EA (the movement) has a number of very serious problems, but I'd argue that the Sams are pretty much the least of them. Unless we're only talking about public opinion/PR, anyway.
The article is paywalled, but more generally, can someone defend EA to me? Does EA offer a specific idea that we should adopt in our thinking?
When I think of other, recent intellectual movements, I can clearly state "why" I should support them.
For example, YIMBY, the "idea" that producing more housing will lead to lower housing costs.
Or, Energy Transition, the "idea" that the falling costs of solar, renewables, and batteries will enable everyone to enjoy the fruits of industrial civilization, at lower costs, and without CO2 emissions.
Or, 15-minute cities, the "idea" that if I can walk to everything, I'll be healthier, it'll be more convenient, and we can save money driving less - with smaller roads and less parking.
But what does EA offer? How should we act differently? And if so, what will get better? I associate EA with concerns about AI safety, but I've never seen a coherent plan to address AI safety. And, if not AI safety - what's the "idea" that EA offers?
EA (the philosophy) is simple and reasonable: the idea is to engage in altruism that has the greatest impact possible. Essentially, getting the most bang for your buck.
EA (the movement) takes that simple and reasonable idea and extrapolates it out to support some truly crazy and/or questionable efforts, some of which seem very harmful. In my view, the movement is uncomfortably cultish.
There is no greater impact than government action, that is taxation and redistribution. It will reduce inequality and raise the standard of living precisely because the government is cost effective.
> There is no greater impact than government action, that is taxation and redistribution.
The issue is not which has the "greater impact" in the absolute (where government action may win) but which has the greater impact for spending extra money/resources. It's quite clear that EA's are getting a lot more bang for the buck there than, e.g. government foreign aid ever did.
First, the effective altruists did a bunch of harm to get all the money, then put it on something focused that not everyone agrees is the best thing to do with it.
They can measure the results of the spending as being higher, but don't account for the harm they needed to do to accumulate the money.
How do you measure spending that will only affect people living in the distant future? This is the problem with the longtermist end of the EA movement.
> (the movement) takes that simple and reasonable idea and extrapolates it out to support some truly crazy and/or questionable efforts
You've perfectly described the fundamental problem with many recently popular philosophical propositions. The starting premise sounds reasonable because it is reasonable. Yet that basic premise is then extended to include much more controversial things. These initially reasonable things range from "It would be good to incorporate some safety measures as we implement AI" to "Racism is bad." Obviously, reasonable propositions it's hard to disagree with. Yet somehow we end up with influential social movements LARPing Terminator II scenarios and serious people arguing the police should be defunded.
I'm starting to suspect this repeating dark pattern isn't happening by accident. If your movement has some pretty extreme ideas most people wouldn't agree with, it's more effective to start with ideas almost no one will disagree with. Then, after a lot of reasonable people have bought-in to that, slowly increment toward justifying increasingly extreme measures to achieve "justice". It's essentially bait-and-switch. The unfortunate side effect is now I've learned to be leery of agreeing with even seemingly agreeable things because it's unclear what other things might later be included.
This puts me in the kind of bizarre conversations where someone says "I'm an anti-racist, are you?" and I have to craft a weirdly qualified response like, "Well, I'm certainly opposed to racism but I need to understand what 'anti-racism' includes in this context because some people include policies under that term which have net effects that seem to get uncomfortably close to racism."
I think what you described is something that happens inherently in movements that become echo chambers. As the movement gets more extreme more and more of the reasonable people are driven out or silenced. Eventually the movement is controlled by a tiny minority of extremists and the majority of reasonable people feel compelled to agree with them or leave. Then the movement collapses.
This phenomenon has taken place many times in history. Look at how the ideas of socialism became extreme in China during the cultural revolution or in The USSR during the peak of Stalinism.
Anyway eventually the ultra Maoists in China lost control and China went in a different direction. Russia too. So at the most extreme things do seem to collapse and move in the opposite direction…
As I understand it, EA is about deployment of money, not effort. You accumulate massive money, ostensibly in order to fund things of interest. There is a built-in assumption that other people (labor) will actually perform whatever work you're interested in funding. You don't go build things or heal people, you accumulate wealth and buy from someone the building or the healing.
And concern about AI safety/ethics is a lot broader than just EA these days. It actually seems to be one of the things that EA got right before they were popular, along with pandemic preparedness and other global risks.
The most basic insight of EA is to be efficient in your altruism.
It’s an extension of the effort years ago for nonprofits to report their management costs: If you want to support cancer research, you don’t want to give it to the charity that uses 60% of funds for marketing, you give it to the charity that sends 90% to quality research.
Similar if your goal is more broad than cancer research, like say you just want fewer children dying. Well then you can to the research and find that mosquito nets are dollar-for-dollar the best use of money for reducing child mortality, better than for example paying for more doctors. (*These are just examples)
The more “famous” conclusions of EA are actually less core to the movement, from what I understand. For example, the advise that one can do the most good by getting a high paying job and donating the surplus.
If you’re curious, I would start by listening to William MacAskill’s appearance on “Conversations with Tyler”[1]. It predates the whole Sam Bankman-Fried thing so you can make your mind up purely on the intellectual merits rather than being coloured by that mess.
Then if you’re serious about understanding it further, the logical next step would probably be to read “Doing Good Better”, where MacAskill really sets the whole thing out.
There’s something about the sort of “fundamentalist” utilitarianism of his ideas that really troubles me especially when taken to its logical conclusion so even when I logically agree with them I don’t personally find it easy to actually “sign up” so to speak, but it might resonate with you idk.
That podcast is from August 2022, which is well after SBF began making major contributions to EA organizations using misbegotten funds. In fact it’s likely that his giving helped pay for the entire MacAskill book promotion effort, which could well have been connected to that podcast (which is dated six days before the book publication date.) It’s reasonable to assume that MacAskill and other EA luminaries didn’t know that SBF was funding their organizations with misbegotten funds, but on the other hand there’s no evidence they looked very closely either. SBF’s close involvement and financially-based access certainly do color my impression of the movement.
Those things should definitely colour your opinion[1]. What I’m saying is, if you want to give the EA ideas the benefit of the doubt, this is the strongest possible version that I can think of as an intro - one of the main dudes given a chance to really explain his point of view while being gently challenged intellectually, but without just guilt by association with SBF. Then if it survives that, you can investigate further if curious. If it doesn’t survive that, then the SBF thing might be the thing that finally sinks the ship but it was already taking on a lot of water already.
[1] In my opinion at least- they certainly do for me. I already had some qualms about it before that but it sort of confirmed a queasiness that I had about how it all seems pretty convenient. You can be megarich and live in a very selfish way and that’s alright because you’re earning to give etc. While I like the fact that they try to measure (and optimise) impact, I also have a big problem with the way MacAskill in particular seems to only value impacts he can measure, so philanthropy with social benefits - donating to things like art, culture, museums etc he basically says are completely worthless. That just seems very obviously wrong to me. From a personal point of view, I have seen the social mobility benefits that arise from things like art or music scholarships, where the benefit accrues to the family of the scholarship student even if (maybe especially if) they don’t go on to pursue a career in the arts.
I’ve read some of MacAskill’s writings, and my objection to his philosophy is twofold. First, while he talks about the value of “measurable contributions”, many of the things he finds important aren’t measurable. Longtermist investments and things that maximize the utility of future humans are almost by definition impossible to measure. This undermines the entire founding ethos of EA. What substitutes for true measurement in these cases is something more like opinion, where major EA orgs can then justify buying a luxury retreat for meetings since it will “help to produce an environment conducive to good ideas, and those ideas could save countless lives” and other nonsense like that.
Secondly there’s the obsession with cultivating and catering to high-earning donors, which in this case has a long and “rational” justification. But is, in the end, largely similar to what all traditional (non-effective) charities do. The major difference in EA is that there is even less incentive to look closely at what the donors do, and what kind of human beings they are, since the only measure that “matters” is the donation amount.
In that sense “the SBF fiasco” wasn’t so much an outlier but an inevitable result of the entire approach that specific wing of EA has been taking. I also agree that the focus on measurability harms difficult-to-measure causes like the arts, but I sort of expect that from a tech-nerd focused charity anyway.
> Does EA offer a specific idea that we should adopt in our thinking?
The specific idea is that while people at large may have the best intentions about charity and social good, the impact of their money and effort could be greatly magnified if they deployed it better.
For example most donations happen to already overfunded causes (say cancer research) where your dollar is getting lost among executive bonuses and generally not having the kind of impact you think. There are tons of more neglected areas where you can do much more immediate, quantifiable good.
Another example from their website is that governments worldwide spend $280B per year on counterterrorism and $8B on preventing public health crises. If you look at the relative human impact from the two, those numbers are laughable.
There are many, many such examples of obvious inefficiency in the system that can be solved with a better thought framework and more research.
You can 'turtles all the way down' it all you want (A god fearing person goes to heaven for selfish reasons, blah blah), but how is altruism a myth - Could you possibly elaborate? (TIA)
Anyway, I do agree that altruism is arguably not the correct term for this movement, given that the members probably look after themselves first (subconsciously or otherwise) then are able to engage in giving (and cannot truly be 'selfless' in the hermetically sealed sense of the word).
Much altruism - especially in the Silicon Valley subculture sense - comes with positive externalities. You gain influence, popularity, brand, whatever, from activities that always seem to go well publicized.
If you're gaining something, and you know you're gaining something going into it, can it really be altruistic?
OK thanks yes, that's the interpretation (/thing that undermines true selflessness) that I had (and had poorly worded in my comment). I suppose I thought I was missing something else that OP was implying.
That's a reasonable and necessary question. Of course, the answer hinges on how one chooses to define altruism and any values framework deriving from it (cue Moral Philosophy 101). Like a lot of things which are good within reason, if taken to extremes, unconstrained altruism can lead to some pretty strange moral quicksand and unworkable outcomes. Deciding what is "reasonable" and what is "extreme" falls into the Venn overlap between "It's complicated" and "It depends."
As much shit as Ayn Rand gets, when I actually read her philosophical writing specifically on altruism I thought she got one thing right. If altruism becomes (or is assumed to be) a moral obligation rather than a personal contextual choice, it doesn't lead to a workable society - at least at scale and in the long-term. That had never occurred to me until I read her reasoning about it.
30 comments
[ 4.2 ms ] story [ 68.1 ms ] threadWhen I think of other, recent intellectual movements, I can clearly state "why" I should support them.
For example, YIMBY, the "idea" that producing more housing will lead to lower housing costs.
Or, Energy Transition, the "idea" that the falling costs of solar, renewables, and batteries will enable everyone to enjoy the fruits of industrial civilization, at lower costs, and without CO2 emissions.
Or, 15-minute cities, the "idea" that if I can walk to everything, I'll be healthier, it'll be more convenient, and we can save money driving less - with smaller roads and less parking.
But what does EA offer? How should we act differently? And if so, what will get better? I associate EA with concerns about AI safety, but I've never seen a coherent plan to address AI safety. And, if not AI safety - what's the "idea" that EA offers?
EA (the movement) takes that simple and reasonable idea and extrapolates it out to support some truly crazy and/or questionable efforts, some of which seem very harmful. In my view, the movement is uncomfortably cultish.
Here's a non-paywalled link to the article: https://archive.is/paC4K
What's you're best example of a government that is following your advice?
Do trolls have a right to food? I probably shouldn't be feeding you on off-topic subjects. https://news.ycombinator.com/newsguidelines.html
The issue is not which has the "greater impact" in the absolute (where government action may win) but which has the greater impact for spending extra money/resources. It's quite clear that EA's are getting a lot more bang for the buck there than, e.g. government foreign aid ever did.
First, the effective altruists did a bunch of harm to get all the money, then put it on something focused that not everyone agrees is the best thing to do with it.
They can measure the results of the spending as being higher, but don't account for the harm they needed to do to accumulate the money.
> (the movement) takes that simple and reasonable idea and extrapolates it out to support some truly crazy and/or questionable efforts
You've perfectly described the fundamental problem with many recently popular philosophical propositions. The starting premise sounds reasonable because it is reasonable. Yet that basic premise is then extended to include much more controversial things. These initially reasonable things range from "It would be good to incorporate some safety measures as we implement AI" to "Racism is bad." Obviously, reasonable propositions it's hard to disagree with. Yet somehow we end up with influential social movements LARPing Terminator II scenarios and serious people arguing the police should be defunded.
I'm starting to suspect this repeating dark pattern isn't happening by accident. If your movement has some pretty extreme ideas most people wouldn't agree with, it's more effective to start with ideas almost no one will disagree with. Then, after a lot of reasonable people have bought-in to that, slowly increment toward justifying increasingly extreme measures to achieve "justice". It's essentially bait-and-switch. The unfortunate side effect is now I've learned to be leery of agreeing with even seemingly agreeable things because it's unclear what other things might later be included.
This puts me in the kind of bizarre conversations where someone says "I'm an anti-racist, are you?" and I have to craft a weirdly qualified response like, "Well, I'm certainly opposed to racism but I need to understand what 'anti-racism' includes in this context because some people include policies under that term which have net effects that seem to get uncomfortably close to racism."
This phenomenon has taken place many times in history. Look at how the ideas of socialism became extreme in China during the cultural revolution or in The USSR during the peak of Stalinism.
Anyway eventually the ultra Maoists in China lost control and China went in a different direction. Russia too. So at the most extreme things do seem to collapse and move in the opposite direction…
And concern about AI safety/ethics is a lot broader than just EA these days. It actually seems to be one of the things that EA got right before they were popular, along with pandemic preparedness and other global risks.
It’s an extension of the effort years ago for nonprofits to report their management costs: If you want to support cancer research, you don’t want to give it to the charity that uses 60% of funds for marketing, you give it to the charity that sends 90% to quality research.
Similar if your goal is more broad than cancer research, like say you just want fewer children dying. Well then you can to the research and find that mosquito nets are dollar-for-dollar the best use of money for reducing child mortality, better than for example paying for more doctors. (*These are just examples)
The more “famous” conclusions of EA are actually less core to the movement, from what I understand. For example, the advise that one can do the most good by getting a high paying job and donating the surplus.
Then if you’re serious about understanding it further, the logical next step would probably be to read “Doing Good Better”, where MacAskill really sets the whole thing out.
There’s something about the sort of “fundamentalist” utilitarianism of his ideas that really troubles me especially when taken to its logical conclusion so even when I logically agree with them I don’t personally find it easy to actually “sign up” so to speak, but it might resonate with you idk.
[1] https://conversationswithtyler.com/episodes/william-macaskil...
[1] In my opinion at least- they certainly do for me. I already had some qualms about it before that but it sort of confirmed a queasiness that I had about how it all seems pretty convenient. You can be megarich and live in a very selfish way and that’s alright because you’re earning to give etc. While I like the fact that they try to measure (and optimise) impact, I also have a big problem with the way MacAskill in particular seems to only value impacts he can measure, so philanthropy with social benefits - donating to things like art, culture, museums etc he basically says are completely worthless. That just seems very obviously wrong to me. From a personal point of view, I have seen the social mobility benefits that arise from things like art or music scholarships, where the benefit accrues to the family of the scholarship student even if (maybe especially if) they don’t go on to pursue a career in the arts.
Secondly there’s the obsession with cultivating and catering to high-earning donors, which in this case has a long and “rational” justification. But is, in the end, largely similar to what all traditional (non-effective) charities do. The major difference in EA is that there is even less incentive to look closely at what the donors do, and what kind of human beings they are, since the only measure that “matters” is the donation amount.
In that sense “the SBF fiasco” wasn’t so much an outlier but an inevitable result of the entire approach that specific wing of EA has been taking. I also agree that the focus on measurability harms difficult-to-measure causes like the arts, but I sort of expect that from a tech-nerd focused charity anyway.
The specific idea is that while people at large may have the best intentions about charity and social good, the impact of their money and effort could be greatly magnified if they deployed it better.
For example most donations happen to already overfunded causes (say cancer research) where your dollar is getting lost among executive bonuses and generally not having the kind of impact you think. There are tons of more neglected areas where you can do much more immediate, quantifiable good.
Another example from their website is that governments worldwide spend $280B per year on counterterrorism and $8B on preventing public health crises. If you look at the relative human impact from the two, those numbers are laughable.
There are many, many such examples of obvious inefficiency in the system that can be solved with a better thought framework and more research.
Anyway, I do agree that altruism is arguably not the correct term for this movement, given that the members probably look after themselves first (subconsciously or otherwise) then are able to engage in giving (and cannot truly be 'selfless' in the hermetically sealed sense of the word).
Much altruism - especially in the Silicon Valley subculture sense - comes with positive externalities. You gain influence, popularity, brand, whatever, from activities that always seem to go well publicized.
If you're gaining something, and you know you're gaining something going into it, can it really be altruistic?
That's a reasonable and necessary question. Of course, the answer hinges on how one chooses to define altruism and any values framework deriving from it (cue Moral Philosophy 101). Like a lot of things which are good within reason, if taken to extremes, unconstrained altruism can lead to some pretty strange moral quicksand and unworkable outcomes. Deciding what is "reasonable" and what is "extreme" falls into the Venn overlap between "It's complicated" and "It depends."
As much shit as Ayn Rand gets, when I actually read her philosophical writing specifically on altruism I thought she got one thing right. If altruism becomes (or is assumed to be) a moral obligation rather than a personal contextual choice, it doesn't lead to a workable society - at least at scale and in the long-term. That had never occurred to me until I read her reasoning about it.