This tech note is significant for two reasons... one is that it shows Gustafson had fleshed out the concept of scaled speedup long before the Sandia results made it well-known. The other is that Gustafson proposed that supercomputers should be constructed by filling a building with low-cost cabinets based on commodity VLSI, connected by message passing. This view was greeted with laughter at the time, but was prescient; now, every large-scale computer uses this approach.