Dead on. Preach it, brother.

A backlog of stories, describing work to be done, "happy" and "sad" paths, and acceptance criteria is a good thing. Estimating the relative level of effort to deliver each story is a good thing. Filling sprints with stories based on a team's average velocity is a good thing.

But none of this takes the place of planning or documentation. AND WAS NEVER INTENDED TO.

"Ready, fire, aim" is a stupid, wildly inefficient way to build software, especially mission-critical software.