In the last 80.000 podcast Hugh White made a compelling case that the US cares less about Taiwan than China. Given shifts in relative might, I wonder how feasible it is as a long-term strategy to hope to deter China forever.
This article seems to assume under current deterrence China sees the costs of a Taiwan invasion as higher than the long-term benefits. Still, I'm a little afraid part of their calculus is an expectation that the costs will be lower in the future as their relative might increases and Taiwan might opt for a more peaceful resolution. If this is the case any build-up in deterrence might only provoke China into an invasion, as the expectation of future costs increase and long-term Taiwanese independence is unacceptable to them.
More military power could also increase the costs in case of an invasion, so you need to be quite confident that deterrence will work and continue working indefinitely.
To me, the most tractable part of this equation seems to be the chance that the USA intervenes, but I'm profoundly uncertain if it would be better for the world if this were higher or lower. Do you have any views on that?
I'm a little nervous around geopolitics as a philanthropic cause area cause it seems so easy to harm when information and calculations are imperfect (and our information will be extremely imperfect). Decreasing the likelihood of conflict often comes from either sacrificing a lot of well-being by conceding important points (if Taiwan became part of China that would surely reduce the probability of great power conflict, at least in the short-medium term, but be awful in many other ways), increasing the cost of conflict (as I think your suggestion would do) and decreasing the cost often seems to come from increasing the likelihood/conceding important points. Do you think there are clear things we can do in this space to mitigate downside risks?
I fully understand your anger, but it feels misdirected. If EA or GiveWell could have successfully protected democracy in the US or prevented Trump from getting elected we should have done that. However it's very unclear that EA as a whole and especially GiveWell as an organization could have prevented this mess . Given the mess we are in what GiveWell is doing will still make it so many people survive who otherwise would have died. Every one of those people who might get to make friends, find love and meaning matter a ton. And that is not a solution to the problem, but it still good work.
So no millions cannot fill a gap of billions, but millions is still a lot better than nothing. If somehow those millions could be spent in a way that would have prevented the billions from disappearing, that would obviously have been more effective. But many millions have been spent on trying to prevent Trump 2.0 (including by EA's) and it's not clear what EA could have done.
It can not be simultaneously true that the awefulness of the current situation is a consequence of EA 'forgetting about democracy', and that protecting democracy is not tractable. If EA fully focussing on democracy as it's singular cause area would have prevented the current situation protecting democracy would be tractable, that is what tractability means. The consequence you are drawing requires a lot more evidence than you are providing.
I'm sure many people at GiveWell are just as angry about the current situation as you are, they see the suffering every day. The argument that this is partly EAs fault requires a lot more evidence.
I wonder if this is mostly an effect of total effective giving growing less than GiveWell anticipated, or if more money is going to other effective giving opportunities.
Some evidence for the latter would be that in Europe different effective giving initiatives are growing (for example effektiv-spenden in Germany raised 13 million in 2022, and 19 million in 2023, doneer-effectief in the Netherlands raised 700.000 in 2022, and 2.5 million in 2023, I don't know if there is a similar trend in other EU-countries). I sense that within EA there is some shift in focus to causes such as animal wellbeing. Anecdotally I shifted my giving away from GiveWell towards the Doneer-Effectief animal wellbeing fund.
However, these trends are not totally unexpected and in size might only explain a fraction of the 75 million difference between the median estimate and actual data of non-open philanthropy funding. Does anybody have a good sense of how effective giving as a whole changed between 2022 and 2023?
I think you are too optimistic about how much the average person in the global north care about people in the global south (and possibly too pessimistic about how much people care about animals, but less confident about that).
"Saving children from malaria, diarrheal disease, lead poisoning, or treating cataracts and obstetric fistula is hard to argue against without sounding like a bad person."
The argument that you should help locally instead (even if the people making that argument don't do so) is easily made without sounding like a bad person. I live in the Netherlands and any spending on developmental cooperation or charity work tends to be very unpopular, our current government also pledged to cut a lot of that spending. Pushback on 'giving money to Africa' is something I certainly encounter, it might be less than for being vegan but I'm not sure and also not sure by how much. I would want better data on this, this piece seems to assume a big difference in how socially acceptable, moderate and politizised global health is compared to animal welfare. I would like to see better data on whether that assumption is true or not.
I find the slippery slope argument pretty weak, historically expanding the moral circle (to women, slaves, people of color, lhbtqi+ people) has been quite important and still requires a lot of work. It seems much easier to not go far enough than to go too far, and expanding the moral circle to animals or people in the global south is both an expansion. A core strength of EA is that there is a lot of attention to groups that are outside the moral circle of many comperatively affluent and wealthy people or institutions (the global poor, animals, and future generations). The slippery-slope argument is only meaningful if going down the slope would be bad, which is unclear to me.
It makes me quite sad that in practice EA has become so much about specific answers (work on AI risk, donate to this charity, become vegan) to the question of how we effective make the world a better place, that not agreeing with a specific answer can create so much friction. In my mind EA really is just about the question itself and the world is super complicated so we should be skeptical of any particular answer.
If we accidentally start selecting for people that intuitive agree with certain answers (which it sounds like we are doing, I know people that have a deep desire to make a lot of counterfactual impact, but were turned of because they ‘disagreed’ with some common EA belief, and sounds like if you read superintelligence earlier that would have been the case for you as well) that has a big negative effect on our epistemics and ultimately hurts our goal. We won’t be able to check each others biases and have a less diverse set of views and viewpoints.
I think many of these lessons have more merrit to them than you assume. To speak specifically about the ‘earning to give’ one, yes EA has pointed out that you should not do harm with your job to give it away. However I also think it is a bit psychologically naïeve to think that what happened with FTX is the last time that giving people the advice of earning to give is the last time it will lead to people doing harm to make money.
Trade-offs between ethical principles and monetary gain are not rare, and once we have established making as much money as possible (to give it away) as a goal in itself and something that gives status, it can be hard to make these trade-offs the way you are supposed to. It is not easy to accept a setback in wealth, power and (moral) status so lying to yourself or others to think that what you are doing is ethical becomes easy. It is also generally risky for individuals to become incredibly rich or powerful, especially if that depends on a misguided believe that some group membership (ea) makes you inherently ethical and therefore more trustworthy, since power tends to corrupt.
At the minimum I would like EA to talk more about how to jointly maximize the ethics of how you earn and spend your money, making sure that we promote people to gain their wealth in ways that add value to the world.
I do think there is something here, while I do not agree with everything taking away the lesson that we should think more about power and how to prevent it from being concentrated seems good. If you look at what EA has written it is clear that what SBF did was against almost everything that has been said about how to do good (do not do harm, ends don't justify the means, act with integrity and in accordance with common-sense altruism). However, he was originally inspired by EA, and might have started out following the principles but when things went sour abandoned them. It is common psychological knowledgde that power tends to corrupt, so in a sense him not just conceding power when and thinking 'a hit and a miss' might not be that unexpected. In this sense instead of writing better advices of what to do when you have a lot of power (either political or wealth) me might need to focus more on making sure power is not concentrated in the first place.
Thanks for the responses. I was not aware of the article on harmful careers and think it is very good (I recognize that many of these issues are hard, so even tough I might be a bit more skeptical about some of these high paying jobs and examples I could easily be wrong). Thanks for bringing it to my attention, it shows that some of my criticism was a bit misguided.
Agree with the postmortem process, there is a reasonable chance that SBF used EA type thinking to justify his behaviour and we certainly celebrated him as some kind of hero. I think it is important to not just condemn fraud but also really try to figure out if there is stuff EA did or advice it gives that incentivizes this kind of behaviour.
In the last 80.000 podcast Hugh White made a compelling case that the US cares less about Taiwan than China. Given shifts in relative might, I wonder how feasible it is as a long-term strategy to hope to deter China forever.
This article seems to assume under current deterrence China sees the costs of a Taiwan invasion as higher than the long-term benefits. Still, I'm a little afraid part of their calculus is an expectation that the costs will be lower in the future as their relative might increases and Taiwan might opt for a more peaceful resolution. If this is the case any build-up in deterrence might only provoke China into an invasion, as the expectation of future costs increase and long-term Taiwanese independence is unacceptable to them.
More military power could also increase the costs in case of an invasion, so you need to be quite confident that deterrence will work and continue working indefinitely.
You also mentioned this equation:
"37% chance of invasion × 50% USA intervenes × 5% nuclear risk = 0.9% nuclear war"
To me, the most tractable part of this equation seems to be the chance that the USA intervenes, but I'm profoundly uncertain if it would be better for the world if this were higher or lower. Do you have any views on that?
I'm a little nervous around geopolitics as a philanthropic cause area cause it seems so easy to harm when information and calculations are imperfect (and our information will be extremely imperfect). Decreasing the likelihood of conflict often comes from either sacrificing a lot of well-being by conceding important points (if Taiwan became part of China that would surely reduce the probability of great power conflict, at least in the short-medium term, but be awful in many other ways), increasing the cost of conflict (as I think your suggestion would do) and decreasing the cost often seems to come from increasing the likelihood/conceding important points. Do you think there are clear things we can do in this space to mitigate downside risks?
I fully understand your anger, but it feels misdirected. If EA or GiveWell could have successfully protected democracy in the US or prevented Trump from getting elected we should have done that. However it's very unclear that EA as a whole and especially GiveWell as an organization could have prevented this mess . Given the mess we are in what GiveWell is doing will still make it so many people survive who otherwise would have died. Every one of those people who might get to make friends, find love and meaning matter a ton. And that is not a solution to the problem, but it still good work.
So no millions cannot fill a gap of billions, but millions is still a lot better than nothing. If somehow those millions could be spent in a way that would have prevented the billions from disappearing, that would obviously have been more effective. But many millions have been spent on trying to prevent Trump 2.0 (including by EA's) and it's not clear what EA could have done.
It can not be simultaneously true that the awefulness of the current situation is a consequence of EA 'forgetting about democracy', and that protecting democracy is not tractable. If EA fully focussing on democracy as it's singular cause area would have prevented the current situation protecting democracy would be tractable, that is what tractability means. The consequence you are drawing requires a lot more evidence than you are providing.
I'm sure many people at GiveWell are just as angry about the current situation as you are, they see the suffering every day. The argument that this is partly EAs fault requires a lot more evidence.
I wonder if this is mostly an effect of total effective giving growing less than GiveWell anticipated, or if more money is going to other effective giving opportunities.
Some evidence for the latter would be that in Europe different effective giving initiatives are growing (for example effektiv-spenden in Germany raised 13 million in 2022, and 19 million in 2023, doneer-effectief in the Netherlands raised 700.000 in 2022, and 2.5 million in 2023, I don't know if there is a similar trend in other EU-countries). I sense that within EA there is some shift in focus to causes such as animal wellbeing. Anecdotally I shifted my giving away from GiveWell towards the Doneer-Effectief animal wellbeing fund.
However, these trends are not totally unexpected and in size might only explain a fraction of the 75 million difference between the median estimate and actual data of non-open philanthropy funding. Does anybody have a good sense of how effective giving as a whole changed between 2022 and 2023?
I donated most to the animal welfare fund of 'doneer effectief' and some to their global health and wellbeing fund.
I think you are too optimistic about how much the average person in the global north care about people in the global south (and possibly too pessimistic about how much people care about animals, but less confident about that).
"Saving children from malaria, diarrheal disease, lead poisoning, or treating cataracts and obstetric fistula is hard to argue against without sounding like a bad person."
The argument that you should help locally instead (even if the people making that argument don't do so) is easily made without sounding like a bad person. I live in the Netherlands and any spending on developmental cooperation or charity work tends to be very unpopular, our current government also pledged to cut a lot of that spending. Pushback on 'giving money to Africa' is something I certainly encounter, it might be less than for being vegan but I'm not sure and also not sure by how much. I would want better data on this, this piece seems to assume a big difference in how socially acceptable, moderate and politizised global health is compared to animal welfare. I would like to see better data on whether that assumption is true or not.
I find the slippery slope argument pretty weak, historically expanding the moral circle (to women, slaves, people of color, lhbtqi+ people) has been quite important and still requires a lot of work. It seems much easier to not go far enough than to go too far, and expanding the moral circle to animals or people in the global south is both an expansion. A core strength of EA is that there is a lot of attention to groups that are outside the moral circle of many comperatively affluent and wealthy people or institutions (the global poor, animals, and future generations). The slippery-slope argument is only meaningful if going down the slope would be bad, which is unclear to me.
It makes me quite sad that in practice EA has become so much about specific answers (work on AI risk, donate to this charity, become vegan) to the question of how we effective make the world a better place, that not agreeing with a specific answer can create so much friction. In my mind EA really is just about the question itself and the world is super complicated so we should be skeptical of any particular answer.
If we accidentally start selecting for people that intuitive agree with certain answers (which it sounds like we are doing, I know people that have a deep desire to make a lot of counterfactual impact, but were turned of because they ‘disagreed’ with some common EA belief, and sounds like if you read superintelligence earlier that would have been the case for you as well) that has a big negative effect on our epistemics and ultimately hurts our goal. We won’t be able to check each others biases and have a less diverse set of views and viewpoints.
I think many of these lessons have more merrit to them than you assume. To speak specifically about the ‘earning to give’ one, yes EA has pointed out that you should not do harm with your job to give it away. However I also think it is a bit psychologically naïeve to think that what happened with FTX is the last time that giving people the advice of earning to give is the last time it will lead to people doing harm to make money.
Trade-offs between ethical principles and monetary gain are not rare, and once we have established making as much money as possible (to give it away) as a goal in itself and something that gives status, it can be hard to make these trade-offs the way you are supposed to. It is not easy to accept a setback in wealth, power and (moral) status so lying to yourself or others to think that what you are doing is ethical becomes easy. It is also generally risky for individuals to become incredibly rich or powerful, especially if that depends on a misguided believe that some group membership (ea) makes you inherently ethical and therefore more trustworthy, since power tends to corrupt.
At the minimum I would like EA to talk more about how to jointly maximize the ethics of how you earn and spend your money, making sure that we promote people to gain their wealth in ways that add value to the world.
I do think there is something here, while I do not agree with everything taking away the lesson that we should think more about power and how to prevent it from being concentrated seems good. If you look at what EA has written it is clear that what SBF did was against almost everything that has been said about how to do good (do not do harm, ends don't justify the means, act with integrity and in accordance with common-sense altruism). However, he was originally inspired by EA, and might have started out following the principles but when things went sour abandoned them. It is common psychological knowledgde that power tends to corrupt, so in a sense him not just conceding power when and thinking 'a hit and a miss' might not be that unexpected. In this sense instead of writing better advices of what to do when you have a lot of power (either political or wealth) me might need to focus more on making sure power is not concentrated in the first place.
Thanks for the responses. I was not aware of the article on harmful careers and think it is very good (I recognize that many of these issues are hard, so even tough I might be a bit more skeptical about some of these high paying jobs and examples I could easily be wrong). Thanks for bringing it to my attention, it shows that some of my criticism was a bit misguided.
Agree with the postmortem process, there is a reasonable chance that SBF used EA type thinking to justify his behaviour and we certainly celebrated him as some kind of hero. I think it is important to not just condemn fraud but also really try to figure out if there is stuff EA did or advice it gives that incentivizes this kind of behaviour.