The point I was trying to make was that these things are no longer what determines success or failure. Incremental improvements are possible, but do not dominate.
Consider the fight between a company that's great at RF design, and a company that's on top on software. Which do you think will win in the market for cell phones? (This is a trick question.)
Mature technologies tend to be things that can be outsourced. So one can make some money (if not a lot) as a supplier of these things in a horizontally integrated industry.
In some instances perhaps that's so. And I'd mostly agree with you re mature industries and outsourcing. But in the instances I've given (and they were just off-the-cuff, but there are many more) I'd argue that they are new and developing tech that's nowhere near full maturity. New modulation methods combined with newer polarization techniques have the potential to revolutionize spectrum management (that's where many more channels can be made available in already-overcrowded spectrum space). With new meta materials the sky's the limit—from new types of specialized antennae to optical lenses that are an order of magnitude smaller whilst bettering optical sharpness and abberations, and that's just for starters.
Re cell phones, I'll just say this. RF engineering went through a revolution about 10 to 15 years before cell phones became commonplace. The electronics shrunk in size by an order of magnitude, the RF front ends including its transistors increased their FT (transition/cutoff frequency) by over an order of magnitude, and the noise figure of RF front ends was reduced to well below 1dB.
The net/combined result of these three improvements in RF engineering was a truly remarkable achievement. Fact is, there'd be no smartphones or integrated GPS without these developments all having happened. I'd also add that by and large programmers simply haven't a clue about this very special aspect of RF engineering. They just take for granted that the hardware does what it says, to them it's just a black box.
Programming, which cannot be really called engineering because many decades on, it's still an undisciplined way of working that's more akin to an art form than true professional engineering, can however provide the glitz factor that sells phones, but it's completely lost without the state-of-art RF hardware substrate upon which it sits. Moreover, the next revolution in phone technology such as direct satellite services, will only be practical after a further revolution in RF hardware (what passes for some showy demos today is nowhere near the mark if the tech is to be widely adopted).
Thirty years on essentially programming has adopted some new techniques for developing software but the fundamental underlying problem of the lack of rigorous standards together with proven methods to solve problems by computer remains unresolved, and in some instances the software industry has gotten demonstrably worse.
I've not the time to put current references that support my claim but a search will soon find them—ones written by both authoritative people within the programming industry and elsewhere such as in computer science research.
Consider the fight between a company that's great at RF design, and a company that's on top on software. Which do you think will win in the market for cell phones? (This is a trick question.)
Mature technologies tend to be things that can be outsourced. So one can make some money (if not a lot) as a supplier of these things in a horizontally integrated industry.