Well, some things are just hard, take time and can't be divided into vertical slices. That's usually not the case with web technologies, if my impression is correct, but it certainly is the case in many other disciplines.
Sure. There are certainly some things which don't lend themselves to regular achievable goals. If you were tasked with building a bridge, you'd probably be hard pressed to put out anything within the first few days, let alone a prototype. That can happen in software too, but I suspect that situations where that is the case are in the minority. This isn't fighter jet avionics, it's a glorified discussion board.
Exactly. As someone who has written one in the span of a few hours, it's hard for me to be impressed by this and all the other attempts to reinvent discussion boards over the years - they all seem to lack some very basic functionality that mainstream forum software has had for years and turn out to be overengineered solution-looking-for-a-problem type things.
I guess the real problem here is that the NYT and the WP wants it to be perfect when it's out on their website. They don't want iteration. If my hunch is correct they'll also use some kind of machine learning to filter out bad comments.