Scaling is necessarily a function of learning. At any given time, there is a size above which we can't build. Some of the reasons are external -- a limited market, for example -- but some are clearly internal. We don't know how to build it, as with steel pressure vessels above a certain size and pressure capability; or, if we think we know how to build, we sometimes don't know how to do it economically -- as with sodium-to-steam heat exchanges of large capacity. Sometimes, the connection of risks in one package gets too big to handle -- thus the imposition of a capacity limit on light-water reactors in the United States. And there are doubtless other reasons for things being "too big to build". All these limitations change with time. We can learn how to avoid or overcome them, or in some instances to ignore them as they turn out to be mental rather than physical limits.