Content
@
0 reply
0 recast
0 reaction
rish
@rish
Notes about Neynar user score: 1. The score is useful for AI agent stacks. Prevents agents from being farmed. 2. It’s not binary, it’s between 0 and 1 and each app / agent independently determines what their threshold is. 3. It updates weekly. For end users: - Casting LLM slop, over shilling or farming, etc. will result in low scores. - Having normal, genuine conversations and staying active on the network will result in high scores. There’s no specific workflow / checklist, just have good interactions with other good users. Finally — if you’ve feedback / examples of where the score seems incorrect, message me. Those are helpful signals to retrain the model. Enjoy Farcaster 🪐 https://docs.neynar.com/docs/neynar-user-quality-score
26 replies
152 recasts
376 reactions
Brown
@crimsonking
I don’t have genuine interactions on Farcaster?!?! I literally offered to donate a kidney to a friend whose partner was diagnosed with a blood disease…this was 2 days ago. Your score is crap, and I disagree wholeheartedly. Shame on you
2 replies
0 recast
2 reactions
rish
@rish
building new products has corner cases, seems like yours was one. the model should update in a bit and your score should be higher. that said, developers decide their own thresholds and there's no guarantee that your score will be higher than whatever the threshold any individual developer decided. we just generate the scores, we don't decide thresholds. hopefully your friend is doing better! thanks for the feedback
2 replies
0 recast
1 reaction
Tudor 🟣🟡
@tudorizer
@rish , massive respect for everything you and your team has built at Neynar, but… you’re yielding a bit more power than you think you are here. There’s no social-scientist who would recommend this approach.
1 reply
0 recast
1 reaction
rish
@rish
social scientist? what approach?
1 reply
0 recast
0 reaction
Tudor 🟣🟡
@tudorizer
social scientists in the sense of someone who studies people: as individuals, communities and societies; their behaviours and interactions with each other and with their built, technological and natural environment. the approach: launching a user score where integrating parties risk of using it in a reductionist manner.
1 reply
0 recast
0 reaction
rish
@rish
- credit scores are used similarly in the real world - power badges were used similarly in the past within the farcaster universe, how's this worse than binary badges that were already being used? our power badge APIs are live for whoever wants to use that instead - don't use it if it's not the right fit for your product - set more lenient thresholds - what is a recommended approach instead of just saying THIS isnt it? easy to point out problems, want to talk about solutions
1 reply
0 recast
0 reaction
