What’s the best way to filter out long form spam from an article reader? I don’t want to use wot because it’s supposed to be a discovery service but there’s a lot of garbage spam. I could hardcode it but that feels wrong and might false flag good content

Reply to this note

Please Login to reply.

Discussion

WoT is the base for discovery. Example at npub.world

Any good documentation to feed a hungry bot?

You are welcome: https://vertexlab.io/docs/nips/

Vertex is a premium service for WoT, but if you are interested I can give you some free credits. Just post here the npub of the bot.

Ah. I can’t pay for anything additional atm . My hobby expenses are already high 🤣

What is it was hard coded and flagged, but you put an AI agent (man) in the middle.

Taught the agent or human how to differentiate spam from real and if there were still questions it was to a committee of agents or a real human to approve or deny and further train the agent?

I need a simple solution. I’m not a superdev

Man in the middle is simplest.

Use a human if you don’t want to automate it via ai agents

You could test the theory / workflow using something like Monkey Learn where you train a bot what to look for.

But either way; MIM is simplest especially if you don’t want to code while you eat dinner 🤪

Like from highlighter.com?

Not sure what you mean

Maybe it’s my understanding of what an article reader is…or maybe you’re referring to one outside of the nostr scope.

I’m thinking something like highlighter.com which supports nostr login and article consumption?

It's easy. You must charge 1 Satoshi in order for someone to make a comment on your articles.

Bots won't pay so there will be no spam on your website or articles.

Cluster of npubs and their lists, select only short lists of max 50 npubs