It isn't non-human algorithm which chooses which evidence the jury sees.
In a real trial, when the defendant takes the stand and lies about what they said (i.e. they changed the post after a buddy warned them their post had been alerted on), the attorney for the plaintiff would present evidence of what was actually said (i.e. the screen capture of the post before editing). That can't happen here because what the jury sees is simply the mechanical selection of up-thread posts. The offensive comment is not visible to anyone who is on the jury - even though it is still visible on DU to anyone reviews the edit history of the post
Further - a post can only be adjudicated once. It would be very easy to edit an incredibly offensive post once a buddy alerts the poster that their post is being juried, wait until the jury leaves the post, then re-edit the post back to the original. Since a post can't be adjudicated more than once, the offensive post will stand.
A second problem is people who accidentally reply to the wrong post. This makes the mechanical selection of posts visible to the jury non-sensical because the conversation makes no sense because a significant portion of the conversation is missing - even though it makes perfect sense when you can view the entire thread.
You do you. But if the post I'm asked to adjudicate looks completely innocuous, or the conversation I'm shown is complete nonsense, I'm going to the thread to see if it was edited from something which should be hidden, or if there was an accidental reply to the wrong thread (or anything else which is only apparent in the context of the entire conversation).