I don’t know about that—the term “expert” corresponds to a reasonably coherent group of people, at least enough so that if the post’s claim were really true (i.e., that “experts” are a less good source of information/advice than a smart layperson reading academic journals), it would be very surprising and interesting.
Of course, what really isn’t an interesting reference class is the complement of experts: non-experts. If the experts are no good, we need better experts, not to look to non-experts. (In LW-speak: reversed stupidity isn’t intelligence.)
I disagree. Anyone can join or exile themselves from the group at any time. Far too nebulous to be able to talk about trusting them as a group besides when personally defined.
“experts” is not an interesting reference class.
I don’t know about that—the term “expert” corresponds to a reasonably coherent group of people, at least enough so that if the post’s claim were really true (i.e., that “experts” are a less good source of information/advice than a smart layperson reading academic journals), it would be very surprising and interesting.
Of course, what really isn’t an interesting reference class is the complement of experts: non-experts. If the experts are no good, we need better experts, not to look to non-experts. (In LW-speak: reversed stupidity isn’t intelligence.)
I disagree. Anyone can join or exile themselves from the group at any time. Far too nebulous to be able to talk about trusting them as a group besides when personally defined.
This objection collapses if you don’t think about it as a binary trusting-condition but instead as incremental changes in priors.