Voice Agents Now Retrieve Context in <150ms with On-Device Predictiontwitter.com1 point·shawkatkabbara··0 commentsOpen articleSaveView on HN