Less Meta

Eliezer Yudkowsky26 Apr 2009 5:38 UTC
17 points
21 comments1 min readLW linkArchive

My recent sequence on the craft and the community is highly forward-looking—not an immediately whole recipe, but a list of action items and warnings for anyone setting out in the future. Having expended this much effort already, it seems worthwhile to try and leverage others’ future efforts.

That sequence seemed like an appropriate finale, but putting it last had some side effects that I didn’t expect. Thanks to the recency effect, people are now talking as if the entire œuvre had all been anticipation of future awesomeness, with no practical value in the present...

Okay, seriously, if you look over my posts on Overcoming Bias that are not just a couple of months old, you really should see quite a lot of practical day-to-day stuff. Yes, there’s a long sequence on quantum mechanics, but there really is plenty of day-to-day stuff, right down to applying biases of evaluability to save money on holiday shopping. (Not to mention that I finally did derive real-world advice out of the QM detour!)

I suspect there may also be a problem here with present schools not teaching people the experience of creating new craft. What they present you with, is what you learn. So when I talk about my belief that we could be doing better, they look around and say: “But we aren’t doing that well!” rather than “Hm, how can we make progress on this?”

And then your current accomplishments start to pale in the light of grander dreams, etcetera. One of the great lessons of Artificial Intelligence is that no matter how much progress you make, it can always be made to appear slow and unexciting just by having someone else quoted in the newspapers about much grander promises on which they fail to deliver.

Anyway. I think that discussion here is going a bit too meta, too much about the community, and that was not the final push I had planned to deliver. (I know, I know, obvious in retrospect, yes, but I still didn’t see it coming.) Hence the quick add-on about practical advice backed by experimental results or deep theories—thankfully we are going back to those again in recent posts.

So if it’s all right with you, dear readers, after the end of April, I will not promote more than one meta post unless at least four nonmeta posts have appeared before it—does that sound fair?

Part of the sequence The Craft and the Community

Next post: “Go Forth and Create the Art!”

Previous post: “Practical Advice Backed By Deep Theories”

Eliezer Yudkowsky26 Apr 2009 5:38 UTC
17 points
21 comments1 min readLW linkArchive