Despite widespread adoption, Patrick Collison notes that AI has not yet produced measurable gains in macroeconomic productivity. He points to recent studies and the lack of corresponding GDP growth outside the U.S. as evidence that the diffusion of these technologies through the economy is slow and complex.
Related Insights
Contrary to the feeling of rapid technological change, economic data shows productivity growth has been extremely low for 50 years. AI is not just another incremental improvement; it's a potential shock to a long-stagnant system, which is crucial context for its impact.
Contrary to a popular narrative, the surge in AI investment has not yet contributed measurably to US GDP growth. This is because the investment largely consists of imported goods, creating a neutral GDP effect, and accounting rules misclassify key semiconductor components as intermediate goods rather than final investment.
Stanford economist Erik Brynjolfsson argues that a major downward revision of 2025 job numbers, while GDP figures remained strong, mathematically implies a massive productivity surge. This suggests AI's economic impact is finally visible in macroeconomic data, moving beyond anecdote and theory.
Traditional metrics like GDP fail to capture the value of intangibles from the digital economy. Profit margins, which reflect real-world productivity gains from technology, provide a more accurate and immediate measure of its true economic impact.
Despite strong productivity numbers alongside flat job growth, economists believe it is too early for AI to be the primary driver. The gains are more likely attributable to businesses becoming more dynamic and achieving better labor-market matches following the pandemic disruptions, rather than a widespread technological revolution.
The anticipated AI productivity boom may already be happening but is invisible in statistics. Current metrics excel at measuring substitution (replacing a worker) but fail to capture quality improvements when AI acts as a complement, making professionals like doctors or bankers better at their jobs. This unmeasured quality boost is a major blind spot.
The argument is that "economic diffusion lag" is an excuse for AI's current limitations. If AI models were truly as capable as human employees, they would integrate into companies instantly—far faster than human hiring. The slow rollout proves they still lack core, necessary skills for broad economic value.
Economist Tyler Cowen argues AI's productivity boost will be limited because half the US economy—government, nonprofits, higher education, parts of healthcare—is structurally inefficient and slow to adopt new tech. Gains in dynamic sectors are diluted by the sheer weight of these perpetually sluggish parts of the economy.
The slow adoption of AI isn't due to a natural 'diffusion lag' but is evidence that models still lack core competencies for broad economic value. If AI were as capable as skilled humans, it would integrate into businesses almost instantly.
General-purpose technologies like AI initially suppress measured productivity as firms make unmeasured investments in new workflows and skills. Economist Erik Brynjolfsson argues recent data suggests we are past the trough of this "J-curve" and entering the "harvest phase" where productivity gains accelerate.