I want faster AI transcription latency for individual videos. Currently, AI generation (Whisper V3 Turbo) takes 20+ seconds, but I need it to be under 10 seconds to match competitors like tokscript.com. This is important for my workflow which requires near-instant results for individual social media clips.