Brian_Tomasik

Why Char­i­ties Usu­ally Don’t Differ Astro­nom­i­cally in Ex­pected Cost-Effectiveness

[Question] How much fung­ing is there with dona­tions to differ­ent EA an­i­mal char­i­ties?

• Brian_Tomasik 20 Jan 2023 21:16 UTC

  79 points

  25 ∶ 13

  on: Thread for dis­cussing Bostrom’s email and apology

  Bostrom was essentially still a kid (age ~23) when he wrote the 1996 email. What effect does it have on kids’ psychology to think that any dumb thing they’ve ever said online can and will be used against them in the court of public opinion for the rest of their lives? Given that Bostrom wasn’t currently spreading racist views or trying to harm minorities, it’s not as though it was important to stop him from doing ongoing harm. So the main justification for socially punishing him would be to create a chilling effect against people daring to spout off flippantly worded opinions going forward. There are some benefits to intimidating people away from saying dumb things, but there are also serious costs, which I think are probably underestimated by those expressing strong outrage.

  Of course, there are also potentially huge costs to flippant and crass discussion of minorities. My point is that the stakes are high in both directions, and it’s very non-obvious where the right balance to strike is. Personally I suspect the pendulum is quite a bit too far in the direction of trying to ruin people’s lives for idiotic stuff they said as kids, but other smart people seem to disagree.

  As some others have noted, probably the best approach we can take to the question of a genetic racial IQ gap is to voluntarily avoid mentioning the idea or at least to portray the question as relatively uninteresting (as Bostrom said in his apology), especially since the policy implications of such a gap wouldn’t be that big anyway. However, trying to suppress inquiries by others into the topic seems to me like it may cause more harm than good, via the Streisand effect. If it looks like there’s secret knowledge that the scientific community has informally conspired to suppress, the topic suddenly becomes way more interesting. There was also a genuine conspiracy (both in terms of meetings among scientists and tacit peer pressure) to suppress discussion of the lab-leak theory of COVID-19 in 2020, and that fact made the question much juicier.

  I admire that Bostrom’s apology didn’t take the easy way out by lying to pretend he thinks there’s no possibility of a genetic IQ gap. I see that as a positive sign about his intellectual honesty. (Actually, it’s obvious that between any two groups of humans, there will be some difference in the distribution of genes and therefore some nonzero genetically caused difference in the average of almost any trait you wish to measure. The real question is whether those differences are big or negligible. And even if the differences are non-negligible, I suspect it’s usually bad to highlight them anyway, such as due to stereotype threat and the possibility of encouraging violence toward and persecution of outgroups. In medical contexts, such as when people with African ancestry are genetically more at risk from certain diseases than other groups, highlighting these differences seems useful.)

• Brian_Tomasik 17 Feb 2023 4:21 UTC

  76 points

  16 ∶ 1

  on: Why I No Longer Pri­ori­tize Wild An­i­mal Welfare (ed­ited)

  Great post!

  As comments by Max and Vasco hint at, I think it might still be the case that considering effects on wild animals is essential when evaluating any short-termist intervention (including those for farmed animals and human welfare). For example, I remain uncertain whether vegetarianism increases or decreases total suffering because of wild-animal side effects, mainly because beef may (or may not!) reduce a lot of suffering even if all other meat types increase it. (I still hope people avoid eating chicken and other small farmed animals.)

  In my opinion, the most important type of WAW research is getting more clarity on big questions, like what the net impact is of cattle grazing, climate change, and crop cultivation on total invertebrate populations. These are some of the biggest impacts that humanity has on wild animals, and the answers would inform analysis of the side effects of various other interventions like meat reduction or family planning.

  I haven’t followed a lot of the recent WAW work, but my experience is that many other people working on WAW are less focused on these questions about how humans change total population sizes. Researchers more often think about ways to improve welfare while keeping population size constant. Those latter interventions may have more public support and are more accommodating to non-suffering-focused utilitarians who don’t want to reduce the amount of happiness in nature. But as you mention, those interventions also seem more subject to cluelessness (is vaccination net good or bad considering side effects? it’s super unclear) and often target big animals rather than invertebrates. From this perspective, I think efforts to improve the welfare of farmed chickens and fish may be more cost-effective (though it’s still worth exploring wild-animal interventions too, as you say). Research and advocacy of less painful killing of wild-caught fish also seems extremely important, and it’s unclear whether to count this as a farmed-animal or wild-animal intervention.

  Apart from less painful killing of wild animals (fish, rodents, insect on crop fields), or maybe some other large-scale interventions like reducing aquatic noise, I think the cost-effectiveness of work on wild animals would come from trying to reduce (or avoid increasing) population sizes, via reducing plant productivity. Reducing the amount of plant growth in an ecosystem helps invertebrates (including mites, springtails, and nematodes, which are extremely numerous but also hard to help in ways other than preventing their existence) and is somewhat less subject to cluelessness problems because you don’t have to model internal ecosystem dynamics as much—you just have to reduce the productivity of the first trophic level. But I haven’t found a lot of people who are interested in working on population-reduction interventions. This apparent lack of interest in reducing populations is one reason I’ve done less thinking about WAW in recent years. Another reason is that I respect the efforts of WAW organizations to put a more mainstream face on the WAW movement, and I wonder if my continuing to harp on why we should actually be focusing on reducing populations would seem to them as counterproductive.

  I compiled a list of possible interventions to reduce total invertebrate populations that could possibly be lobbied for at the government level in some fashion. Some of them don’t seem super cost-effective, but some might be, such as trying to reduce irrigation subsidies, which is an intervention that could be argued for on other grounds as well. Taxing fertilizer and/​or water use on crop fields, pastures, and/​or lawns might be pretty valuable if it could be achieved. (Some local regions do have subsidies for people who reduce their lawn’s water use.) If geoengineering to fertilize oceans ever happens, opposing it would be extremely important, though doing a campaign about that now might be net harmful via increasing the salience of the idea.

• Brian_Tomasik 22 Jan 2023 22:43 UTC

  61 points

  10 ∶ 0

  in reply to: Brian_Tomasik’s comment on: Thread for dis­cussing Bostrom’s email and apology

  Some notes on the last paragraph in my above comment:

  When I used the phrase “SJWs”, I intended it to have either neutral valence or a valence of friendly teasing. I agree with some amount of the SJW agenda myself. However, Wikipedia says that since 2011, the term is primarily used as an insult and is associated with the alt-right, which was not an implication I had in mind. Like Bostrom’s 1996 email and 2023 apology, this example is an illustration that it can be difficult to realize exactly how a given word or statement may be perceived, especially if people are reading it as if it were a dog whistle.

  Part of my reason for using the term “SJW” was that I didn’t want to say merely “leftist” or “progressive”. I was a strong leftist and progressive in the early aughts, and back then, people with that ideology were, in my experience, generally more focused on trying to improve people’s welfare via economic and other government-level policy. Progressives didn’t spend as much time as they do now on shaming individual people or groups. I think the woke-ward shift of the last decade, while it raises some important issues that were less highlighted in the past, is plausibly overall less useful for improving total human welfare than the earlier economic and policy focus was. So I don’t like conflating “woke” with “progressive”. (That said, I think some progressive economic policy positions, such as against outsourcing American jobs to developing nations, may be net bad for short-term human welfare.)

  A more neutral phrasing than “SJW” could have been just “social-justice activist”.

  As far as the other part of my phrasing, when I said that most social-justice activists “think it’s fine to enslave and murder” non-human animals, I was in part being deliberately provocative to make a point. Bostrom is right when he says in his apology: “I do think that provocative communication styles have a place”. If I had instead written that most social-justice activists “think it’s acceptable for farmers to raise and slaughter livestock”, the use of those conventional euphemisms would have dulled what I was trying to convey, which is that this practice is actually really awful. (BTW, I should also acknowledge that I myself pay for some amount of enslavement and murder of dairy cows, via eating cheese and ice cream. However, I think the total amount of harm this causes is much lower than the harm caused by eating meat from smaller farm animals.)

  Provocation can shock people out of their normal way of seeing the world into looking at some fact in a different light. This seems to be roughly what Bostrom was saying in the first paragraph of his 1996 email. However, in the case of that email, it’s unclear what socially valuable fact he was trying to shock people into seeing in a new way.

  One function of comedy is to do roughly the same thing: stating some true fact in an unconventional way in order to make people see the world through a new lens. However, an important principle in comedy is the distinction between “punching up” and “punching down”, and if we interpret Bostrom’s 1996 statement as analogous to provocative humor, it would clearly be punching down.

  It’s fairly common and even celebrated in modern Western society to hear statements like “women are more productive than men” or “girls are smarter at language than boys”. Many of these statements are made in fairly blunt language, similar to Bostrom’s 1996 statement. I assume most people think these statements about female superiority are pretty harmless, both because they’re seen more as “punching up” (given the history of men dominating women in much of the world until the late 20th century) and because the hypothesis of biological gender differences is less taboo and more scientifically established. But I do think the contrast in people’s reactions between saying “boys are worse at language than girls” versus Bostrom’s 1996 statement is interesting, and it shows that the degree of outrage a statement provokes is often not obvious unless you have a lot of experience with a specific culture’s norms.

  I do worry a bit that the casual misandry that society often seems to celebrate may be detrimental to the self-esteem of boys, though I’m also not interested in trying to police such language. It’s plausible to me that some amount of humorous mocking between different groups is actually helpful, by showing people that we can laugh together, rather than priming ourselves to interpret any offensive statement as an act of aggression.

• Brian_Tomasik 12 Jan 2022 8:44 UTC

  45 points

  0 ∶ 0

  on: Why Char­i­ties Usu­ally Don’t Differ Astro­nom­i­cally in Ex­pected Cost-Effectiveness

  Eight years later, I still think this post is basically correct. My argument is more plausible the more one expects a lot of parts of society to play a role in shaping how the future unfolds. If one believes that a small group of people (who can be identified in advance and who aren’t already extremely well known) will have dramatically more influence over the future than most other parts of the world, then we might expect somewhat larger differences in cost-effectiveness.

  One thing people sometimes forget about my point is that I’m not making any claims about the sign of impacts. I expect that in many cases, random charities have net negative impacts due to various side effects of their activities. The argument doesn’t say that random charities are within ¹⁄₁₀₀ times as good as the best charities but rather that all charities have a lot of messy side effects on the world, and when those side effects are counted, it’s unlikely that the total impact that one charity has on the world (for good or ill) will be many orders of magnitude higher than the total impact that another charity has on the world (for good or ill).

  I think maybe the main value of this post is to help people keep in mind how complex the effects of our actions are and that many different people are doing work that’s somewhat relevant to what we care about (for good or ill). I think it’s a common trend for young people to feel like they’re doing something special, and that they have new, exciting answers that the older generations didn’t. Then as we mature and learn about more different parts of the world, we tend to become less convinced that we’re correct or that the work we’re doing is unique. On the other hand, it can still be the case that our work may be a lot more important (relative to our values) than most other things people are doing.

  What links here?

  • Re­sults from the First Decade Review by Lizka (13 May 2022 15:01 UTC; 163 points)
  • What con­crete ques­tion would you like to see de­bated? I might or­ganise some de­bates. by Nathan Young (18 Apr 2023 16:13 UTC; 35 points)
  • Hauke Hillebrandt's comment on Nathan Young’s Quick takes by Nathan Young (17 Feb 2023 12:32 UTC; 2 points)

• Brian_Tomasik 25 Jun 2019 13:56 UTC

  43 points

  1 ∶ 0

  on: How Much Do Wild An­i­mals Suffer? A Foun­da­tional Re­sult on the Ques­tion is Wrong.

  Congrats on fixing the error!

  When I first discussed Ng (1995)’s mathematical proof with some friends in 2006, they said they didn’t find it very convincing because it’s too speculative and not very biologically realistic. Other people since then have said the same, and I agree. I’ve cited it on occasion, but I’ve never considered the mathematical result of that particular model to be more than an extremely weak argument for the predominance of suffering.

  I think the intuition underlying the argument—that most offspring die not long after birth—is one of the reasons many people believe wild-animal suffering predominates. It certainly might be the case that this intuition is misguided, such as based on what you said: “when the probability of suffering increases, the severity of suffering should decrease.” I have an article that also discusses theoretical reasons why shorter-lived animals and animals who are less likely to ever reproduce may not feel as much pain or fear as we would from the same kinds of injuries.

  While I think these kinds of arguments are interesting, I give them relatively low epistemic weight because they’re so theoretical. I think the best way to assess the net hedonic balance of wild animals is to watch videos and read about their lives, seeing what kinds of emotions they display, and then come up with our own subjective opinions about how much pain and pleasure they feel. This method is biased by anthropomorphism, but it’s at least somewhat more anchored to reality than simple theoretical models. We could try to combat anthropomorphism a bit by learning more about how other animals make tradeoffs between combinations of bad and good things, and so on.

  For me, it will always remain obvious that suffering dominates in nature because I believe extreme, unbearable suffering can’t be outweighed by other organism-moments experiencing pleasure. In general, I think most of the disagreement about nature’s net hedonic balance comes down to differences in moral values rather than disagreements about facts. But yes, it remains useful to improve our frameworks for thinking about this topic, as you’re helping to do. :)

• Brian_Tomasik 21 Jan 2023 6:49 UTC

  38 points

  9 ∶ 0

  in reply to: Matthew_Barnett’s comment on: Thread for dis­cussing Bostrom’s email and apology

  Thanks. :) I mainly had in mind something more like wisdom, rather than intelligence. Social norms on particular topics are often not what you would expect by armchair reasoning. In many cases, you have to directly encounter people expressing those norms, or see news stories /​ hear gossip about people who have run afoul of those norms, to know what they are. Nerds who are very interested in science/​math/​theoretical things may be less likely to learn about these norms than the average person, despite having high fluid intelligence. (BTW, this is one reason I’ve updated toward thinking reading some amount of news is important.) I imagine that people told Bostrom that what he said in 1996 wasn’t cool, and if so, that was a useful learning experience for him. The only problem was that it was written down for posterity to see.

  I think cultural context is also relevant to judging these things. Most young people today (even most nerds) know that what Bostrom said (even though it was in the context of giving an example of what you shouldn’t say) would elicit strong negative reactions, given how much media attention these things receive. I assume this was less obvious to nerds in the 1990s (though it was probably fairly predictable even back then).

  For what it’s worth, my fifth-grade class was assigned to read The Great Gilly Hopkins, which includes a tasteless joke about the N-word (though in the context of suggesting the person making it was being an asshole). And in high school, in 2005, when reading Adventures of Huckleberry Finn, we used the N-word in relation to Jim without any problems, because that’s the word Twain used. The degree of sensitivity around these things has changed a lot in the last 10-20 years.

• Brian_Tomasik 22 Jan 2023 2:49 UTC

  37 points

  6 ∶ 0

  in reply to: Ives Parr’s comment on: Thread for dis­cussing Bostrom’s email and apology

  We ought to move back to the attitude that it is an ideal to not care about race, sex, gender, sexual orientation etc rather than that we need to always be thinking about these things.

  I plausibly agree. There are times and places to bring up racism and sexism, their historical contexts, and instances where they still exist today. But I also get the sense that people would generally be happier (plausibly even many minorities(?), though I’m not at all sure about that) if they ruminated on these ideas less often. Rumination can both exacerbate the pain of actual injustices and make one perceive injustices where they may not actually exist or don’t exist much (manspreading, Shirtgate, etc). Note that this point can also apply to anti-woke people: focusing a bit less on the perceived wrongs of cancel culture might make them happier.

  Believers in genetic racial IQ gaps often say their viewpoint is needed in cases like affirmative action, to show that it’s not necessarily discriminatory if the demographic composition of some elite group doesn’t match the demographic composition of the whole population. But if we were more race-blind and didn’t think much about demographic composition to begin with, then talking about the possibility of genetic racial IQ gaps would also be less relevant. In order for this to work, society would also have to improve on providing better education, nutrition, income, etc to poorer parts of society, because otherwise not thinking much about the demographic composition of elite groups could lead to not noticing the substantial environmental and cultural causes of inequality that definitely still remain and that affirmative action aims to overcome. OTOH, maybe it’s naive to expect significant improvement in society’s motivation to actually raise the living standards of poor people, in which case the “kludge” of affirmative action might be better than nothing.

  I think having a few visible examples of minorities in powerful positions, such as the first black or female president of the USA, can be pretty valuable as inspiring role models. It may also be the case that people would be deterred from entering workplaces where there are too few “of one’s own kind” (race, gender, etc), such as because of fearing harassment. Maybe most humans are actually too tribalistic to pull off race-blindness, gender-blindness, etc. IDK.

  In any case, I suspect it’s generally more fruitfuil to focus on helping poor people (of all races) economically than to focus on, say, discrimination in hiring, because I think economic and cultural inequalities drive a lot of the inequality in outcome that we observe. In elite settings, my experience is that hiring discrimination is often in the direction of favoring black, women, gay, etc candidates, though I’m sure discrimination against such groups still happens somewhat too.

  Even fewer probably think it is acceptable to commit acts of violence against a group because they are a member of a groups with a lower than average level of cognitive ability. I really never see these attitudes.

  I imagine those attitudes were common during slavery and colonization. For example, the Europeans who arrived in the New World in the centuries after 1492 probably considered the indigenous people inferior and therefore didn’t feel as guilty about enslaving or murdering them.

  In the contemporary West, I agree that the view you mention seems rarer. Society tends to take less action against violence or hardship endured by very poor people, and poverty correlates with lower cognitive ability, but this isn’t an intentional part of society’s ideology so much as a byproduct of apathy.

  In the case of non-human animals, even most SJWs think it’s fine to enslave and murder them, and probably a main reason SJWs would cite for this is that non-human animals have different brains than humans do, though it’s unclear how much this reason is about intelligence per se versus sentience.

• Brian_Tomasik 23 Jan 2023 5:07 UTC

  35 points

  5 ∶ 0

  in reply to: Pablo’s comment on: Thread for dis­cussing Bostrom’s email and apology

  Thanks. :) I feel somewhat bad about spending time on this topic rather than my usual focus areas, especially since many of my points were already made by others. Plus, as I mentioned and as Bostrom learned, anything you say about controversial topics online is fodder for political enemies to take out of context. But I have a (maybe non-utilitarian) impulse to stick up for what I think is right even if some people will dislike me for doing so. (For a time, my top-level comment here had a net agreement of −10 or so. Of course, maybe the downvoters were correct and I’m wrong.)

• Brian_Tomasik 21 Jan 2023 8:53 UTC

  34 points

  9 ∶ 0

  in reply to: Matthew_Barnett’s comment on: Thread for dis­cussing Bostrom’s email and apology

  Interesting point! I hadn’t even heard of “stating one’s pronouns while introducing oneself”, although maybe that’s because I rarely meet anyone in person.

  As you said, there’s a tension between young people having the cutting edge of norms versus older people knowing a greater quantity of norms, even though some may be stale.

  I think the obsession among young people with political correctness increased dramatically in the last 10 years, and it was barely a discussion topic when I was in pre-college school. Usually it seemed to be teachers and administrators trying to inculcate anti-bullying lessons into the students. At the anti-bullying workshops, students often rolled their eyes. So I’m not sure how true it would have been to say that students were at the vanguard of social norms in my school. (I went to a pretty liberal public school in upstate New York.)

  I may also be generalizing too much from my own past self, since I was often called “oblivious” at Bostrom’s 23-year age and wasn’t that well informed about scandals, maybe because I thought they were too gossip-y and not as important as “serious” topics. (Now I realize that gossip is actually very important.)

  If Bostrom learned anything—and indeed, he apologized within 24 hours—it was that saying something like that can be inadvisable even among friends.

  Yeah. He also said he only “recently” began to believe that speaking flippantly is unsuccessful, which I think jibes with my hypothesis of him being fairly oblivious. Many people would consider the ineffectiveness of speaking flippantly so obvious as to not be worth mentioning as any kind of realization.

• Brian_Tomasik 6 Sep 2022 6:00 UTC

  27 points

  1 ∶ 0

  in reply to: CarlShulman’s comment on: Cri­tique of MacAskill’s “Is It Good to Make Happy Peo­ple?”

  Good points!

  situations where it is easier for individuals to experience things multiple times in easy-to-process fashion and then form a behavioral response

  It’s not obvious to me that our ethical evaluation should match with the way our brains add up good and bad past experiences at the moment of deciding whether to do more of something. For example, imagine that someone loves to do extreme sports. One day, he has a severe accident and feels so much pain that he, in the moment, wishes he had never done extreme sports or maybe even wishes he had never been born. After a few months in recovery, the severity of those agonizing memories fades, and the temptation to do the sports returns, so he starts doing extreme sports again. At that future point in time, his brain has implicitly made a decision that the enjoyment outweighs the risk of severe suffering. But our ethical evaluation doesn’t have to match how the evolved emotional brain adds things up at that moment in time. We might think that, ethically, the version of the person who was in extreme pain isn’t compensated by other moments of the same person having fun.

  Even if we think enjoyment can outweigh severe suffering within a life, many people object to extending such tradeoffs across lives, when one person is severely harmed for the benefit of others. The examples in David’s comment were about interpersonal tradeoffs, rather than intrapersonal ones. It’s true that people impose small risks of extreme suffering on some for the happiness of others all the time, like in the case of driving purely for leisure, but that still begs the question of whether we should do that. Most people in the West also eat chickens, but they shouldn’t. (Cases like driving are also complicated by instrumental considerations, as Magnus would likely point out. Also, not driving for leisure might itself cause some people nontrivial levels of suffering, such as by worsening mental-health problems.)

• Brian_Tomasik 12 Apr 2023 11:13 UTC

  26 points

  5 ∶ 0

  in reply to: yefreitor’s comment on: No­body’s on the ball on AGI alignment

  I think humans may indeed find ways to scale up their control over successive generations of AIs for a while, and successive generations of AIs may be able to exert some control over their successors, and so on. However, I don’t see how at the end of a long chain of successive generations we could be left with anything that cares much about our little primate goals. Even if individual agents within that system still cared somewhat about humans, I doubt the collective behavior of the society of AIs overall would still care, rather than being driven by its own competitive pressures into weird directions.

  An analogy I often give is to consider our fish ancestors hundreds of millions of years ago. Through evolution, they produced somewhat smarter successors, who produced somewhat smarter successors, and so on. At each point along that chain, the successors weren’t that different from the previous generation; each generation might have said that they successfully aligned their successors with their goals, for the most part. But over all those generations, we now care about things dramatically different from what our fish ancestors did (e.g., worshipping Jesus, inclusion of trans athletes, preventing children from hearing certain four-letter words, increasing the power and prestige of one’s nation). In the case of AI successors, I expect the divergence may be even more dramatic, because AIs aren’t constrained by biology in the way that both fish and humans are. (OTOH, there might be less divergence if people engineer ways to reduce goal drift and if people can act collectively well enough to implement them. Even if the former is technically possible, I’m skeptical that the latter is socially possible in the real world.)

  Some transhumanists are ok with dramatic value drift over time, as long as there’s a somewhat continuous chain from ourselves to the very weird agents who will inhabit our region of the cosmos in a million years. But I don’t find it very plausible that in a million years, the powerful agents in control of the Milky Way will care that much about what certain humans around the beginning of the third millennium CE valued. Technical alignment work might help make the path from us to them more continuous, but I’m doubtful it will avert human extinction in the long run.

• Brian_Tomasik 13 Apr 2023 10:06 UTC

  23 points

  0 ∶ 0

  on: All AGI Safety ques­tions wel­come (es­pe­cially ba­sic ones) [April 2023]

  Suppose that near-term AGI progress mostly looks like making GPT smarter and smarter. Do people think this, in itself, would likely cause human extinction? How? Due to mesa-optimizers that would emerge during training of GPT? Due to people hooking GPT up to control of actions in the real world, and those autonomous systems would themselves go off the rails? Just due to accelerating disruptive social change that makes all sorts of other risks (nuclear war, bioterrorism, economic or government collapse, etc) more likely? Or do people think the AI extinction risk mainly comes when people start building explicitly agentic AIs to automate real-world tasks like making money or national defense, not just text chats and image understanding as GPT does?

• Brian_Tomasik 23 Jan 2023 5:21 UTC

  23 points

  5 ∶ 1

  in reply to: lastmistborn’s comment on: Thread for dis­cussing Bostrom’s email and apology

  I see. :) I would think people would consider biological differences much more plausible in the gender case than the race case. I’ve heard several people say that when you’re a parent to both a boy and a girl, the differences between them are unmistakeable even in the first ~2 years. I think many American adults at least privately understand that there are big biological differences between the brains of men and women, while most American adults probably expect no non-trivial biological racial brain differences. But yeah, any particular gender difference, such as the language gap, could be mostly or all environmental.

• Brian_Tomasik 31 Mar 2023 1:48 UTC

  22 points

  6 ∶ 1

  on: Paus­ing AI Devel­op­ments Isn’t Enough. We Need to Shut it All Down by Eliezer Yudkowsky

  I’m not particularly well informed about current EA discourse on AI alignment, but I imagine that two possible strategies are

  1. accelerating alignment research and staying friendly with the big AI companies

  2. getting governments to slow AI development in a worldwide-coordinated way, even if this angers people at AI companies.

  Yudkowsky’s article helps push on the latter approach. Making the public and governments more worried about AI risk does seem to me the most plausible way of slowing it down. If more people in the national-security community worry about AI risks, there could be a lot more attention to these issues, as well as the possibility of policies like limiting total computing power for AI training that only governments could pull off.

  I expect a lot of AI developers would be angry about getting the public and governments more alarmed, but if the effort to raise alarm works well enough, then the AI developers will have to comply. OTOH, there’s also a possible “boy who cried wolf” situation in which AI progress continues, nothing that bad happens for a few years, and then people assume the doomsayers were overreacting—making it harder to ring alarm bells the next time.

• Brian_Tomasik 9 Oct 2022 3:45 UTC

  22 points

  1 ∶ 0

  on: Can we ac­tu­ally con­firm that a ve­gan diet is com­pat­i­ble with high in­tel­lec­tual out­put?

  Maybe you could eat beef and dairy to hedge your bets nutritionally while minimizing your animal-suffering impact? Of course, your question can also be interesting to discuss in its own right, but to the extent you worry about the issue, this could be a solution.

• Brian_Tomasik 16 May 2023 10:27 UTC

  21 points

  4 ∶ 0

  in reply to: James Özden’s comment on: How much fung­ing is there with dona­tions to differ­ent EA an­i­mal char­i­ties?

  Your question is fairly relevant to the discussion because if I thought there was net positive value in the lives of wild animals, then I would have a lot fewer concerns about non-welfare-reform animal charities.

  I’ve had it on my todo list to check out that video and paper, but I probably won’t get to it any time soon, so for now I’ll just reply to the slides you asked about. :)

  Personally I would not want to live even as the two surviving adult fish, because they probably experience a number of moments of extreme suffering, at least during death if not earlier. They may be fearful of predators, they face unpleasant temperatures or other bad environmental conditions without being able to control them the way humans do with air conditioning and heating, they may face long periods of low food, there might be intraspecific aggression and sexual harassment (I don’t know if those behaviors apply to Atlantic cod, but they are common in some fish species), and there would be many other hardships. Most of these moments of suffering probably wouldn’t feel that bad, but a few of them might be unbearably awful.

  I said that I “personally” wouldn’t want to live as one of these surviving fish, but you might say that the real question is whether they would want to have these lives rather than not existing. We can’t ask fish that question, but if we imagine humans having similar lives as these fish, we could ask such humans that question. Maybe many of those humans would say during many moments of their lives that they were on balance glad to exist. However, I suspect that during some moments, such as the peak pain of dying, they would often change their minds and wish they hadn’t existed. Therefore, there’s no single individual whom we can ask whether his/​her life was net positive; there are multiple “individuals” within the animal’s life, some of whom are glad to exist and some of whom are horrified to exist. How we weigh up these conflicting opinions is ultimately a judgment call, and no amount of further empirical data on wild-animal welfare will resolve it. I take a suffering-focused approach to this dilemma and say that it’s not acceptable for the happier moments of the animal’s life to impose unbearable suffering on some other moments of the animal’s life. So for any animal that has moments of unmitigated, unbearable agony (as most animals do, if only when dying), its life is net negative in my view.

  But most people don’t take this suffering-focused approach. Many people think enough happy moments of life can outweigh something as awful as being eaten alive. So next I’ll discuss the specific numbers in those slides.

  If a baby fish is only enduring 10 seconds of agony when dying, as the first slide suggests, then it’s presumably dying from predation (or maybe a severe physical injury like being crushed or something). The next slide suggests that maybe the pain of predation is ¹⁰⁰⁄₁₀₀, compared against a presumed positive welfare of 0.1/​100 for ordinary life. So getting eaten alive is only 1000 times worse than the goodness of a typical moment of life. That might seem plausible if we only glance at the numbers, but it’s not at all plausible if we actually think about what it implies. Imagine that you endure getting eaten alive for 1 minute. These numbers say that a mere 1000 minutes of ordinary life could compensate for that. 1000 minutes is 16.7 hours, slightly more than the amount of time a typical person is awake in a day. So this ratio says that even if you spend a minute every day experiencing what it’s like to be eaten alive, then your life can still be welfare-neutral. I wonder if anyone would actually sign up for that. One of the least suffering-averse trade ratios I’ve heard someone endorse was that he’d be willing to experience being eaten alive for an extra week of life (IIRC; that conversation was a long time ago). (I guess there are also a few people who say even more extreme things like “I’d rather be alive and tortured forever than not exist”, though I expect they’d change their minds pretty quickly when the torture started.)

  One possible argument is that it’s illegitimate to rely on our human intuitions about this tradeoff, because r-strategists may have evolved different pain-pleasure trade ratios based on the situation they face. For example, almost all fish babies will die by default, so if there’s an opportunity to take a dangerous risk in order to gain some slight advantage, they should probably take it, since they have almost no chance of winning otherwise. Therefore, maybe they need to be less averse to suffering (or at least less afraid of suffering) than we would be. This might be the case, but it’s a very theoretical argument, so I’m wary of putting too much stock in it. Of course, any estimate we have of how much suffering and pleasure exist in nature will be very speculative, so if I were a classical utilitarian who thought a minute of extreme suffering might be outweighable by a few days, weeks, or months of ordinary life in the wild, then I would have some uncertainty about the net hedonic balance of nature. But in my own case, I don’t think it’s ok to force extreme suffering on one for the pleasure of another—much less imposing extreme suffering on 1,999,998 for the pleasure of 2. (If we assume a 10% hatch rate and a 10% chance of sentience, then this comparison is actually 19,999.98 vs 2. And if we look at individual organism-moments of experience, the 2 surviving fish have a lot of organism-moments.)

  As I mentioned, the slides seem to be assuming deaths by predation given how short the duration of suffering is. Death by almost anything else would probably take hours, days, weeks, etc, although the intensity of pain during that time would usually be a lot lower than the intensity of pain during predation. This article says:

  A new study has uncovered the reason why 90 percent of fish larvae are biologically doomed to die mere days after hatching. This understanding of the mechanism that kills off the majority of the world’s fish larvae may help find a solution to the looming fish crisis in the world. The research suggests that “hydrodynamic starvation,” or the physical inability to feed due to environmental incompatibility, is the reason so many fish larvae perish.

  So maybe rather than 10 seconds, the period of pain while dying should be measured in hours or days? 1 day = 86,400 seconds. Of course, the badness of most of those seconds would be a lot less than ¹⁰⁰⁄₁₀₀.

  See also: “Is There More Suffering Than Happiness in Nature? A Reply to Michael Plant”.

• Brian_Tomasik 21 Jan 2023 0:10 UTC

  21 points

  9 ∶ 1

  in reply to: Jason’s comment on: Thread for dis­cussing Bostrom’s email and apology

  Fair enough. :) Some headlines called the FTX leadership “a gang of kids”, which I think isn’t unreasonable, even though they were in their late 20s or early 30s. The main thing I wanted to convey is that people at this age often have limited life experience or understanding of how the world works and so often do dumb things. Youth is a time to explore weird ideas and make mistakes. Therefore, I would agree that 23-year-olds generally shouldn’t be entrusted to hold important decision-making positions unless they’ve shown a track record of unusual maturity.

Suffer­ing in An­i­mals vs. Humans