Title and excerpt aside, this isn't really a retraction of the actual test, just of the author's overly-aggressive interpretation thereof. Basically he says that the test does have some predictive power but not enough to prove anything about who can or can't pass a programming class.