The recent AKT-IRC meeting in Southampton was, for me, thought provoking. There are numerous niches within the AKT view of knowledge management lifecycles for the methods, techniques and tools which we produce but exactly how these are realised depends on one's view of the goal of the exercise. I intend to discuss this, very informally, and give my own view - one of many equally legitimate views. To provoke discussion I will make the (deliberately exaggerated) distinction between the extreme version of "smart stuff happens" (where we discover surprising knowledge from unexpected sources) and the opposite extreme where "useful stuff is delivered" (where we engineer into the fabric of our lifecycles the ability to deliver certain kinds of information reliably).