I was reading the Singularity Institutes Q1 2008 Update, and noticed the following proud news item:
[Eliezer Yudkowsky] has written 300,000 words as of March on the Overcoming Bias blog. Select material is intended for a popular book; the rest can be used by SIAI for years to come.
When I read this, I couldn’t help chuckling to myself. For those of you who aren’t familiar with the blog, here is how it describes itself.
Over the last several decades, new research has changed science’s picture of how we succeed or fail to seek the truth. The heuristics and biases program, in cognitive psychology, has exposed dozens of major flaws in human reasoning. Microeconomics, through the power of statistics, has shown that many facets of society don’t work the way we thought.
Overcoming Bias aims to bring the implications home.
Personally, I find these topics fascinating. The implications of the heuristics and biases program have not permeated society nearly as much as is due, and economic theory still sorely needs more input from behavioral economics and the like to give us better models. Meanwhile, I am passionate about many of the other topics that are discussed on Overcoming Bias, like philosophy of mind and Bayesian statistics and prediction markets.
But despite it’s being on my feed reader, I hardly ever read full posts on that blog.
Contrast with Robert Hanson’s latest quip about prediction markets, which adheres much better to the blog’s own style standards:
Ideal posts are short, direct, have a clear thesis, and clear support such as a real-life example, a quote, an analysis, or a pointer to longer treatment.
The portability of text and the efficiency with which it can be transported on the internet make it incongruous to publish it in large chunks. On the contrary, in today’s world, Twitter makes sense. It takes a breed of arrogance to believe that others will want to spend more than two minutes discovering what you have to say.
So mere word count is a terrible measure of one’s contribution to the writing available on the web. That raises an interesting question though. What is a better measure? I’d speculate about an answer here, but this post is already too long.