|
|
|
|
|
|
by bickfordb
196 days ago
|
|
|
I'm not an expert in this area but from what I understand what was once novel content spam filtering is not at all novel now, there are easily trainable model strategies (BERT?) that get you to 99%. A whitelist, auth/contact is ideal for messaging without spam and is more workable with a large federated group that can adopt an evolving open source protocol. |
|
|