Those questions they gather I think will be useful to input to their algorithm in the end.
j 下一段next speechk 上一段previous speech
