I enjoyed this post and this series overall. However, I would have liked more elaboration on the section about EA’s objectionable epistemic features. Only one of the links in this section refer to EA specifically; the others warn about risks from group deliberation more generally.
And the one link that did specifically address the EA community wasn’t persuasive. It made many unsupported assertions. And I think it’s overconfident about the credibility of the literature on collective intelligence, which IMO has significantproblems.
Thanks for your question, Nathan. We were making programmatic remarks and there’s obviously a lot to be said to defend those claims in any detail. Moreover, we don’t mean to endorse every claim in any of the articles we linked. However, we do think that the worries we mentioned are reasonable ones to have; lots of EAs can probably think of their own examples of people engaging in motivated reasoning or being wary about what evidence they share for social reasons. So, we hope that’s enough to motivate the general thought that we should take uncertainty seriously in our modeling and deliberations.
I enjoyed this post and this series overall. However, I would have liked more elaboration on the section about EA’s objectionable epistemic features. Only one of the links in this section refer to EA specifically; the others warn about risks from group deliberation more generally.
And the one link that did specifically address the EA community wasn’t persuasive. It made many unsupported assertions. And I think it’s overconfident about the credibility of the literature on collective intelligence, which IMO has significant problems.
Thanks for your question, Nathan. We were making programmatic remarks and there’s obviously a lot to be said to defend those claims in any detail. Moreover, we don’t mean to endorse every claim in any of the articles we linked. However, we do think that the worries we mentioned are reasonable ones to have; lots of EAs can probably think of their own examples of people engaging in motivated reasoning or being wary about what evidence they share for social reasons. So, we hope that’s enough to motivate the general thought that we should take uncertainty seriously in our modeling and deliberations.