The Bayesian spam filtering set-up I’ve got (thanks Scott) is working pretty well. In slightly under a month and a half, I’ve gotten more than 2000 spam e-mails. There are still a few trickling in to my inbox, but not too many and there are now very few false-positives. More and more I’m seeing the ineffective ‘hey, thought you might like to see this’ spam that was predicted in A Plan for Spam.