Ignoring e-mail content and throwing Naive-Bayes on the header alone is pretty much hove we got amazing spam filters about 15 years ago. All of course using a m

Ignoring e-mail content and throwing Naive-Bayes on the header alone is pretty much hove we got amazing spam filters about 15 years ago. All of course using a millionth or less of the resources a large language model would use.