On: G Suite Horror Story

[derefr]

> filtering incoming spam/malware/philishing attempts

Is there not just an open-source pre-trained spam-filtering model (using whatever is the state of the art for these—guessing classical Bayesian filtering on word-tokens is no longer at all what Gmail is doing)? It’d be a nice complement to the open RBLs, no?

[ivan_ah → derefr]

There are, but they are not state of the art.

There is also the benefit of the spam filter service being "live updated," e.g. if a new type of spam message comes up it will get flagged since lots of users.