Needs More Ratings: Publishing Community Notes by Precedent

Community-driven content moderation has become commonplace on platforms such as X and Meta's Threads, where regular users can add helpful context to misleading posts. Recently, LLMs have also been employed to write and submit notes faster, but note selection remains a pressing bottleneck. X's Community Notes' bridging-based algorithms publish only one in ten proposed notes and often do so hours after the original post's peak engagement, limiting the reach and corrective value of the note. Drawing from social choice, we propose a guarded precedent-closure rule, which publishes notes if their rater evidence is stronger, on multiple tracked dimensions, than the evidence of earlier published notes. In a chronological offline replay on X's public data (857,256 notes created Nov 2025-Aug 2026), the rule recovers 84.0% of the incumbent's CRH decisions, reaches shared notes a median 1.5 hours earlier, and adds 50,156 publications. The rule lets helpful notes, written by humans and LLMs, reach readers sooner and more often.