It would be amazing if, in multitrack, Descript could notice when two people are talking at once and separate their speech out, ideally into some kind of logical order.