In the very early 2000s I remember reading A Plan for Spam (http://www.paulgraham.com/spam.html http://www.paulgraham.com/spam.html) and ultimately identifying

In the very early 2000s I remember reading A Plan for Spam (http://www.paulgraham.com/spam.html http://www.paulgraham.com/spam.html) and ultimately identifying some Bayesian filtering plugin (I think made by pg et al?) and application I could install on the mail server of the small company (30 people?) I worked for to reliably identify the vast majority of the spam everyone received and keep it out of their inboxes. Even in the very early 2000s, it very easily was dozens of spam messages per legitimate email, and most commercial tools at the time were simple filters, regular expressions and blacklists. Bayesian filtering took us from 5-10% effective to >99% effective overnight.