Oddbean new post about | logout
▲ ▼
Michael J | 10 months ago (raw) | root | parent | reply | flag +2 
 To prevent gaming the algorithm, you'd want filters to cut out notes below a certain length that are more likely to be bots (remember "Happy New Year"?).

It would be really cool to do a social graph visualization that shows which notes are seeing the most interaction.
▲ ▼
Laeserin | 10 months ago (raw) | root | parent | reply | flag 
 Yes, I mention that in the other comment. Nothing short or too repetitive or similar. "GM" "Good morning" "Hi" "Yo" is all the same thing.

You would also want to filter out anything too reply-guy-ish. Sometimes discussions break out in GM threads, but sometimes it's just a string of GMs, so you'd filter out the second type.
▲ ▼
Laeserin | 10 months ago (raw) | root | parent | reply | flag 
 Here is another example. This note shot up into the trends, but it has a flat structure and the original poster has never replied or written anything else. No human interaction, despite so many replies. We all just look silly.

nostr:nevent1qqstd2mffmavec70lf3q66ycm550c9tyh364quzev9c3pakdfr7kcegpzamhxue69uhkummnw3ezu6r4vfkkz6m9wghxjmczyrw587khexkvw6wl3ku80860crt0fktldg9z0rk05gvrukjc9u5y7qcyqqqqqqgwdv7xx
▲ ▼
Laeserin | 10 months ago (raw) | root | parent | reply | flag +1 
 Well, that's what nostr.band has done, but my idea is a novel type.

Patent it. 😂