Related to this: I was at UC Berkeley in the early 80's (and did the port of Vaxima (the VAX version of Macsyma, running on Franz Lisp) to the Motorola 68000). There was definitely a rivalry between UCB and Caltech in the symbolic algebra area, and I think the "100x slower" was just an excuse to do it differently. I don't think he would have wanted to appear to be following anyone at UCB.