This part makes sense to me. One can do the initial system design up front and avoid considerable refactoring, or one can rush to code and re-design it a few times. This might just be my style of thinking. I know a few people that can just write code and build useful applications. Others that do this end up having to throw away lots of code before having something that works well. The key is to find the method that works best for you and use it.

"We don't need to make all sorts of fancy diagrams, UML charts, story cards, or any of the other things that delay coding."

Actually, I do need those diagrams. To figure out any system that someone wants me to build, I need to make a high-level sequence diagram and sometimes also an object diagram. These are quick drawings with paper and pencil and help me figure out the system in my head. That's the hard part, actually, as once it's all in designed the implementation is generally simple. Things like interfaces to commercial and open source software, cluster applications, and the like can complicate things, but in general the implementation isn't bad.