Comments: Meme and (tech.)

"All roads lead to Rome, all links lead to the A-list ..."

Welcome to the A-list. I came here from Memeorandum.
You're one of the elite. Congratulations.

Posted by Norm Schwartz at October 30, 2005 03:04 AM

Nice post, Seth. It's always a good sign when your first commenter misses the point.

But I wonder if your Central Pundit Theorem isn't merely a restatement of the Power Law. Of course, I don't even know if there is such a thing as the "Power Law," so maybe it isn't a restatement.

Maybe it's "cream always rises to the top."

And doesn't the cream love that?

Posted by dave rogers at October 30, 2005 08:05 AM

I think it completely depends on what your definition of "cream" and especially "top" is. If you're trying to get A-listers to take notice of you or link to you, or to show up in something like tech-memeorandom, then maybe "top" is going to be a problem.

But if a definition of "top" is to "find readers who find value in your blog", then yes, I believe the "cream" (value as perceived by the readers) most definitely rises to the top. Isn't that what Shelley's blog has done?

Posted by Kathy Sierra at October 30, 2005 06:17 PM

Dave: I think the Central Pundit Theorem definitely applies where there's a Power Law of attention, but there are other circumstances where it would be applicable too. So it's not really a restatement in that way.

Kathy: Unfortunately, that tends to be an unfalsifiably belief.

Posted by Seth Finkelstein at October 31, 2005 01:23 AM


I was - obviously, I hope - speaking with my tongue firmly in my cheek, suggesting how those who find themselves at the very top of the hierarchy might choose to regard their presence there.

I don't really know how to take your comment. It seems as though you're saying something to the effect that criticism of the shortcomings of hierarchy are somehow inappropriate because we can all find "value" in whatever portion of attention we can garner to ourselves. I don't think the two are mutually exclusive.

I think the top of the hierarchy is always an appropriate target for criticism because by virtue of their "authority" attendant to their rank, they tend to dominate the choices of topics for public discourse. And since many of them share much in common with one another, they don't often disagree with one another, the exception being political topics on political blogs.

I don't really get your objection to criticism, except that perhaps you don't like the "negativity." To some extent, it's a result of some frustration on the part of those of us who have been reiterating these issues for some time. I myself grow rather tired of the negativity, but I find it all too easy to summon up in response to the breathless hyperbole offered by would-be "visionaries." But that's just me.

If you don't like the criticism, by all means, feel free to criticize it. But don't take it personally when your criticisms are ignored as roundly as those we've been making. It's all part of the fractal nature of the "conversation." (Self-similar at all scales.)

Posted by dave rogers at October 31, 2005 05:10 AM

Seth, I'm reading Linked by a guy whose name I can't spell without looking at the title... wait, I can use technology! A little copy and paste... Linked by Albert-László Barabási, which is supposed to help explain the Power Law as a network effect.

I'd be interested in your thoughts on the other circumstances part of the Central Pundit Theorem. (Love that name, btw.)

Posted by dave rogers at October 31, 2005 05:15 AM

Dave, what I meant was that I was trying to apply the ideas from statistics that given many types of distributions, just a small sampling (which could be described as bottom-up construction) is sufficient to approximately construct the whole distribution, as well as the idea that many independent samples can converge to a distribution. This actually is a different property than power laws. Maybe I'm being too pedantic, in that I was merely trying to convey I didn't view the Central Pundit Theorem as about *exponential* distribution, but that any "top-heavy" system, exponential or not, was going to have general sampling basically point to the top part of the curve again. The top pundits will tend to be the "modal" (most common) results.

The idea that often small samples aren't very different from enormous datasets is sometimes very surprising to people, it can go against some wrong intuitive notions.

Posted by Seth Finkelstein at November 1, 2005 02:31 AM