I get the strong feeling I've read this article before, but I can't find a strong indicator this is an old article. Did someone else write about the leaked script?
You may've read the much shorter original version back when it lived in http://www.gwern.net/Notes or read it sometime in the past 3-4 months when I was working on expanding it.
(The first version was basically the non-quantitive first third, but about half that size, so I added a ton.)
It's a bit hard to find, but in the left sidebar there's a date string "2 Nov 2009". This would match up to the article's timeline; script appears on the Internet in May, add a few months for the author to discover it and perform extensive research. Inspecting the HTML verifies that that date is probably when the article was first published:
<meta name="dc.date.issued" content="2 Nov 2009"/>
This particular post is a new version of an older article, not a completely new article, so it seems mostly likely that you read the old version already.