Show HN: Dia, an open-weights TTS model for generating realistic dialogue

pretty cool - love seeing open stuff like this come together so fast. you think all this will ever match what a real voice actor can pull off, or something totally new comes out of it?