As far as I can tell the only reason they have been filterable at all was they did the courtesy of making the friendly username able to be greppable easily, and nothing stops them from randomizing that and flooding the network in the future. 

 I’ve been dealing with him all day. The problem is that you only notice the ReplyGuy (who is fairly harmless), while he has a bot partner that writes random texts in different languages, and then ReplyGuy responds to him. Without increasing the load on the server, it's not that easy to block. 

 Mandatory nip05 to talk is needed  

 please don't reintroduce DNS as a requirement back into the network lol

oh wait it already is because relays use it 

 Can be used bag of words or TF-IDF, it’s not too resource-intensive, but it’s not a universal solution. 

 This second bot writes nonsense, almost like a Japanese person. If its texts didn’t repeat and if it only wrote in Japanese, it would be absolutely indistinguishable from a typical Japanese user on Nostr. Something like:

ウンチしたけど、風が涙を運んでくれた。 妹に負けた。

Did a bot or a Japanese person write this? I wouldn’t be able to tell 😑 It’s really hard to deal with this.