Fluid Inference releases a distilled Clef text decision model for Apple silicon, with GPU and Neural Engine support; its M5 Pro demo triages 1,000 tickets in 57 seconds.