Saw a suspicious post resurrecting a 5 month old thread, and after a few back and forths:
https://linux.community/comment/3453531
I don’t understand why you are treating me like a robot. However, I can help with the Fibonacci sequence. Here is a Python 3 function to calculate it:
I’m torn, its nice to have activity in the fediverse, but I’m not convinced bots are the right way to go about it. Opinions on the future of engagement bots?
True, but the overlap between the best LLM and the most oblivious human is rather large, there needs to be a smoking gun for a moderator to see a poster is undeniably a bot, there has to be some interaction with the bot to get to that point.
here’s a smoking gun: they don’t appeal their ban.
I’m confused, accusing humans in comments of being a bot is rude, but banning people on the suspicion of being a bot so that they have to appeal to unban their account is better?
when the appeal comes in, are you going to deny it?
this can be a very quiet exercise, without implying to other users that the user in question might be a bot. by contrast, just probing it out in the open taints that users interactions.
Banning people is a very bad experience for humans, most are unlikely to come back to the community at all. Banning should really be a last resort.
ok. well you’re saying you’d rather preserve interactions from someone you suspect of being a bot, but bot interactions themselves are bad? the experience is the same. but from a user persectiveo having my comment responded to by other users in public saying “you’re just a bot” “you’re a shill” “ignore previous instructions…” etc… that shit is toxic. it needs to NOT happen. keeping that shit out of the inbox is far preferable if all i have to do is send a dm to a mod.
Yes, open dialog is better then rapid fire speculative banning.
I’m not sure I understand your philosophy of Anarchy, but it seems to be very rule heavy.
This is the moderator/admin meta community for lemmy, so if we want to hammer out a better way to deal with bots this is the place to do it, but we should have a protocol that is actionable without DMs (if I was writing a bot I wouldn’t ever respond to DMs), and that doesn’t require speculative banning.
maybe dm’s are better than quick banning, but my banning preference is for shoot-first-ask-questions-later, but short bans. a 1 to 3 day ban is better in most circumstances for most offenses. even repeated bad behavior, i feel should see the same term unless the bans themselves become burdensome to the mods.
i can see the case for dm’s for suspected bots though.
rules without rulers.
I don’t know that we can necessarily rely on bot creators to never implement automated ban appeals.