> Curiously enough, looking at the numbers now, I realize that the base speed of the LARC was only 20x the Elliott 903C, though with floating point, etc.—a fact

> Curiously enough, looking at the numbers now, I realize that the base speed of the LARC was only 20x the Elliott 903C, though with floating point, etc.—a factor that pales in comparison with the 500x speedup in computers in the 40 years since I started working on cellular automata. 500x? Shouldn’t the speedup over 40 years be something between 10,000x (increase in single-threaded performance) and 1,000,000x (increase in number of transistors).