Indeed. Ultimately we're still relying on blacklists, Bayesian filters, and hueristics to detect actual SPAM.