4 ms·

Then the followup question is "What is being effective to you? Based on your second assumption, it appears that being effective to you means, at worst, only hel

Then the followup question is "What is being effective to you? Based on your second assumption, it appears that being effective to you means, at worst, only helping those close to you, and at best, helping in ways that are visible to you. Do you think those in your community are the most in need of assistance, and even if they are, do you think that your view of what they need is more insightful than their opinion of what they need?" The way you've defined things, asking a homeless person for what they believe would most help them is not effective, because it is second hand.

keybored 4y ago
> Then the followup question is "What is being effective to you? See above. > Based on your second assumption, it appears that being effective to you means, at worst, only helping those close to you, and at best, helping in ways that are visible to you. Do you think those in your community are the most in need of assistance, and even if they are, do you think that your view of what they need is more insightful than their opinion of what they need?" It’s not (under those assumptions) about helping those close to me per se. It’s about doing good with a high probability. I know what they need so I am more likely to actually help them, instead of causing unintentional harm. It’s funny that this author chose the tower analogy. EA is like a skyscraper of assumptions, built with paper: send money through three continents in order to help Africans in a very specific way without necessarily knowing anything about their specific needs. Maybe the malaria nets will help save millions. Or maybe they will just end up polluting the waterways. > The way you've defined things, asking a homeless person for what they believe would most help them is not effective, because it is second hand. If I ask that person then I get first-hand experience with that person.
- joshuamorton 4y ago
  > EA is like a skyscraper of assumptions, built with paper: send money through three continents in order to help Africans in a very specific way without necessarily knowing anything about their specific needs. Maybe the malaria nets will help save millions. Or maybe they will just end up polluting the waterways. I don't follow this line of reasoning. It's basically anti-empiricism which can be used to support pretty much any ideology you want. There's evidence that the malaria nets help save lives. We can see that fewer people get malaria. We can see death-rates declining. You can argue that that doesn't actually improve life/quality of life. You can argue that there are better ways, but the claim that "maybe they will just end up polluting the waterways" is based, fundamentally, on an intentional lack of knowledge. Calling donating money to malaria nets or deworming "a skyscraper of assumptions built with paper" is like calling modern physics or modern cancer treatment "a skyscraper of assumptions built with paper". Sure, yeah, it's complex and you don't understand all of it, and neither do I, but we have very clear and obvious evidence that it works! > If I ask that person then I get first-hand experience with that person. No you haven't, you've been told something through their biases. It's second hand experience, which you claim is unreliable. They could be lying to you! If what matters is your direct experience and observation, a study on what is most beneficial to homeless people, and a recorded interview with a homeless person, and a direct in person conversation with a homeless person (or a study on malaria nets, or a recorded interview with someone whose describes how their town has far fewer malaria related deaths than a decade ago) are all equally bad and less reliable than your preconceived beliefs and in-the-wild observations of what people struggle with. > See above. You don't actually define "effective" anywhere. I'm asking you to do that. You say it's a foundational assumption, and that sticking with things you know well helps you be effective, but we still don't know what effective means, and certainly not what it means to you, hence my request, which your above post doesn't address. Consider at least two possibilities: 1. To you, "effective" means provably done. That is, if you do not have incontrovertible (say, directly observable evidence) that your work has done good, you assume the value of your investment is zero. 2. To you, effective means something akin to utilitarianly efficient, e.g. most utils per $. It sounds to me like you initially claimed (2) and had a number of higher level assumptions about low trust, but it appears now you're claiming (1), and perhaps there are other possibilities that I haven't thought of. So once again I'd implore you to actually say what you mean.
  - keybored 4y ago
    > Calling donating money to malaria nets or deworming "a skyscraper of assumptions built with paper" is like calling modern physics or modern cancer treatment "a skyscraper of assumptions built with paper". Physics manages simple models and the the efficacy of cancer treatment is (I guess) pretty measurable. You can’t compare these things to the Nerds in the West → Africa pipeline because of the mountains of variables that are layered on top. Just the sheer amount of humans that have to be involved, for one. I’m glad that we were able to figure out that those “fundamental assumptions” were not, in fact, the bare fundamental assumptions. They were just the fundamental assumptions that EA proponents thought that they needed because they thought that things like Utilitarianism, empiricism, etc. just followed immediately from them. Hence how you are trying to logic nerd my simple, primitive, chimp-like, intuitive list of assumptions and get me to either write whole thesis of axioms or to concede that Utilitarianism+empiricism etc. are the one true tools to rule all. But I am unfortanetely too low-IQ and not committed enough to some silly nerd morality hobby to write a whole thesis in order to try to convince people that trying to prevent the AI Apocalypse should not be the topmost concern of people’s precious 10%.

keybored 4y ago

> Then the followup question is "What is being effective to you? See above. > Based on your second assumption, it appears that being effective to you means, at worst, only helping those close to you, and at best, helping in ways that are visible to you. Do you think those in your community are the most in need of assistance, and even if they are, do you think that your view of what they need is more insightful than their opinion of what they need?" It’s not (under those assumptions) about helping those close to me per se. It’s about doing good with a high probability. I know what they need so I am more likely to actually help them, instead of causing unintentional harm. It’s funny that this author chose the tower analogy. EA is like a skyscraper of assumptions, built with paper: send money through three continents in order to help Africans in a very specific way without necessarily knowing anything about their specific needs. Maybe the malaria nets will help save millions. Or maybe they will just end up polluting the waterways. > The way you've defined things, asking a homeless person for what they believe would most help them is not effective, because it is second hand. If I ask that person then I get first-hand experience with that person.

joshuamorton 4y ago
> EA is like a skyscraper of assumptions, built with paper: send money through three continents in order to help Africans in a very specific way without necessarily knowing anything about their specific needs. Maybe the malaria nets will help save millions. Or maybe they will just end up polluting the waterways. I don't follow this line of reasoning. It's basically anti-empiricism which can be used to support pretty much any ideology you want. There's evidence that the malaria nets help save lives. We can see that fewer people get malaria. We can see death-rates declining. You can argue that that doesn't actually improve life/quality of life. You can argue that there are better ways, but the claim that "maybe they will just end up polluting the waterways" is based, fundamentally, on an intentional lack of knowledge. Calling donating money to malaria nets or deworming "a skyscraper of assumptions built with paper" is like calling modern physics or modern cancer treatment "a skyscraper of assumptions built with paper". Sure, yeah, it's complex and you don't understand all of it, and neither do I, but we have very clear and obvious evidence that it works! > If I ask that person then I get first-hand experience with that person. No you haven't, you've been told something through their biases. It's second hand experience, which you claim is unreliable. They could be lying to you! If what matters is your direct experience and observation, a study on what is most beneficial to homeless people, and a recorded interview with a homeless person, and a direct in person conversation with a homeless person (or a study on malaria nets, or a recorded interview with someone whose describes how their town has far fewer malaria related deaths than a decade ago) are all equally bad and less reliable than your preconceived beliefs and in-the-wild observations of what people struggle with. > See above. You don't actually define "effective" anywhere. I'm asking you to do that. You say it's a foundational assumption, and that sticking with things you know well helps you be effective, but we still don't know what effective means, and certainly not what it means to you, hence my request, which your above post doesn't address. Consider at least two possibilities: 1. To you, "effective" means provably done. That is, if you do not have incontrovertible (say, directly observable evidence) that your work has done good, you assume the value of your investment is zero. 2. To you, effective means something akin to utilitarianly efficient, e.g. most utils per $. It sounds to me like you initially claimed (2) and had a number of higher level assumptions about low trust, but it appears now you're claiming (1), and perhaps there are other possibilities that I haven't thought of. So once again I'd implore you to actually say what you mean.
- keybored 4y ago
  > Calling donating money to malaria nets or deworming "a skyscraper of assumptions built with paper" is like calling modern physics or modern cancer treatment "a skyscraper of assumptions built with paper". Physics manages simple models and the the efficacy of cancer treatment is (I guess) pretty measurable. You can’t compare these things to the Nerds in the West → Africa pipeline because of the mountains of variables that are layered on top. Just the sheer amount of humans that have to be involved, for one. I’m glad that we were able to figure out that those “fundamental assumptions” were not, in fact, the bare fundamental assumptions. They were just the fundamental assumptions that EA proponents thought that they needed because they thought that things like Utilitarianism, empiricism, etc. just followed immediately from them. Hence how you are trying to logic nerd my simple, primitive, chimp-like, intuitive list of assumptions and get me to either write whole thesis of axioms or to concede that Utilitarianism+empiricism etc. are the one true tools to rule all. But I am unfortanetely too low-IQ and not committed enough to some silly nerd morality hobby to write a whole thesis in order to try to convince people that trying to prevent the AI Apocalypse should not be the topmost concern of people’s precious 10%.