I feel like he is laying out an interesting methodology here for making logical predictions about the future. Starting with the idea of evaluating the future of technological development by
Understanding how each technology functions. What it is what it's made if how ot works methods etc. then understanding its limitations, what is preventing it from being better at what it does (does he miss the strain of what it could be?-maybe make another model with additions). And then evaluates how likely and easily that will be overcome (match with the processing chip doubling equation).