I'm building a public Autospam detection service that will be compatible with Pixelfed, Mastodon and other projects!
I'm leveraging my expertise* in Autospam, along with my 3 year old dataset to build the best fediverse Akismet alternative.
Are you interested? Boosts appreciated!
@dansup This sounds good and reminds me of the early days of fighting email spam with Spamassasin.
@dansup I don't know if this is relevant, but as one of the moderators on IOCX I see an immense amount of SEO spam signups. We ban them on signup, and thus very very few such posts get onto the local timeline. I'm certain the exact same account is created on multiple Mastodon-servers, so we should really have a sort of "two instances have banned this very same account so block it"-functionality.
@dansup @thisismissem related work?
@dansup very interesting! awesome! something we should have had from day one, since we should have learned something from the email days, but it's here now at least. You say it'll work with pixelfed and mastodon, that's awesome! could it work similarly with go to social?
Now, imagine such a service, which doesn't only mark something as spam, but can be used to mark a message as different things, one of them would be spam, but also politics, racism, or even make lists of good stuff, doesn't have to be all negative, then users could decide what to do with something labeled, for example, politics. Crucially, this is on a per-user basis, and import/export can only be done between users, but you know, would be interesting to think about that kind of system too, now that this exists in a fediverse context.
@dansup got some yesterday
@dansup That sounds pretty cool. I've often wondered whether I could pipe mastodon posts into rspamd to classify them, leveraging existing email filters. It seems like it'd be the same basic idea.
@dansup nice! is this on the pixelfed github somewhere?
@dansup You are a freaking rock star coding god!
