I consider you a smart guy -- but when I wrote a couple of front page posts about Bayesian statistics, you made a few comments that revealed notable Dunning-Kruger effect with respect to the topic. Since that time, my observations of your interactions with other subject matter experts about topics in their domains of expertise have only reinforced this impression.

My current understanding of you is that you are a smart guy, but posting on LW is often unrewarding for you because (i) your high intelligence is a core part of your self-image, and (ii) you're not as smart as you think you are. I hope this info can be of use to you; apologies for any narcissistic injury caused by this comment;.

Assuming that I disagreed with you re. Bayesian statistics, our positions are symmetric--you believe I am overconfident, and I believe you are over-confident. I have a degree in math and have taught basic Bayesian statistics at a university, and used Bayesian statistics successfully to get correct results in computer programs many times, so I have some reason for my confidence. Have you made use of this information in re-evaluating your own confidence?

12Anatoly_Vorobey6yOut of curiosity, did you consider sending this comment via PM, and if so, what made you decide to post it publicly?
8Kaj_Sotala6yIf you're going to make a comment like this, you really have to provide specific examples.

Using vs. evaluating (or, Why I don't come around here no more)

by PhilGoetz 1 min read20th Jan 201438 comments


[Summary: Trying to use new ideas is more productive than trying to evaluate them.]

I haven't posted to LessWrong in a long time. I have a fan-fiction blog where I post theories about writing and literature. Topics don't overlap at all between the two websites (so far), but I prioritize posting there much higher than posting here, because responses seem more productive there.

The key difference, I think, is that people who read posts on LessWrong ask whether they're "true" or "false", while the writers who read my posts on writing want to write. If I say something that doesn't ring true to one of them, he's likely to say, "I don't think that's quite right; try changing X to Y," or, "When I'm in that situation, I find Z more helpful", or, "That doesn't cover all the cases, but if we expand your idea in this way..."

Whereas on LessWrong a more typical response would be, "Aha, I've found a case for which your step 7 fails! GOTCHA!"

It's always clear from the context of a writing blog why a piece of information might be useful. It often isn't clear how a LessWrong post might be useful. You could blame the author for not providing you with that context. Or, you could be pro-active and provide that context yourself, by thinking as you read a post about how it fits into the bigger framework of questions about rationality, utility, philosophy, ethics, and the future, and thinking about what questions and goals you have that it might be relevant to.