I have to commend Sam Harris for featuring a guest post by security expert Bruce Schneier on profiling, especially since he backs up everything I’ve said so far. 1) Profiling people who “look Muslim” will have a high false positive rate, 2) “looking Muslim” is a hopelessly indefinable criterion, 3) terrorists will use profiling to avoid detection, and 4) it’s a strategy to alienate those who could be on our side.
I do recommend his blog, Schneier on Security, for your regular reading (also because he has Friday Squid Blogging — he’s a man of excellent taste).
I’m also going to recommend this paper that Brownian mentioned in the comments here: Strong profiling is not mathematically optimal for discovering rare malfeasors. Profiling doesn’t add up; the numbers don’t work.
Are we done now?