Seedance 2.0 on live–their strongest multimodal AI video model with native audio | Hacker News Reader