5 ms·

This is all a strawman. When people (me included) criticize EA for it's questionable priorities, it's an expression of exasperation, not an existential attack

by ajross 4y ago

This is all a strawman. When people (me included) criticize EA for it's questionable priorities, it's an expression of exasperation, not an existential attack on the movement or its practitioners. We like EA. We think it's a good idea. We also think it, on the whole, has made some outrageously bad analysis about what it should work on, and is perpetutating that nonsense in an echo chamber of sci-fi speculation. And that really sucks, because we see all that energy that could be spent on climate issues or public health or democratization instead aimed at predicting when WOPR or SkyNet will achieve sentience. It's... frustrating.

not-my-account 4y ago
I think that EA is a lot more decentralized than you think it is. You could totally be an effective altruist devoted to climate issues, public health, and democratization. Just pop down (or move sideways from) the pillar that states that AI Safety is paramount and chug away!
- joshuamorton 4y ago
  I think the issue is that while the majority of funding through EA and EA-adjacent orgs still goes to poverty-reduction work, the majority of mindshare and interest (among EA!) is in primarily AI- but also pandemic-risk. The first of these especially avoids addressing many real, pressing, ML/AI problems in the world today for philosophizing (and getting paid to do it by people who were previously donating their money to global poverty reduction). If you consider EA Global as a charity, and the money donated to longtermist AI to be, functionally, salaries paid to employees, the overhead of EA jumps by like 15%. That's not admirable, by EAs own metrics.
  - not-my-account 4y ago
    Sure, but these people and others truly believe that AI safety is an important issue. There are a lot of really smart people who have thought a lot about it, and I have not looked deeply into it so I can't say, but it doesn't seem like that would invalidate EA entirely. I.E. I still don't understand why you can't just dislike the fact that some people are devoting time, energy, and money to AI safety, and just be an effective altruist for the issues that you think are most important. In other words, don't throw the EA baby out with the AI Safety bathwater.
    - joshuamorton 4y ago
      > In other words, don't throw the EA baby out with the AI Safety bathwater. I don't! I donate (lots of) money to GiveWell, and I credit EA members for convincing me to do that. The parts of EA that are functionally "hey donate money to effective charities" are good things that make the world better. I do think that it is dangerous to EA as a movement to be discolored by self-dealing AI longtermism.
      - not-my-account 4y ago
        Ah! Gotcha, I misunderstood your position. Sorry for the misunderstanding! I think we are in more-or-less the same / similar position.
lrobinovitch 4y ago
> When people (me included) criticize EA for it's questionable priorities, it's an expression of exasperation, not an existential attack on the movement or its practitioners. I have had different experiences with critics of EA. Many critics come off as passive, finding one or two basic disagreements with institutions or moonshot/futurist ideas that they use to justify their total disengagement with their potential positive impact in the world beyond themselves and their friends/families. You might not even call these folks critics - maybe they're just "excuse-seekers"? It sounds like you are a critic who is actively engaged at the second and third blocks of the tower, which I assume the author would have no issue with.
notahacker 4y ago
This. And the analysis (good and bad) is the movement. It's great that they're encouraging people to donate money, and have goals in doing so, but that idea's been around since ancient times. (And they're far the first people to apply utilitarian frameworks or RCTs to philanthropy either, so it's really the detail that's novel) And since the movement set itself up with the goal of debating the effectiveness of altruistic interventions (including explicit suggestions that others weren't doing altruism effectively), it seems ludicrously antithetical to that objective to shield it from criticism by telling us to shut up about the bits we think are ineffective and go away and give 10% of our own income instead. (Apart from anything else, 10% of the average sceptic's lifetime income probably isn't worth nearly as much in QALY as a throwaway comment convincing a single SV billionaire with exit money to spend on the future of humanity that they might be better off listening to boring old aid agencies that don't even do RCTs, rather than being seduced by arguments in favour of donating to a more EA-approved cause like the nonprofit marketing wing of OpenAI Inc)
- zozbot234 4y ago
  AI research is not "more EA-approved" than global development - any such claim is peddling gross misinformation. Most EA's donate a lot to global development, healthcare research, animal rights, disaster preparedness and other very mainstream causes.
  - notahacker 4y ago
    > AI research is not "more EA-approved" than global development - any such claim is peddling gross misinformation I haven't made that claim. I have accurately stated that EA organisations have both explicitly criticised many/most aid agencies for not providing enough evidence of efficacy in the form of RCTs etc and explicitly endorsed and funded OpenAI. In my opinion, the latter cause is a less effective and in-need-of-funds one than the average "do not recommend" charity on GiveWell, and so I think human QALYs would benefit from high net worth individuals regarding the core of the EA movement's opinions on effectiveness as highly debatable and listening to voices outside it too. (And obviously whilst many EAs recommend and contribute significantly to global development causes, it's also true that this cause is extremely highly prioritised outside the movement whereas AI risk and "longtermism" - an increasingly emphasised if not universally approved part of the movement - is pretty much exclusive to it, so I think the net direction of most people already committed to human-life centred philanthropy paying attention to the EA movement is going to be to fund the former less and the latter more)

by ajross 4y ago

not-my-account 4y ago

I think that EA is a lot more decentralized than you think it is. You could totally be an effective altruist devoted to climate issues, public health, and democratization. Just pop down (or move sideways from) the pillar that states that AI Safety is paramount and chug away!

joshuamorton 4y ago
I think the issue is that while the majority of funding through EA and EA-adjacent orgs still goes to poverty-reduction work, the majority of mindshare and interest (among EA!) is in primarily AI- but also pandemic-risk. The first of these especially avoids addressing many real, pressing, ML/AI problems in the world today for philosophizing (and getting paid to do it by people who were previously donating their money to global poverty reduction). If you consider EA Global as a charity, and the money donated to longtermist AI to be, functionally, salaries paid to employees, the overhead of EA jumps by like 15%. That's not admirable, by EAs own metrics.
- not-my-account 4y ago
  Sure, but these people and others truly believe that AI safety is an important issue. There are a lot of really smart people who have thought a lot about it, and I have not looked deeply into it so I can't say, but it doesn't seem like that would invalidate EA entirely. I.E. I still don't understand why you can't just dislike the fact that some people are devoting time, energy, and money to AI safety, and just be an effective altruist for the issues that you think are most important. In other words, don't throw the EA baby out with the AI Safety bathwater.
  - joshuamorton 4y ago
    > In other words, don't throw the EA baby out with the AI Safety bathwater. I don't! I donate (lots of) money to GiveWell, and I credit EA members for convincing me to do that. The parts of EA that are functionally "hey donate money to effective charities" are good things that make the world better. I do think that it is dangerous to EA as a movement to be discolored by self-dealing AI longtermism.
    - not-my-account 4y ago
      Ah! Gotcha, I misunderstood your position. Sorry for the misunderstanding! I think we are in more-or-less the same / similar position.

lrobinovitch 4y ago

> When people (me included) criticize EA for it's questionable priorities, it's an expression of exasperation, not an existential attack on the movement or its practitioners. I have had different experiences with critics of EA. Many critics come off as passive, finding one or two basic disagreements with institutions or moonshot/futurist ideas that they use to justify their total disengagement with their potential positive impact in the world beyond themselves and their friends/families. You might not even call these folks critics - maybe they're just "excuse-seekers"? It sounds like you are a critic who is actively engaged at the second and third blocks of the tower, which I assume the author would have no issue with.

notahacker 4y ago

This. And the analysis (good and bad) is the movement. It's great that they're encouraging people to donate money, and have goals in doing so, but that idea's been around since ancient times. (And they're far the first people to apply utilitarian frameworks or RCTs to philanthropy either, so it's really the detail that's novel) And since the movement set itself up with the goal of debating the effectiveness of altruistic interventions (including explicit suggestions that others weren't doing altruism effectively), it seems ludicrously antithetical to that objective to shield it from criticism by telling us to shut up about the bits we think are ineffective and go away and give 10% of our own income instead. (Apart from anything else, 10% of the average sceptic's lifetime income probably isn't worth nearly as much in QALY as a throwaway comment convincing a single SV billionaire with exit money to spend on the future of humanity that they might be better off listening to boring old aid agencies that don't even do RCTs, rather than being seduced by arguments in favour of donating to a more EA-approved cause like the nonprofit marketing wing of OpenAI Inc)

zozbot234 4y ago
AI research is not "more EA-approved" than global development - any such claim is peddling gross misinformation. Most EA's donate a lot to global development, healthcare research, animal rights, disaster preparedness and other very mainstream causes.
- notahacker 4y ago
  > AI research is not "more EA-approved" than global development - any such claim is peddling gross misinformation I haven't made that claim. I have accurately stated that EA organisations have both explicitly criticised many/most aid agencies for not providing enough evidence of efficacy in the form of RCTs etc and explicitly endorsed and funded OpenAI. In my opinion, the latter cause is a less effective and in-need-of-funds one than the average "do not recommend" charity on GiveWell, and so I think human QALYs would benefit from high net worth individuals regarding the core of the EA movement's opinions on effectiveness as highly debatable and listening to voices outside it too. (And obviously whilst many EAs recommend and contribute significantly to global development causes, it's also true that this cause is extremely highly prioritised outside the movement whereas AI risk and "longtermism" - an increasingly emphasised if not universally approved part of the movement - is pretty much exclusive to it, so I think the net direction of most people already committed to human-life centred philanthropy paying attention to the EA movement is going to be to fund the former less and the latter more)