Innovation Economics · Historical overlay
The GPT J-curve.
General purpose technologies — electricity, the semiconductor, the internet — diffuse across the whole economy, generate complementary innovations, and reorganise production. The productivity gains arrive late, after firms restructure around the technology. Plotted on a common axis of years-since-introduction, all three GPTs trace the same J. AI is at year 4. The historical pattern says the steep ramp begins after firms reorganise, not when the technology arrives.
The thesis
Solow's famous quip — "you can see the computer age everywhere except in the productivity statistics" — was published seven years before the IT productivity acceleration actually arrived. The historical lesson is that the firms that capture GPT gains are the ones that reorganise themselves around the technology, not the ones that simply buy it. AI is a strategic and organisational question, not a procurement one.