Why AI is in a particular club
The idea
In a 1995 paper called Engines of Growth, Timothy Bresnahan and Manuel Trajtenberg coined the term general-purpose technology (GPT) for a rare kind of innovation that reshapes the whole economy. Classical examples are the steam engine, electricity, the internal combustion engine, and the computer. GPTs spread through nearly every sector rather than stopping at one.
The three criteria
Bresnahan and Trajtenberg proposed three tests. First, pervasiveness — the technology spreads across most sectors. Second, improvement over time — it gets cheaper and better for decades. Third, innovation spillovers — it unlocks new technologies and new ways of doing unrelated things. A true GPT scores on all three.
Why AI qualifies
By the mid-2020s, large language models and their derivatives pass all three tests. They are used in law, medicine, education, construction, design, journalism, and software engineering. Their capabilities keep improving on a curve that shows no sign of flattening. And they have already unlocked new products (AI-native drug discovery, code-writing code) that did not exist before.
Why it matters for this course
Most past economic transitions failed to produce mass long-term unemployment because displaced workers found new work created by the same technology. If AI is a GPT, the same long-run logic may apply. But GPT transitions also produce severe short-term dislocation. Both halves of that sentence are real, and both will appear again in this course.
Can you name three technologies that reshaped more than one industry? What did they have in common?