vote for Vegan Outreach.

Thanks for the suggestion, brightmidnight.

A number of people have suggested New Harvest, and I agree that it's a solid organization. My primary concern is that, as I suggested in the main post, I fear vegetarianism could -- depending on the balance of suffering in nature and the animal-population impacts of climate change -- cause more animal suffering than it averts. The main reason I favor Vegan Outreach and The Humane League is actually not the direct reductions in meat consumption that they effect (whose net animal-suffering impact is, I fear, ambiguous to me), but the long-term ideological concern for animal suffering in general that they help to build among the public. The latter might, I hope, eventually translate into concern for wild animals. In contrast, the technological fix that New Harvest would facilitate produces comparatively less ideological meme propagation. (An exception is if you think that a significant amount of the resistance to giving animals ethical consideration is the fact that people don't want to give up eating them. A technological fix might allow people to realize the barbarity of their previous attitudes toward factory-farmed livestock, similarly to the way in which the industrial revolution made it easier for people to see the barbarity of slavery.)

I wonder how hard it would be to find the place you want to donate to, and just advertise for them.

DanielLC, interesting idea. Advertising can indeed be effective.

I guess my question would be, Why not donate to the organization and let it decide whether advertising is the most cost-effective use of funds? Doing so might not be a good idea, though, if (1) I didn't trust the organization's judgment on resource allocation for fundraising / advertising (e.g., because the people who work for the organization were too focused on tangible results to be persuaded by cold figures about fundraising ROI), or (2) the organization was hesitant to devote more funding to advertising for fear of jeapordizing it's statistics about "% of budget spent on programs."

In any event, the suggestion about paying for advertising doesn't work in my particular case, because my donation needs to go directly to a 501(c)(3) for various reasons.

Alan, I could sum up your view of in vitro meat by providing a technological solution to the problem of meat industry, we would avoid the root problem of animal suffering. The current state where people feel ethically conflicted between promoting their own interest in tasty food and animals interests will eventually be overturned anyway, and when it does, there may be an anti-animal suffering movement with sufficient momentum to reach into the wild.

While I think that's a decent suggestion, I think the more intuitive and opposite suggestion is far more likely. That providing in vitro meat will sharpen the distinction between ethical and unethical behaviour. That adoption of veganism will not dampen a movement towards concern for animal welfare nor will it energise such a movement. Rather, it will cause the movement to have less work left to do since the views of the public have converged towards the utilitarian point of view.

To DanielLC, I think advertising for an organisation instead of donating to an organisation is a sensible idea. It's for Alan's reason 2. For organisations to spend their revenue on advertising is difficult. The % of budget reaching those in need decreases. However, Alan needs to choose between the 501c3 organisations in existence in this instance.

Alan, what are the consequences of not giving to a 501c3 (and what is the category)? Presumably they're not absolute, so they might be worth taking a hit on for a good enough cause.

Re your dollar worth essay, I'm wary that you seem to have taken a lot of Vegan Outreach's claims as read. From some of the quotes, I suspect they're strongly overstating the case - perhaps not deliberately. For example, the claim that 10% of the US population would be willing to read them and that 2.5% of the readers will turn vegetarian are - they admit - basically fabrications. The costs per leaflet are suspicious, too - as you print more leaflets, the cost of effective distribution is likely to rise sharply, since you won't be able to get rid of them all on university stands any more.

That said, there's a lot of leeway for them to be much less effective than they claim and still much better than most other charities. And they definitely seem more convincing a cause than the SIAI (I'll try and post a few last thoughts on that over the weekend).

Thanks for the continued comments.

RyanCarey, I agree it's not obvious that in-vitro meat would be beneficial from an ideological perspective. The fact that many people disapprove of animal suffering in factory farms may provide a wedge to encourage them to care about animals in some context, which is then perhaps more likely to spill over into other areas (like the wild). Without factory farms, the amount of massive-scale mistreatment of animals by humans that's not potentially justified for utilitarian reasons would be much smaller.

Arepo, I won't go into the details of my situation, since I have at this point already made a commitment regarding SIAI. But you're right that the consequences are not absolute.

Vegan Outreach's 10% claim doesn't seem intuitively unrealistic to me -- for instance, roughly three times that many (28% in 2004) have a college degree -- but I could be wrong. I agree the 2.5% figure sounds high, but that's why it's an upper bound. What do you think of Jack Norris's 1 in 235 point estimate? That could still be a little high due to random noise, I suppose.

I think the biggest unfounded part of the calculation is that more money alone will translate into more booklets distributed. It may have been that VO distributed 3,879,329 as of January 31, 2005 at a cost of roughly $5 million, but that doesn't count the large unpaid efforts of volunteer leafleters. In order for more money to translate into more booklets handed out, VO would presumably have to hire paid leafleters, which would increase the marginal cost.

But I agree with your later comment, "That said, there's a lot of leeway for them to be much less effective than they claim and still much better than most other charities." That's the basic point of the paper, even if the exact numbers are off by as much as an order of magnitude or two.

Alan Dawrst wrote:Arepo, I won't go into the details of my situation, since I have at this point already made a commitment regarding SIAI.

How long term is the commitment? I was going to argue further against them specifically in the other thread, but it's prob not worth it if you're bound (either emotionally or legally) to giving them money for a long time to come.

Vegan Outreach's 10% claim doesn't seem intuitively unrealistic to me -- for instance, roughly three times that many (28% in 2004) have a college degree -- but I could be wrong. I agree the 2.5% figure sounds high, but that's why it's an upper bound. What do you think of Jack Norris's 1 in 235 point estimate? That could still be a little high due to random noise, I suppose.

The problem is basically the same problem I have with SIAI - the error with fudged numbers can accumulate rapidly, and disguise omissions. For eg how many of the people who came up to Jack Norris were already on the verge of vegetarianism, such that the leaflet itself only did a fraction of the real work by pushing them past a tipping point? If that's the case, such people might have just gone vegetarian anyway (and if the tipping point is fine, they might go back to eating meat shortly), and the expected # of animal saved by happening to nudge them over the line would be much lower. It's alsy quite possible that Norris suffered confirmation bias, hearing/remembering 'I've stopped eating meat' rather than 'I presently intend to stop eating meat' and 'I intend to stop eating meat' rather than 'I'd like to stop eating meat'. As far as I understand he had no formal way of recording this, and has just come up with the 1-per-uni figure after the fact.

Another issue: given that he'd probably given the leaflets to the keenest people, it might well be that about 1 person at each uni was persuadable by such methods, and many others will take a leaflet because it's there. If so, money put into VO could be completely wasted, and perhaps we should be encouraging them to spend less on distribution.

I think the biggest unfounded part of the calculation is that more money alone will translate into more booklets distributed. It may have been that VO distributed 3,879,329 as of January 31, 2005 at a cost of roughly $5 million, but that doesn't count the large unpaid efforts of volunteer leafleters. In order for more money to translate into more booklets handed out, VO would presumably have to hire paid leafleters, which would increase the marginal cost.

Precisely.

But I agree with your later comment, "That said, there's a lot of leeway for them to be much less effective than they claim and still much better than most other charities." That's the basic point of the paper, even if the exact numbers are off by as much as an order of magnitude or two.

The thing is they could easily be off by multiple orders of magnitude.

I'm not really trying to make a point about VO specifically, so much as overcredulity towards Pascal's button causes. VO's case actually seems quite simple - there aren't that many places where their estimates have known unknowns, and they've actually given some numbers that, if correct, have a heavy bearing on their effectiveness. There's still a huge amount of room for donations to them to be much less effective than they claim them to be. But compared to a group like SIAI, they're virtually transparent.

Arepo wrote:How long term is the commitment? I was going to argue further against them specifically in the other thread, but it's prob not worth it if you're bound (either emotionally or legally) to giving them money for a long time to come.

The commitment is for one year only. I would be very curious to hear your further criticisms, but I also suspect the probability of their changing my behavior will be relatively low. Still, I have been persuaded to change my mind on many occasions in the past, so it might be worth your effort.

Arepo wrote:For eg how many of the people who came up to Jack Norris were already on the verge of vegetarianism, such that the leaflet itself only did a fraction of the real work by pushing them past a tipping point? If that's the case, such people might have just gone vegetarian anyway (and if the tipping point is fine, they might go back to eating meat shortly), and the expected # of animal saved by happening to nudge them over the line would be much lower.

A good point. FWIW, I did assume in the lower-bound estimate that only 1/3 of people would remain vegetarian. The concern about just tipping over the edge for people already most of the way there is an important one.

Arepo wrote:The thing is they could easily be off by well over an order of magnitude.

They could be, sure. But what's your probability that they are? If there's, say, a 10% chance they're only off by at most a factor of 10, then the expected value can be at most two orders of magnitude away from the estimates given.

Another factor we didn't consider suggests a higher-than-calculated cost-effectiveness: Namely, the long-term impact of creating new vegetarians on things other than the factory-farmed animals they personally avoid creating. These impacts include spreading vegetarianism to their friends, possibly lobbying for improved animal-welfare standards, and maybe also donating back to Vegan Outreach or becoming a volunteer. Plus, there's value to the other work that Vegan Outreach does as an organization, like reaching people through their website, and generally promoting a utilitarian ethic among the animal-advocacy and progressive communities.

By the way, thanks for your skepticism about these figures -- I agree they're perhaps less trustworthy than the piece implies. Skepticism is good, though only to the extent that it helps correctly calibrate your probability estimates. I don't think it's inherently the case that people should only donate to organizations that can prove beyond a shadow of a doubt that they accomplish positive outcomes. If I had to donate tomorrow, and I had a choice between a charity that I knew would accomplish +10 utilons vs. a charity that I thought might be a fly-by-night organization which would accomplish nothing (80%) but might actually be effective and produce +60 utilons, I would choose the latter. Of course, in practice, the right action here would probably be to do more detailed research before making the donation, but sometimes there are inherent limits to doing so (like quantifying the expected impact of SIAI, which can only be done to some extent, and never to the point of convincing proof).

Alan Dawrst wrote:The commitment is for one year only. I would be very curious to hear your further criticisms, but I also suspect the probability of their changing my behavior will be relatively low. Still, I have been persuaded to change my mind on many occasions in the past, so it might be worth your effort.

We're not built to be very good at changing our mind. If I'm arguing against you it's either for my own benefit, or because I think even with the low odds of changing your mind, the stakes are high enough that I still have a high expected gain

They could be, sure. But what's your probability that they are? If there's, say, a 10% chance they're only off by at most a factor of 10, then the expected value can be at most two orders of magnitude away from the estimates given.

My probability that they are would be a number plucked from the nebulous interrelations of my brain, information, and instinct, so I don't want to give it more credence than it deserves by specifying it. That said, in a worst case scenario, if everything I've said were true, then the expected value might be modified something like as follows (using the upper bounds my intuition will give, since they've done the same):

Confirmation bias leading to overreporting: x 1/10 (if this seems harsh, I'm also thinking the neo vegetarians might be overly willing to praise an activity they see as worthwhile. I think this could plausibly be higher still, given the power of this bias when there's nothing checking it)

Increasing cost of distribution: x 1/2

Risk of leaflets already reaching all the relevant targets x 1/20 (again, if this seems harsh it's because they're extrapolating from a sample of 1 per uni. Again, I suspect this could be much higher. If they start distributing further afield then this factor might diminish, but the cost of distributing one would raise much higher)

Environmental harm from using up paper and coloured ink: x 9/10

Risk of them decreasing some amount of total happiness due to logic-of-the-larder considerations, given that we already know they're quite ethical people who might source their meat humanely: x 14/15

Diminishing human contact per leaflet given reducing the power of the leaflet's information: x 1/4

Risk of provoking an adverse reaction by groups threatened by veganism itself, or the philosophies it entails: x 4/5

Extra time taken to distribute the extra leaflets by unpaid volunteers who might have put it into other activities: x 9/10

Total modifier to expected value: x 0.000378, ie what you thought would cost $1 might cost $2645.

Another factor we didn't consider suggests a higher-than-calculated cost-effectiveness: Namely, the long-term impact of creating new vegetarians on things other than the factory-farmed animals they personally avoid creating. These impacts include spreading vegetarianism to their friends, possibly lobbying for improved animal-welfare standards, and maybe also donating back to Vegan Outreach or becoming a volunteer.

Granted. I'm really not attacking VO - despite all of the above, I imagine they're an extremely good cause.

Plus, there's value to the other work that Vegan Outreach does as an organization, like reaching people through their website, and generally promoting a utilitarian ethic among the animal-advocacy and progressive communities.

But if you're donating money specifically to the manufacture of leaflets which require time and resources to distribute, you're marginally reducing the time they have to spend on other activities, so I don't think you should claim this as a side benefit.

I don't think it's inherently the case that people should only donate to organizations that can prove beyond a shadow of a doubt that they accomplish positive outcomes.

Obviously I agree with this in principle. But in practice, obvious uncertainties tend to come with hidden ones. Say you're choosing for some reason between giving $10 to a (genuinely starving) beggar today and giving $21 to a courier to give to him tomorrow. A mutual friend tells you with 50% probability the courier will take all of the money to the destination, and with 50% probability he will keep it all as a fee.

In a clean thought experiment, you obviously want to use the courier. In the real world, all sorts of factors creep in to the adjust the expected value:

Mutual friend has fooled you, and is actually working with the courier to scam you (x 499/500)

Mutual friend has misunderstood the courier service terms (could go either way, but given that we can selectively rule out cases where he misunderstood it to be much worse than it is - since he would have told you not to use it - x 19/20)

Courier is genuinely waylaid by some mishap (x 9/10)

Courier has always been honest in your friends' dealings, but is tempted by the money to go dishonest (x 39/40)

Beggar finds a way to get food between now and tomorrow, heavily devaluing the effectivenness of your money to him when it arrives (x 1/2)

And so on. Now you're only expecting him to receive the equivalent of $4.37, and in practice there would probably be dozens more marginally modifying factors, and perhaps a couple more heavily modifying ones.

So I claim that you're not paying nearly enough attention to the practicalities of real world long-shots. As R-WLSs go, the SIAI seems to be a prime example. One could spend days listing and weighting the potentially complicating factors of donations to them, because they're so far removed from the eventual good they're supposed to be causing. I wasn't going to do so here, but while I'm in the mentality, I might as well kickstart a list (the contents of this will depend on what the utility payoff is and why, but I can start with plenty of generic factors). Here I'll assume the alternative is a group like Fred Hollows, who have relatively few steps between you giving the money and it doing its job, and I'll go with numbers that sound plausible to me, rather than upper bounds. And I'll take the modified figure to be something like (((expected increase to utilons given that the SIAI make a key contribution to preventing extinction at the hands of an AI) x (the SIAI's estimate of the probability of them making their key contribution))/(their expected funding between now and then + your potential contribution)):

Cognitive biases causes SIAI to overestimate the size of the threats they're addressing: x 2/3

AIs will not be developed for much longer than we expect, meaning that the money would be better invested closer to the time: x 1/2

The specific properties of AIs will differ enough that their solutions become irrelevant: x 1/3

The HR at SIAI suffer cognitive bias and/or a heavy weighting of applications from people who share their mindset, causing them to have too homogenous a staff and subsequently failing to consider vital alternatives: x 1/5 (they don't have nearly enough real mathematicians from what I've seen of them, and hence sometimes write claims that I've seen extremely competent mathematicians dismiss as trivially false or fatallly vague, eg Bostrom on the Doomsday Argument, Eliezer Yudkowsky on Newcomb's Paradox)

The work they're doing is simple enough/difficult enough to distribute among multiple researchers/rewarding enough at a low pay rate that they already have adequate funding or will be able to find it: x 1/8 (they have plenty of extremely wealthy benefactors)

Any intelligence capable of the calculation necessary to eliminate the human race will either be benevolent towards us or a utility monster capable of giving itself far more utilons than the human race can generate, and in either case is not a bad scenario: x 1/2

An AI wiping out all life in the solar system would actually be a good thing because net utility if it doesn't is negative: x 3/4 (this number must be closely related to the initial number - the larger number number of utilons SIAI guesstimate preventing an extinction would save, the more probable it will be that they've overestimated the happiness of life)

Funding more immediate causes will generate more public interest and provoke more donations, eg under a well-publicised and permanent donation-encouraging scheme along the lines I suggested here, or better: x 3/4

Major economic upheaval, caused eg by peak oil or climate change, devalues your currency before it's actually been used and prevents it from causing any or as much good as money which is spent immediately: x 2/3

The SIAI will find an optimal response to the AI problem, which will turn out to have lower expected utility per resource spent on it than other issues, therefore effectively being a resource sink even conditional on its payoff: x 1/100 (this should probably be even higher - the claim SIAI and their donors are making is that something they haven't almost no idea of the probability of - and little idea of its magnitude - is a bigger concern than any other in the world)

Global extinction/societal collapse before, during or shortly after development of AIs meaning the money is wasted: x 1/2 (Nick Bostrom's guesstimate)

The permanent colonisation of nearby worlds reduces the existential risk of a near-worst-case AI: x 9/10

Independent AI researchers turn out to be smarter than the SIAI gives them credit for, and, on first creating a real AI, do so in a way that doesn't cause massive global catastrophe: x 2/3

Your own estimate of the good SIAI will do being biased upwards, given that they're the cause you've come to support: x 2/3

Even if the SIAI researchers do everything as efficiently and skillfully as humanly possible, they critically misguess key risks/benefits because so many of the numbers they use are unscientifically concocted: x 1/20 (the irony is partly intentional )

Greater risk of unknown unknowns getting in the way than more immediate causes: x 2/3 (this should prob be quite a bit higher, but I'm immodestly assuming that I've managed to think of most relevant factors)

Iff the Christian god who you suppose created us all and let us live this long in the first place exists, him planning to recreate or replace us even if we wipe ourselves out: x 1/5 (conditional on your belief in his existence)

(if some of these overlap, I apologise)

I might add more if you know the number I'm modifying (perhaps a more relevant one to you, given that it's you I'm discussing this with), and why.

Regarding your last point, Arepo:

Iff the Christian god who you suppose created us all and let us live this long in the first place exists, him planning to recreate or replace us even if we wipe ourselves out: x 1/5 (conditional on your belief in his existence)

surely if one of us took up a belief in christianity, we would still recognise the probability of his existence to be <1% and therefore make him irrelevant to this calculation. However, even if we disregard that last point, Arepo has made a strong case for adjusting your perception of SIAI.

What is more important to me about SIAI is not that you have the audacity to radically depart from behaviour that we intuit to be ethical, but that you do it in the name of utilitarianism! If you would just promote SIAI on a deontological basis instead, I would hugely prefer that!

RyanCarey wrote:Regarding your last point, Arepo:

surely if one of us took up a belief in christianity, we would still recognise the probability of his existence to be <1% and therefore make him irrelevant to this calculation.

I agree in principle, but this was intended specifically for Alan, and is a bit fiddly. As far as I remember Alan's Christianity (please let me know if I grossly misrepresent you here, A), he's accepted a form of Pascal's wager that says given any finitely low probability of Christian God (CG)'s existence, the expected utility from believing in him is infinite.

But 'believing' isn't just recognising that 0.00001 x infinity is infinity, it's actually having some reasonably high (>50% by the definition I expect most would use) credence that the proposition in question is true. So one of the common objections to Pascal's wager is that you can't make yourself believe something - IIRC Alan denies this, and despite having come from a position of epistemological scepticism, has said he has made himself believe in some sense.

I don't really understand what that means to him, but IMO to 'believe' something at least entails acting in a manner consistent with it being true. So if he were to act consistently with the proposition that God exists, he would give full (or at least strong) credence to the idea that CG is a factor when considering where to donate. Therefore not discounting for the possibility of CG's involvement is not compatible with believing in him, and thus not with Alan going to heaven.

What is more important to me about SIAI is not that you have the audacity to radically depart from behaviour that we intuit to be ethical, but that you do it in the name of utilitarianism! If you would just promote SIAI on a deontological basis instead, I would hugely prefer that!

I wouldn't go this far. I'd prefer he was doing the 'wrong' thing for the right reason, since a) it means he might eventually realise he should do otherwise and b) he's more likely than a deontologist to actually be doing the 'right' thing, and simply have some insight that we don't.

Thanks, Arepo, for giving some explicit out-of-thin-air multiplier numbers. We can all agree that they shouldn't be taken too seriously, but being concrete gives us more to talk about than trying to draw verbal distinctions with words like "small" and "most."

Your enumeration of ways things could go wrong was very thoughtful, and I agree with several of them. But even with a very small multiplier (say, 10^-15), I think that actions to affect the future of humanity still come out far ahead of things like Fred Hollows. The numbers of potential sentients affected by working on futuristic scenarios is just too huge.

Arepo wrote:But if you're donating money specifically to the manufacture of leaflets which require time and resources to distribute, you're marginally reducing the time they have to spend on other activities, so I don't think you should claim this as a side benefit.

Only if the money doesn't go toward hiring more total staff.

Arepo wrote:AIs will not be developed for much longer than we expect, meaning that the money would be better invested closer to the time: x 1/2

Given the internal rates of return of the activities that charities do (e.g., spreading their memes and gaining members), it's not clear to me that investing is usually the best choice. (Of course, curing blindness has internal rates of return of its own.) As far as when AI is developed, I think the right time to start serious work on the problem is hundreds / thousands of years ahead of time -- some of the thorny philosophical issues involved need that much reflection.

Arepo wrote:they don't have nearly enough real mathematicians from what I've seen of them

If SIAI is strong in anything at all, I'd say it's math skills. I agree they don't have well-established mathematics professors working for them full-time, but nearly everyone involved has an excellent mathematical background, and a large fraction of the researchers are top scorers on, e.g., the Putnam exam.

If anything, i sometimes worry that SIAI may be too math-heavy, to the exclusion of other ways of looking at the world.

Arepo wrote:I've seen extremely competent mathematicians dismiss as trivially false or fatallly vague

Well, I'd personally go with the philosophers over the mathematicians in that case. There are a lot of very smart people who don't think these problems are trivial. I guess we just differ in our outlooks here....

Arepo wrote:Major economic upheaval, caused eg by peak oil or climate change, devalues your currency before it's actually been used and prevents it from causing any or as much good as money which is spent immediately

As far as I can tell, much of the donation will be used right away, to pay for more researcher hours that would otherwise be spent at regular income-earning jobs. In general, I agree that economic collapse is an important factor to include in these calculations.

Arepo wrote:Your own estimate of the good SIAI will do being biased upwards, given that they're the cause you've come to support

FWIW, I started out extremely skeptical (and remain fairly skeptical, actually). That my position shifted in their favor might suggest using a multiplier >1.

RyanCarey wrote:What is more important to me about SIAI is not that you have the audacity to radically depart from behaviour that we intuit to be ethical

Well, that's what utilitarianism often is, no? At least until we re-adjust our intuitions in light of new data or thought experiments.

Of course, utilitarianism is also about healthy debate about philosophy and practical realities, so I appreciate all of your comments here.

By the way, I should add that I don't think reliance on intuition is a bad thing. Fundamentally, our foundational intuitions are all there is (and all there can be) to ethics.

Arepo wrote:But 'believing' isn't just recognising that 0.00001 x infinity is infinity, it's actually having some reasonably high (>50% by the definition I expect most would use) credence that the proposition in question is true.

I may have made suggestions to that effect at some point, but I now disagree. I think it's really dangerous to corrupt your probability distribution (at least, say, until you're near death) because the expected value of new information on questions like these is enormous. My probability of Christianity is extremely small (<< 1%).

Arepo wrote:I don't really understand what that means to him, but IMO to 'believe' something at least entails acting in a manner consistent with it being true. So if he were to act consistently with the proposition that God exists, he would give full (or at least strong) credence to the idea that CG is a factor when considering where to donate. Therefore not discounting for the possibility of CG's involvement is not compatible with believing in him, and thus not with Alan going to heaven.

That's a valid point, yes. "Acting as though Christianity is true" in its entirety would seem to require that. So I am risking hell here in order not to selfishly hurt others. (As I may have mentioned before, I personally don't care much about heaven.)

Alan Dawrst wrote:Your enumeration of ways things could go wrong was very thoughtful, and I agree with several of them. But even with a very small multiplier (say, 10^-15), I think that actions to affect the future of humanity still come out far ahead of things like Fred Hollows. The numbers of potential sentients affected by working on futuristic scenarios is just too huge.

This is obviously the key point, but I think it’s worth a separate topic. I was going to cross-post, but it’s not finished and I need to wind up. I’ll try and post it tomorrow.

Only if the money doesn't go toward hiring more total staff.

Well no, but this discussion is all about the value of money spent on flyers. If it’s a discussion about something else then I’ll reformulate the criticism after you’ve reformulated the argument

Arepo wrote:As far as when AI is developed, I think the right time to start serious work on the problem is hundreds / thousands of years ahead of time -- some of the thorny philosophical issues involved need that much reflection.

I don’t see how you can possibly make this claim with any confidence, least of all if it’s true.

Arepo wrote:If SIAI is strong in anything at all, I'd say it's math skills. I agree they don't have well-established mathematics professors working for them full-time, but nearly everyone involved has an excellent mathematical background, and a large fraction of the researchers are top scorers on, e.g., the Putnam exam.

I’m not familiar with the Putnam exam. As for ‘excellent mathematical background’, I must admit ignorance. What are we talking about? Anything less than postgraduate research isn’t really worth comparing against professional mathematicians who, in my experience, consider a maths undergrad relatively simple.

Arepo wrote:Well, I'd personally go with the philosophers over the mathematicians in that case.

That strikes me as seriously wrong. Philosophy - almost by definition - has no mechanism for recognising bad research or researchers. Mathematicians on the other hand are tested every step of the way. Tell a philosopher a high-level maths problem and 99 times out of 100 he'll get quickly stuck. Tell a mathematician a high-level philosophy problem and 100 times out of 100 he'll opine as confidently as the philosopher.

There are a lot of very smart people who don't think these problems are trivial.

While I think argument from authority is an undervalued approach, I don’t think it applies here. The problems with these arguments are relatively simple, and basically handwaved away by those who support them - it’s not like they give counterpoints to my objections that I can’t comprehend.

Also, intelligent people believe many things which I think are obviously implausible - ghosts, the sorites 'paradox', most continental philosophy etc. Those things don’t necessarily affect the SIAI’s mission in the way that bad reasoning about existential threats do, though.

As far as I can tell, much of the donation will be used right away, to pay for more researcher hours that would otherwise be spent at regular income-earning jobs. In general, I agree that economic collapse is an important factor to include in these calculations.

Ok.

FWIW, I started out extremely skeptical (and remain fairly skeptical, actually). That my position shifted in their favor might suggest using a multiplier >1.

I don’t think this is very relevant. People change their mind about things all the time. Still more often they believe they’ve changed their mind, sometimes exaggerating the extent to which they actually did. Perhaps there’s a slight tendency for changes of minds to be in the ‘right direction’, but I doubt it’s anywhere near as significant as confirmation bias.

By the way, I should add that I don't think reliance on intuition is a bad thing. Fundamentally, our foundational intuitions are all there is (and all there can be) to ethics.

I think most of us will agree that a) some sort of philosophical scepticism is appropriate and b) util isn’t a creed written into the universe, but I think philosophers heavily overuse and equivocate on the concept of ‘intuition’. For example, creating numbers based on your gut feeling of something’s probability is a completely alien activity to assuming some of the basic axioms necessary to functioning in day-to-day society. But philosophers often conflate the two (and I think you’re doing so here, to some degree).

My probability of Christianity is extremely small (<< 1%).

That’s interesting. Do you still consider yourself a Christian nonetheless? It seems incongruous with the bible/mainstream Christianity to suppose that Yahweh will let people in to heaven if they think him so improbable...

Arepo wrote:Well no, but this discussion is all about the value of money spent on flyers.

Ah yes, good point. I was thinking about donations to Vegan Outreach as a whole (which I used for the lower-bound cost-effectiveness estimate), but you're right that the "How Much Is a Dollar Worth?" piece focuses on fliers specifically.

Arepo wrote:What are we talking about?

I was talking more about raw mathematical talent and insight, which the Putnam exam measures. But I agree that familiarity with lots of relevant background literature is also very important -- perhaps more so, since the smartest genius can't reproduce all of the insights of the past by her own sheer brain power.

Arepo wrote:The problems with these arguments are relatively simple, and basically handwaved away by those who support them - it’s not like they give counterpoints to my objections that I can’t comprehend.

Agreed. I haven't been specific about how I think the problems are tricky. So, for instance, on Newcomb, do you one-box or two-box? Why?

Arepo wrote:Perhaps there’s a slight tendency for changes of minds to be in the ‘right direction’, but I doubt it’s anywhere near as significant as confirmation bias.

So the multiplier should be a (perhaps very small) amount greater than 1, right?

Alan Dawrst wrote:I was talking more about raw mathematical talent and insight, which the Putnam exam measures. But I agree that familiarity with lots of relevant background literature is also very important -- perhaps more so, since the smartest genius can't reproduce all of the insights of the past by her own sheer brain power.

Ok, but see the above point about maths being a constantly testing process. Unlike philosophy you can't become successful just by reading literature and finding a small group of people who like the sound of your interpretation. You also have to understand and apply it. Net result is that people employed as as mathematicians, especially by prestigious institutions tend to be (provably) extremely good at what they do. You say the Putnam exam 'measures' talent, but what is it benchmarked against?

Agreed. I haven't been specific about how I think the problems are tricky. So, for instance, on Newcomb, do you one-box or two-box? Why?

New thread? I don't want to derail this one. I doubt we'll have much to discuss, though - I don't either box until the question is defined better. Those who think it a paradox just seem to assert that it's already adequately defined.

Arepo wrote:Perhaps there’s a slight tendency for changes of minds to be in the ‘right direction’, but I doubt it’s anywhere near as significant as confirmation bias.

So the multiplier should be a (perhaps very small) amount greater than 1, right?

Not in my view. Once someone's reached a conclusion, I believe cognitive biases tend to lead them to overjudge its probability (I can't source this, it's just my speculation - I vaguely remember Toby alluding to studies that purported to measure the extent of some biases, but it doesn't seem worth tracking down for this). This would heavily outweigh the very slight tendency I also suspect we have to change our minds in the right direction.

Arepo wrote:You say the Putnam exam 'measures' talent, but what is it benchmarked against?

I don't claim to know much more about the exam than, say, Wikipedia does, so I'll defer that question to it.

Arepo wrote:Not in my view. Once someone's reached a conclusion, I believe cognitive biases tend to lead them to overjudge its probability (I can't source this, it's just my speculation - I vaguely remember Toby alluding to studies that purported to measure the extent of some biases, but it doesn't seem worth tracking down for this). This would heavily outweigh the very slight tendency I also suspect we have to change our minds in the right direction.

I'm puzzled. Cognitive biases suggest that the amount by which the factor exceeds 1 should be small, but surely it shouldn't be < 1 as compensation?

Suppose I was at the scene of a crime, and as part of the sequence of events, I thought I saw the murderer flip a coin that came up heads. I repeated my testimony of the murder many times to criminal investigators, and each time my tale became more elaborate and fanciful. It's two years later and I'm telling my story to you. I say I seem to recall that the coin came up heads. Conditional on there having been a coin flip, do you -- on account of my elaborations and fabrications -- assign probability less than 50% to the coin having come up heads?

Alan Dawrst wrote:I'm puzzled. Cognitive biases suggest that the amount by which the factor exceeds 1 should be small, but surely it shouldn't be < 1 as compensation?

I was puzzled too, but I think we're having each other's conversation. I'm discussing how much you personally should adjust the (initial) expected value of your contribution to SIAI. Given that cognitive biases generally lead people to overestimate the likelihood of their views, it follows that whatever you think the expected value of what you think is the best value contribution is, it's slightly lower.

From my perspective though, the fact that you've come to support the group should raise my estimation of contributions to SIAI.

Arepo wrote:I was puzzled too, but I think we're having each other's conversation.

Yes, that's exactly it. I agree with everything you said subsequently on this point.

LOVE this thread, should've read it a while ago.

Alan Dawrst wrote:On the other hand, this option is wanting in precisely the domain where Vegan Outreach is strongest: Changing society's long-term attitudes toward animal suffering. While the impact here is still probably positive (for instance, because changed laws or corporate policies will tend to normalize the notion that animals deserve humane treatment, similarly to the way in which anti-discrimination laws have helped to normalize concern for people of all races), I would guess it's considerably smaller than the direct, person-to-person influence that leafleting has.

I would intuitively think it's the other way around. It's very hard to change people's minds. Growing up with a law that says "x is wrong" is much more influential than having a conversation with a leafleter who is trying to convince you that "x is wrong".

Another point on humane slaughter (or, even better, higher welfare in general) vs. Vegan Outreach that I don't think you mentioned when comparing the two - encouraging higher welfare standards is in line with the utilitarian type of thinking that also supports doing something about wild animal suffering. Trying to create vegans sometimes has to involve arguments about non-intervention and "animal rights".

What about donating to BLTC Research?

I have at this point already made a commitment regarding SIAI.

What are your current thoughts on where's best to donate? I'm not sure how up-to-date everything is that I've read by you.
edit: used quotation marks to fix the code for Alan Dawrst's quote ~Ryan

P.S. How do you make it so that it says "Alan Dawrst wrote" at the start of quotes? :p

Actually, two more things:

What is more important to me about SIAI is not that you have the audacity to radically depart from behaviour that we intuit to be ethical, but that you do it in the name of utilitarianism! If you would just promote SIAI on a deontological basis instead, I would hugely prefer that!

RyanCarey, I'm surprised! Supporting SIAI strikes me, and many utilitarians I know, as a very utilitarian thing to do (though I'll admit initially it all sounded a bit ridiculous, as did most Very Utilitarian Things To Do, as Alan Dawrst pointed out). Has your mind changed since you made this post?

By the way, I should add that I don't think reliance on intuition is a bad thing. Fundamentally, our foundational intuitions are all there is (and all there can be) to ethics.

"all there is" links to the Wikipedia entry on Emotivism...are you an emotivist, Alan Dawrst?

LadyMorgana wrote:Trying to create vegans sometimes has to involve arguments about non-intervention and "animal rights".

Hi LadyMorgana. It's true that a lot of vegans defend non-intervention and harcore rightism. But non-intervention and rightism isn't innate in veganism. I know several vegans who reject rightism and defend interventions in the wild.

It's possible that a animal charity doesn't promote rightism nor non-intervention. Pieces and websites can focus in suffering and speciesism. I guess this would be easier if more utilitarians get involved in animal charities.

LadyMorgana wrote:I would intuitively think it's the other way around. It's very hard to change people's minds. Growing up with a law that says "x is wrong" is much more influential than having a conversation with a leafleter who is trying to convince you that "x is wrong".

Even if so, leafletting is an excellent way to increase the numbers of people who care about the issue and so can eventually lobby for such laws. Even campaigns that are explicitly political spend much time raising popular support.

LadyMorgana wrote:What about donating to BLTC Research?

Is that an active charity? Dave Pearce can comment here.

LadyMorgana wrote:"all there is" links to the Wikipedia entry on Emotivism...are you an emotivist, Alan Dawrst?

I am. However, I don't think it should diminish our enthusiasm for reducing suffering -- indeed, emotivism propounds that we care about our values intensely. If it were any other way -- if values were somehow "absolute" -- then what motivation would we have to care about them?

What about donating to BLTC Research?

I remember reading somewhere that it wasn't accepting small donations, only very large ones from millionaires or something. I'm also concerned with how utilitarian it would be, it's run by utilitarians but that's no guarantee. Oh god, I just rhymed...

I also have just discovered this thread and think it good.

And I second the request for any opinion changes people may've had on this topic. Also, Alan, are you still giving $15,000/yr or are you following the save/invest strategy and why are you following whatever strategy you are following?

Gedusa wrote:Alan, are you still giving $15,000/yr or are you following the save/invest strategy and why are you following whatever strategy you are following?

I still donate ~$15K per year because my employer matches donations up to that amount. The rest I save/invest in my bank account. (Actually, I invest it in stocks, not a bank account, but "stocks" didn't rhyme. Bank accounts have lower expected returns.)

I currently give my yearly donations to Vegan Outreach. The reason for saving the rest is that I may later discover a better charity or, more likely, I may want to use the money for things that aren't funded by any existing charity -- e.g., (a) paying students to research wild-animal welfare, impacts of environmental changes on animal populations, or whether insects can suffer, (b) funding writers/bloggers to popularize existing research on animal welfare and emotions, or (c) starting an organization to promote awareness of wild-animal suffering.

Most of these things could, it seems, be done quite cheaply. When I was in college, students were paid $8 an hour for campus jobs, and this was at one of the best colleges in the country. (Upon graduation, those same students could be earning $30+ per hour.) High-school students are really smart as well and might need even less pay.

And actually, there are lots of people who would do this work for free. (Just look at all of us participating on this forum!) Much of what's required is merely to find interested people and coordinate the details. This could be as simple as suggesting research for bloggers to write about or topics for a grad student in an environmental ethics program to study for her next paper. Thus, some of the savings will just go toward my costs of living so that I can do these other things full-time at some point.

Stocks, as in index funds, or some other investment strategy?

Alan Dawrst wrote:A number of people have suggested New Harvest, and I agree that it's a solid organization. My main concern is that, as I suggested in the main post, I fear vegetarianism could -- depending on the balance of suffering in nature and the animal-population impacts of climate change -- cause more animal suffering than it averts.

Related: http://www.sciencedaily.com/releases/20 ... 101036.htm

Science Daily wrote:'What our study found was that the environmental impacts of cultured meat could be substantially lower than those of meat produced in the conventional way,' said Hanna Tuomisto of Oxford University's Wildlife Conservation Research Unit, who led the research. [...] They also suggest that land freed up from farming could be reforested or used for other carbon sequestration purposes, further lowering the carbon footprint of cultured meat.

Jesper Östman wrote:Stocks, as in index funds, or some other investment strategy?

I buy individual stocks rather than mutual funds for convenience reasons, but philosophically, I think index funds are the way to go. My stock-selection method is essentially to throw darts at a stock chart, although I do prefer stocks with a higher "beta" just in case the CAPM has any validity (see the "Summary" here for links).

[I wrote this before I realized that I was on the first page and reading an old post]

Arepo wrote:Not in my view. Once someone's reached a conclusion, I believe cognitive biases tend to lead them to overjudge its probability

Studies of gambling lead to a conclusion something like this. There is a research paradigm called the "Iowa gambling task" in which people play a game in which they try to win the most money. Of course the cards are rigged to see how people react in different circumstances. Gamblers tend to form a belief about the "quality" of the deck and will continue playing even when the deck is really terrible -- seeming to distort the probability of winning. There is also one called the "Bangor gambling task" (not studies as much as the other one). Another study shows the range of irrational thinking about probability that is observed in gamblers ("The role of cognitive bias and skill in fruit machine gambling" http://scholar.google.com/scholar?clust ... s_sdt=0,29)

If the claim is about making inferences from fact, then different studies cover it. One example is by Peng and Nisbett in which Americans increase their confidence in a prior answer after receiving new evidence that contradicts it while Chinese people express greater uncertainty (http://scholar.google.com/scholar?clust ... s_sdt=0,29). This reflects cultural differences in reasoning.

As for courtroom testimony, there are frequent cases of people being sure of what they saw only to reveal inaccuracies or impossibilities. You can google "testimony accuracy" and find studies on this. There is also a humorous video that shows how people can overlook details even when they are strongly focussed on observing a situation. There are many replications of this study (don't spoil it by revealing the secret!!): http://www.youtube.com/watch?v=7ncJ3Uy13u8&NR
here is another version closer to the original: http://www.youtube.com/watch?v=vJG698U2Mvo
Ryan's edit: fixed the quote code

rehoot wrote:Gamblers tend to form a belief about the "quality" of the deck and will continue playing even when the deck is really terrible

It's interesting to think how these sorts of responses actually make sense within most evolutionary landscapes. In nature, it's usually the case that something which displays high quality initially really has higher quality overall. Situations with fixed random odds (cards, dice, etc.) are pretty rare on the savanna.

rehoot wrote:If the claim is about making inferences from fact, then different studies cover it. One example is by Peng and Nisbett in which Americans increase their confidence in a prior answer after receiving new evidence that contradicts it while Chinese people express greater uncertainty (http://scholar.google.com/scholar?clust ... s_sdt=0,29). This reflects cultural differences in reasoning.

It's an interesting 'culture' that gets more confident of an already adopted answer when receiving new evidence that contradicts it. Did they explain why this happens? (No, I can't be bothered to read the article right now )

tog wrote:Did they explain why this happens?

The difference between the American and Chinese reaction to contradictory evidence suggested a cultural difference, and the difference that the authors suggested was the approach to reasoning. Americans use (what they think is) logic that results in a right or wrong answer and Chinese are more likely to use dialectical reasoning that can allow for contradiction and they are taught to seek a "middle way". If there is a conflict, the right-or-wrong approach means that somebody must be wrong, so of course we (Americans) assume that we are right and the other guy is wrong (as in the political debates about debt and everything else).

I'm not sure if the article mentioned this directly, but Asian cultures are also more focused on cooperation whereas Americans are focused on competition--there is probably some relationship between being competitive and trampling on another person's belief as opposed to being cooperative and allowing for contradiction thereby and thereby closer relationships.

Not sure if this is the right thread, but can you help me out with a bit of logic? I'm thinking about supporting New Harvest in order to make nearer-term in-vitro meat more probable. The terminal goal is to prevent the suffering of non-human animals in factory farming regimes.

However, as a side-effect, it is possible that resource efficiency of meat production is increased, and climate impact is decreased. This would lead to an increase in general ecological sustainability and therefore - indirectly - a reduction in existential risk.

Most people would think that this is a good thing, but I'm not convinced that the existence of sentient life in general is more good than bad in a hedonistic utilitarian sense, and I'm not convinced that this balance will improve in the future with sufficient probability. If it is true that there will be more suffering than pleasure in the future, and it is true that existential risk of the type created by meat production unsustainability can reduce the expected suffering surplus, and it is true that in-vitro meat would reduce existential risk of this type, then it may be wrong to support in-vitro meat.

I want to figure this out before I put more money into it. What are your thoughts?

Hedonic Trader wrote: If it is true that there will be more suffering than pleasure in the future, and it is true that existential risk of the type created by meat production unsustainability can reduce the expected suffering surplus, and it is true that in-vitro meat would reduce existential risk of this type, then it may be wrong to support in-vitro meat.

My short answer is that I have nothing close to a precise answer. I would estimate that the higher-confidence estimates of short-term impacts are beneficial and that the prospect of negative impacts in the future are less certain. Based on that, I would say that the in-vitro project looks good, but I know nothing about possible side-effects like mutations or new diseases that occur as part of the manipulation of cells.

Some things to consider:

a = total harm from meat/fish/poultry industry today
b = expected harm from meat industry after the in-vitro project
c = direct harm-reduction from in-vitro project
d = interaction effects (would reducing actual meat consumption remind people
of other values and have positive effects? Every spoonful is a reminder that
"we don't kill animals" -- would there be conflicts that arise from this?)
e = net value (or disvalue) of increased number of people that can exist
because of the greater efficiencies of the in-vitro project

components of e to consider relative to long-term negative impacts:
* How long before the actual number of humans increases beyond
what would be otherwise impossible?
* Will the effects of the increased number of people (if it ever happens)
be offset by a change in lifestyle that was supported by the in-vitro project?
* How long is "the long-term?" = When will human civilization be destroyed
(giant comet, biological warfare, the sun explodes, war followed by global
anarchy and loss of educational systems and reversion to more primitive
society—all highly uncertain)? I have no idea on this point.

Hi rehoot, thanks for your ideas.

rehoot wrote:My short answer is that I have nothing close to a precise answer. I would estimate that the higher-confidence estimates of short-term impacts are beneficial and that the prospect of negative impacts in the future are less certain.

Yes, I agree. I'm mostly concerned about them because of the large quantities of sentience that can come into existence in the future, e.g. through space-colonization etc.

Based on that, I would say that the in-vitro project looks good, but I know nothing about possible side-effects like mutations or new diseases that occur as part of the manipulation of cells.

Interesting. This didn't really occur to me. I would think that disease risks from cultured meat shouldn't be very much higher than risks from livestock - some pathogens find hosts in intensely farmed animals, and dangerous mutation potential from the human-lifestock interactions (bird flu etc.). I would also assume that society can react relatively quickly to most other health effects if they occur.

Some things to consider:

a = total harm from meat/fish/poultry industry today
b = expected harm from meat industry after the in-vitro project
c = direct harm-reduction from in-vitro project
d = interaction effects (would reducing actual meat consumption remind people
of other values and have positive effects? Every spoonful is a reminder that
"we don't kill animals" -- would there be conflicts that arise from this?)
e = net value (or disvalue) of increased number of people that can exist
because of the greater efficiencies of the in-vitro project

components of e to consider relative to long-term negative impacts:
* How long before the actual number of humans increases beyond
what would be otherwise impossible?
* Will the effects of the increased number of people (if it ever happens)
be offset by a change in lifestyle that was supported by the in-vitro project?
* How long is "the long-term?" = When will human civilization be destroyed
(giant comet, biological warfare, the sun explodes, war followed by global
anarchy and loss of educational systems and reversion to more primitive
society—all highly uncertain)? I have no idea on this point.

Right. e seems to have massive potential. I think the most important factor in e is the probability that sustainability tipping points are affected, with existential risks as dependent outcomes. Examples could be self-reinforcing climate change that cripples the resource base for advanced human civilization, or degradation of topsoil leading to more resource pressures leading to increased war risks leading to higher probability of nuclear winter crippling human civilization. The probability shifts associated with individual sustainability technologies like cultured meat may be small, but the utility shifts are so massive that they would still matter.

E.g., space-colonizing posthuman entities could create massive amounts of sentience, possibly locked in an ever-lasting quasi-darwinian resource struggle on an interstellar scale. But you're right, these outcomes are far less certain. There are many "and ifs" involved here; conjunction fallacy comes to mind.

Hedonic Treader wrote:I'm thinking about supporting New Harvest in order to make nearer-term in-vitro meat more probable. [...] I want to figure this out before I put more money into it. What are your thoughts?

Why not donate to Vegan Outreach instead? It has a similar risk-reward profile as New Harvest in terms of reducing demand for greenhouse-gas-emitting animals, but it also has the (IMO, more important) side effect of increasing concern for animal suffering, which may ultimately translate into awareness of the plight of animals in nature.

(Plus, the more people Vegan Outreach influences, the greater the support for in vitro meat. )

Alan Dawrst wrote:Why not donate to Vegan Outreach instead? It has a similar risk-reward profile as New Harvest in terms of reducing demand for greenhouse-gas-emitting animals, but it also has the (IMO, more important) side effect of increasing concern for animal suffering, which may ultimately translate into awareness of the plight of animals in nature.

Hi Alan, that seems like a very indirect - and therefore statistically small? - side-effect. Generally speaking, I don't like intervention plans that center around convincing people to do something costly that they would not otherwise do (like going vegan).

I'm really pessimistic about most people's ability and willingness to change their minds based on moral arguments alone. A fraction, yes, but majorities? If it doesn't coincide with their selfish incentives? I'm skeptical.

I think there's no more efficient way to increase the readiness to reduce suffering than by supporting veganism. That's because people are involuntarily signaling by not going vegan that they don't care about suffering of others deeply enough to change their behavior so that they refrain from actions that do massive harm while only gaining a moderate advantage for themselves.

If people aren't ready to go vegan, I think it is doubtful that they will care about suffering of wild animals, sentient artificial intelligences, and suffering in lab universes. After all, AIs and lab universes may be way more exciting than eating animal products.

Mike Radivis wrote:If people aren't ready to go vegan, I think it is doubtful that they will care about suffering of wild animals, sentient artificial intelligences, and suffering in lab universes. After all, AIs and lab universes may be way more exciting than eating animal products.

I think you may underestimate the difficulty of the shift in culinary experience and lifestyle that the seemingly simple act of "going vegan" implies for many people. On a daily basis, AIs and lab universes may make for exciting stories, but eating diverse food in habitual ways is probably more important to people. It pays to make that switch easier, e.g. through access to good and easy alternatives. I stopped eating meat only when I started eating at a canteen that routinely offered very good vegetarian meals and I saw footage of animal abuse in factory farming. The vague awareness had been there before, but it didn't result in a switch of habitual consumption.

Furthermore, the problem of wild animal suffering, and potential solutions, have to be addressed directly. They're not implicit in "going vegan" as a lifestyle. I know many vegetarians and vegans who would gladly bring the jungle to other planets if they could, including predation, parasitism, starvation etc. I have found that addressing this problem directly and non-judgmentally (as a simple problem statement, without jumping to suggestions for interventions) in discussions about nature, animal welfare, or environmentalism raises awareness better. On the net, you can also link to Alan's resources or the associated essays on hedweb or Robin Wilbin's site. Many reactions show surprise - people often express they never thought about the problem that way. This seems to be a good start for at least some memetic shift.

Hedonic Treader wrote:I think you may underestimate the difficulty of the shift in culinary experience and lifestyle that the seemingly simple act of "going vegan" implies for many people. On a daily basis, AIs and lab universes may make for exciting stories, but eating diverse food in habitual ways is probably more important to people.

Yes, it's very difficult for some people to go vegan, but I think it's definitely possible. The difficulty stems from being in a culture that is totally accustomed to abusing animals for personal gratification. Imagine our future culture would be accustomed to torturing, abusing, and experimenting with sentient beings in virtual worlds or lab universes. It would be a nightmare scenario that is almost impossible to escape. Such a civilization might be addicted to the core. That's just an extra reason why it's so important to increase awareness and concern for suffering as soon as possible. Nightmare scenarios might just be too attractive, so better not even get close to them!

Hedonic Treader wrote:It pays to make that switch easier, e.g. through access to good and easy alternatives. I stopped eating meat only when I started eating at a canteen that routinely offered very good vegetarian meals and I saw footage of animal abuse in factory farming. The vague awareness had been there before, but it didn't result in a switch of habitual consumption.

My own development was very similar. And yes, it is good to make the switch to veg*ism easier. The contemporary addiction to products of non-human animals is a factor that is very adverse to rationality. And applying utilitarianism the right way requires a high level of rationality.

Hedonic Treader wrote:Furthermore, the problem of wild animal suffering, and potential solutions, have to be addressed directly. They're not implicit in "going vegan" as a lifestyle. I know many vegetarians and vegans who would gladly bring the jungle to other planets if they could, including predation, parasitism, starvation etc. I have found that addressing this problem directly and non-judgmentally (as a simple problem statement, without jumping to suggestions for interventions) in discussions about nature, animal welfare, or environmentalism raises awareness better.

Yes, I absolutely agree. I was just arguing on the basis of already existing charities. What actually might be most effective is an organization that provided all kinds of information and education/edutainment about utilitarianism as well as special related topics like (wild) animal suffering or artificial sentient beings. Better start with the latter topic as soon as possible - it might soon become a real issue, see the post Grandoids! Artificial Life?

Hedonic Treader wrote:Many reactions show surprise - people often express they never thought about the problem that way. This seems to be a good start for at least some memetic shift.

Mike Radivis wrote:Imagine our future culture would be accustomed to torturing, abusing, and experimenting with sentient beings in virtual worlds or lab universes. It would be a nightmare scenario that is almost impossible to escape.

That's one of the reasons I'm wary of humanity's survival.

Hedonic Treader wrote:Furthermore, the problem of wild animal suffering, and potential solutions, have to be addressed directly. They're not implicit in "going vegan" as a lifestyle. I know many vegetarians and vegans who would gladly bring the jungle to other planets if they could, including predation, parasitism, starvation etc. I have found that addressing this problem directly and non-judgmentally (as a simple problem statement, without jumping to suggestions for interventions) in discussions about nature, animal welfare, or environmentalism raises awareness better. On the net, you can also link to Alan's resources or the associated essays on hedweb or Robin Wilbin's site. Many reactions show surprise - people often express they never thought about the problem that way. This seems to be a good start for at least some memetic shift.

Hi HT. You have a good point here. Veganism is usually linked to an environmental and rightist message.

But I think it happens something different when anti-speciesism (and not only veganism) is promoted. Vegan charities in America, United Kingdom and Australia promote... well, just veganism. But several Spanish and South-American vegan charities have a more focused anti-speciesist message. I know a lot of vegan activists in Spain and South America, including several directives, who come to agree that wild animal suffering matters.

So I agree: "the problem of wild animal suffering, and potential solutions, have to be addressed directly." But I think this would be easier in a less speciesist society. In any case there is already an animal rights movement. IMO it would be very important to direct it to a more rational and anti-speciesist approach. If this doesn't happen, to solve the wild animal suffering issue could be even harder.

Alan Dawrst wrote:That's one of the reasons I'm wary of humanity's survival.

Yes. Humanity's existence is quite probably a contributing factor to massive increases in net harm. Alas, as long as no thoughtful one has a genetically engineered supervirus, or something similar, at home, there's not terribly much that can be directly done about that.

Daniel Dorado wrote:several Spanish and South-American vegan charities have a more focused anti-speciesist message. I know a lot of vegan activists in Spain and South America, including several directives, who come to agree that wild animal suffering matters.

! links please:D!

Ruairi wrote:

Daniel Dorado wrote:several Spanish and South-American vegan charities have a more focused anti-speciesist message. I know a lot of vegan activists in Spain and South America, including several directives, who come to agree that wild animal suffering matters.

! links please:D!

http://www.animalequality.net/ (English language)
http://www.equanimal.org/ (Spanish language)
http://www.especismocero.org/ (Spanish language)