Global Feed Post Login
Replying to Avatar Quint

GM!

I've been thinking about the bot issue. It only takes one follower in your WoT who follows a bot (and probably doesn’t realise it) to ruin it. Asking for proof of life when you follow someone is impractical to say the least.

I guess the main way to identify a bot is through their actions — spamming accounts or the same account (aka ReplyGuy). So what if there was an option for auto-unfollowing an npub if they sent the same message to more than [n] accounts or [n] times to the same account? Maybe an option to choose how much of the spam message needs to be identical to qualify.

I realise this doesn’t stop bots who generate random gibberish notes to each spamee (if that’s a word) but it could make a difference. It could also remove a bunch of follows if the tolerance is set too low so maybe a warning would be good if it went below a certain level.

I’m just spitballing. I also have no idea if it’s even possible. What do people think? Could this help?

#asknostr #nostr #spam #bots #grownostr

4a
TheOneWithAReallyLongName 1y ago

The immediate solution to #ReplyGuy and #ReplyGal is client side filtering, though that may not be as universal as some would like. The soon but not quite yet solution is implementation and adoption of NIP-13 for #ProofOfWork. Set the #PoW difficulty high enough to make bulk spamming expensive but low enough that users shouldn't have too much difficulty.

Reply to this note

Please Login to reply.

Discussion

No replies yet.