It isn't exactly "revolutionary". Receiver-based approach has been employed for a while, although with a twist: analogous to methods of identifying DOS-attack

It isn't exactly "revolutionary". Receiver-based approach has been employed for a while, although with a twist: analogous to methods of identifying DOS-attack sources, large ISPs or web mail providers (gmail) establish a network of "honeypots" - thousands of fake mailboxes whose addresses are published on fake web pages, allowing spammers' crawlers to absorb those emails into their databases. Then you just tag IP addresses that attempt to deliver emails to honeypots. What you get is a simplest receiver-based system with just two groups of receivers: 100% spam-receivers and "others".