Nostr Web Client

I only see two ways to stop this spam.

1) filter to see only posts from people in your trust graph

2) impose a real cost to post to all

Anything I’m missing?

Bottom line is that new users will be initially treated like spammers.

This attack is probably just a warm up. I would expect that we’ll soon see AI posts that you can’t pattern match, and overwhelming spam against both replies and DMs. I would expect them to initially target the folks with the largest follower count, then attacking new users.

Reply to this note

Please Login to reply.

Discussion

Leo Fernevak 1y ago

Right.

All effective spam prevention comes with tradeoffs.

Web of trust models should make it very hard for spam to pass. If an algorithm assigns weight to our social network nodes, filtering out spam would be relatively easy. The problem is designing a mechanism allowing newcomers into the web of trust that doesn't come with negative tradeoffs.

A cost-to-post is a possibility that can also help fund relays, yet will add an obstacle for new users and may reduce interaction depending on the costs.

We also have to consider that future censorship mechanisms will likely utilize anti-spam functionality as the primary methods of justification. It is therefore important to consider the possible censorship risks in a given anti-spam strategy.