curious about the toughest technical and linguistic hurdles for real-time voice translation using ai. issues like handling different dialects, latency, preserving tone and emotions, and achieving high accuracy all seem complex. what do you think are the hardest problems to solve?
[link] [comments]