Fair enough. Could you edit your crossposts to the new version of the post that has some clarifications just so there's no misscommunication about this anywhere (you could put the original under a lemmy spoiler block which turns it into a dropdown for people that want to see original wording, details on how to do that here https://join-lemmy.org/docs/users/02-media.html.Or some other way you want)

For context this is for programming.dev specifially and affects our users rather than something on any fediverse software (figured I would mention that since it was cross posted into the main fediverse com with no context)

Reposting my answer to a comment on the main p.d post just so theres no confusion on this

• Processing is local. Were not sending data to an LLM but the method is hidden so bad actors cant bypass it
• The things that are checked are public (posts, comments, and dms that are also sent to other servers we dont control) and can be moderated on regardless of this. There isnt anything that breaches privacy
• Worst thing it can automatically do is temporarily remove a post/comment until an admin can approve or deny it just so its not there for people to run into (for example if someone spams slurs at someone else the someone else doesnt see that instead of it being there until an admin is able to come online)

yeah pretty much exactly that. It's custom built 100% by our team with no external tools or anything (just a way to notify us using apis). it's technically a separate running code since we can't hook it into the base lemmy code that easily but acts the same way

Theres no LLM involved. This version of the post doesnt have the tldr to clarify some things so ill just repost my comment here

• Were not sending data to an LLM but the method is hidden so bad actors cant bypass it (you can disagree on that but the main issue is bots that are able to check blacklisted content and then just edit messages slightly to bypass it)
• The things that are checked are public (posts, comments, and dms that are also sent to other servers we dont control) and can be moderated on regardless of this. There isnt anything that breaches privacy
• Worst thing it can automatically do is temporarily remove a post/comment until an admin can approve or deny it just so its not there for people to run into (for example if someone spams slurs at someone else the someone else doesnt see that instead of it being there until an admin is able to come online)

I do think having options like letting people disable dms, letting people choose to need to approve dms to get notifications about them and then having a whitelist makes sense. I can make tickets for those in the lemmy repo later if they don't exist already

for a couple of those

The things that are checked are public (posts, comments, and dms that are also sent to other servers we dont control) and can be moderated on regardless of this. There isnt anything that breaches privacy

The way its being moderated isnt disclosed so that people cant just bypass it. The processing is local though like said in the original message (so for example were not sending this to some LLM/Ai to then use)

Everything is sent for human review so false positives can be reversed quickly and the worst action that happens is removal of content temporarily (before thats reversed)

This just means someone cant spin up a new account and spam things on the site or on federated sites and have that linger until its able to be moderated

If youre willing to send me the csv I'm interested in seeing the results with a more eurovision style scoring rather than just average rank

if you mention a lemmy community in your post the post gets sent to show up in the lemmy community. Comments are automatically federated if you comment on a lemmy post.

lemmy posts are all federated to mastodon though if the servers know about each other

comment was restored instantly since it shouldn't have hit that message (ill probably try to get piss removed from the check). Just a way for admins to be able to check messages to prevent new accounts from instantly harassing people after sign up