Also, I think this “Well-kept gardens die by pacifism” post might be kind of a good illustration of a problem I have with how the sequences are regarded in general. The epistemological quality of this post seems pretty poor: although it’s discussing phenomena that are best studied empirically (as opposed to phenomena that are best studied theoretically, like math), it cites no studies, and doesn’t make an attempt to become a proto-study itself by trying to, say, find a method to do quasi-random sampling of online communities and figure out whether each community constitutes evidence for or against its thesis. Instead, its argument is based mainly on personal experience (even concrete examples in the form of actual specific anecdotes, like 4chan say, are few).
This phenomena could also have been productively studied theoretically: say, by making references to the expected quality of any given post, thinking about how frequently users are likely to return to a given forum and important it is for them to see new/valuable content each time they return, etc. But EY makes no attempt to do that either. (At least MBlume starts to think about modeling things in a more mathematical fashion.)
And yet it’s a featured post voted up 91 points… as far as I can tell, largely on the strength of the author’s charisma. I’m glad this post was written. I found it valuable to read; it has a couple novel arguments and insights. But it seems suboptimal when people cite it as if it was the last word on the question it addresses. And it seems weird that mostly on the strength of that post’s advice, posts as epistemologically weak as it are no longer being written as often on LW any more. Tossing around novel perspectives can be really valuable even if they aren’t supported by strong theoretical or empirical evidence yet.
Also, I think this “Well-kept gardens die by pacifism” post might be kind of a good illustration of a problem I have with how the sequences are regarded in general. The epistemological quality of this post seems pretty poor: although it’s discussing phenomena that are best studied empirically (as opposed to phenomena that are best studied theoretically, like math), it cites no studies, and doesn’t make an attempt to become a proto-study itself by trying to, say, find a method to do quasi-random sampling of online communities and figure out whether each community constitutes evidence for or against its thesis. Instead, its argument is based mainly on personal experience (even concrete examples in the form of actual specific anecdotes, like 4chan say, are few).
This phenomena could also have been productively studied theoretically: say, by making references to the expected quality of any given post, thinking about how frequently users are likely to return to a given forum and important it is for them to see new/valuable content each time they return, etc. But EY makes no attempt to do that either. (At least MBlume starts to think about modeling things in a more mathematical fashion.)
And yet it’s a featured post voted up 91 points… as far as I can tell, largely on the strength of the author’s charisma. I’m glad this post was written. I found it valuable to read; it has a couple novel arguments and insights. But it seems suboptimal when people cite it as if it was the last word on the question it addresses. And it seems weird that mostly on the strength of that post’s advice, posts as epistemologically weak as it are no longer being written as often on LW any more. Tossing around novel perspectives can be really valuable even if they aren’t supported by strong theoretical or empirical evidence yet.