Abstract: I describe some of the prediction projects I worked on in the late 60's and 70's as a consultant to industry and government, and discuss the prediction methodologies available at that time. On reentering the academic scene at Berkeley in 1980, I encountered and will describe an interesting clash of cultures. In the mid-1980's powerful new non-linear methods became available and engendered a new community, separate from statistics, working on large and difficult prediction problems. Very significant advances in the last five years, contradict much of the wisdom of previous decades. The important names are Rashumon, Heisenberg and Bellman.