Anyone can ask a chatbot for Tech Week events. We publish how reliable ours are. Jev tags every event once: audience, topic, depth, party vs. substance, vendor pitch, builder density. Then Plumloom Autoeval (open source) checks those tags.
30 random events, each tagged 3 times. Share of events where every run agreed.
50 hand-labeled reference events.
Event picks scored by Jev · Checked with Plumloom Autoeval · Built in Lovable