farcasterunion

There is a huge problem with spam on the network. There are too many incentives for them, but many new people get marked as spammer and don’t know why. Maybe we can collectively come up with some working theories on how real people trigger this algo so they can avoid it. 

Every time I check this frame by @mvr, I see the number grow.

Reminder: real people can be spammers! (i.e. not just automated accounts — aka bots — engage in spam)

For starters, would suggest looking at number of casts per day the new accounts that are being marked as spammers are casting

Reminder: real people can be spammers! (i.e. not just automated accounts — aka bots — engage in spam)

For starters, would suggest looking at number of casts per day the new accounts that are being marked as spammers are casting https://warpcast.com/pichi/0x0642ae14

I like running, poker and oysters | Founder, Social Forensics | Berkeley & UPenn | Learning to listen

There needs to be transparency around how the spam detection systems work. Claiming that “we can’t publicize it cuz ppl will game it” is not good enough. Remember the reaction people had when I used my own personal judgment to label spam? 

These days, I think less about how to label spam and more on building trust and understanding who someone is through proof of work. Defining what constitutes proof of work for a community is challenging because it needs to be something that's both hard to fake and easy to track