ciphergoth comments on Mandatory Secret Identities - Less Wrong

28 Post author: Eliezer_Yudkowsky 08 April 2009 06:10PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (177)

You are viewing a single comment's thread. Show more comments above.

Comment author: Eliezer_Yudkowsky 08 April 2009 11:45:22PM 1 point [-]

Hm. Arguably I should only be worried about fast dilution rather than slow dilution. But I'm also worried that the community grows slower if it's inward-looking, and hope for faster growth if it's involved with the outside world.

Is it possible that having taught yourself you haven't so directly experienced that there's not necessarily a correlation between a persons understanding of a subject and their ability to teach it?

Entirely possible. But I'm not sure I have so much faith in the system you describe, either. The most powerful textbooks and papers from which I get my oomph are usually not by people who are solely teachers - though I haven't been on the lookout for exceptions, and I should be.

Comment author: ciphergoth 09 April 2009 09:01:08AM 2 points [-]

Er, I thought the difference between religious and scientific teachings was that scientific teachings didn't have to worry about dilution? It seems like you put a high probability on this community disappearing into a death spiral of some sort without you - I would have thought we should worry more that we're already in one which we haven't picked up on.

Comment author: Eliezer_Yudkowsky 09 April 2009 02:31:33PM 0 points [-]

I thought the difference between religious and scientific teachings was that scientific teachings didn't have to worry about dilution?

More of a difference between things that are hard vs. easy to teach and measure. Businesses have the same problem with a great CEO trying to hire great employees, dilution of corporate culture, etc. - they have highly quantifiable output at the end of the day, but in the middle of the day and the middle steps of the process, it's not as easy to measure.

I anticipate a beginning period extending for at least several years when we don't have good metrics because we're still trying to develop them.