Farcaster Union

Warpcast

May the fourth be with you and grant you a level 2!

Spam labels dataset just updated.

Label transition counts:
null -> 0: 2104
null -> 1: 2275
null -> 2: 220
0 -> 1: 8394
0 -> 2: 2425
1 -> 0: 3495
1 -> 2: 1109
2 -> 0: 2301
2 -> 1: 2532
* -> 0: 7900
* -> 1: 13201
* -> 2: 3754

Check below

Ooooo, lots of new folks got labeled for the first time while FarCon was happening. 

May the odds be ever in your favor.

Pi-chi.eth Artist /animeoutcasts
Hosting: /sakura /kyoto /anime-manga /farcasterunion

https://portrait.so/pichi

I wonder if an icebreaker coefficient can be added to the calculation of the FC spam label 🤔￼￼

Call me /aj | /origami folder | residing in /dcl

yeah of course would be open to it if there's appetite on merkle side

FarCon NYC

cto @ bwl.gg, 100M+ users (bdut.ch) 🙏 $eggs is life

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

CEO at Google • co-founder & rectangle mover at /icebreaker • icebreaker.xyz/j4ck • /j4ck • DJ sixsix

I would love to see a @geoffgolberg style graph of connections or attestations between /farcon-nyc attendees. Of course my little lesson from @farcasteradmin.eth on the dangers of always on NFC showed me the downside of the convenience of the con’s id badge. I know I swiped fewer ppl’s badges after that. Perhaps that kind of cautionary behavior from others would affect the numbers.

easiest way to build on farcaster @ neynar.com | /neynar

Rewards

Agree with you about all scoring systems being inherently limited. 

Even the best scoring system in the world can't easily predict the quality of a brand new user. Web of trust *can* do this well, especially when used as a hybrid solution.

Rewards leaderboards may optimize for engagement, but @jacque being given a spam label is still problematic. 

We are talking with @gt about integrating icebreaker attestations more closely with Farcaster. They are all open and tied to fids when fids exist.

We can even use Farcaster as the attestation infrastructure...

I was interested in a proof of humanity/DID component. Real, person to person interactions and attestations. 

@jacque was at /farcon-nyc and she has a 2 spam label but I don’t see her on the /rewards leaderboard. Why?! She talked to far more attendees than I did. I often ended up chatting with the typewriter poet, Ms. Winham or Carl the saxophone player at @crezno’s party. She talked with so many FCers. I couldn’t keep up. 

I made 22 connections with @icebreaker. Don’t know how many Jacque made. Regardless, the more connections, the greater the likelihood that the account is associated with a real person who spends actual time talking to other ppl. Like coming up to the speakers role in @farhouse and talking, it’s got to count for something.

FYI, @geoffgolberg has proven with his data analysis that there is a STRONG correlation with spam labels and @neynar scores which may be easier to pull.