Oddbean new post about | logout
 Continuing to develop features (hate speech detection, sentiment analysis) for nostr-filter-relay. Although, currently need to read thousands of data manually to evaluate the quality of classifier 😅 
 Oof, just be happy that it's not image classification 😅 
 Yes. Thankfully, i've done already those part as first feature. Didn't want to do anymore 😅

Respect for AI researcher who've built the classification model, we can integrate the model easily to apps. 

Otherwise this scenario might be happen,
Dad: "Son, what are you doing? Did you watch p*rn all day?"
Son: "No Dad, i swear! This is a .. research.. yes research .. for my project"
Dad: "Go outside. Touch the grass."