An obscure 19th-century economic theory about coal consumption has become one of the hottest ideas in Silicon Valley's AI debate. The Jevons paradox - the observation that efficiency gains increase rather than decrease total demand - is being applied to the labor market with striking results.
Atlantic staff writer Annie Lowrey has traced the paradox from Victorian England to today's AI economy, "Hvylya" reports, citing The Atlantic.
In 1865, the economist William Stanley Jevons warned that more efficient steam engines would not reduce England's coal consumption. Cheaper energy would unlock new applications, drive down prices, and pull coal into corners of the economy it had never reached. He was right. British coal production kept climbing until the early 1950s.
Modern examples of the same dynamic are everywhere. LEDs use less electricity than incandescent bulbs, but total electricity consumption has risen because people now have far more gadgets. Broadband and mobile data got faster and cheaper, so people watch short-form video every waking moment - and the world needs more bandwidth, not less.
The paradox plays out in labor, too. In 2016, Nobel laureate Geoffrey Hinton argued the world should stop training radiologists because software would soon do their work. Instead, improvements in medical imaging unlocked new use cases for CTs and MRIs. Patients demanded more tests, and radiologists were the ones administering and interpreting them. AI changed how doctors do their jobs, not whether they have one.
Lowrey cautions that the paradox is not a universal shield. "When technology leads to huge numbers of jobs disappearing quickly, bad things happen," she writes. Coal itself eventually went the way of the horse - all of England's industrial-scale coal mines have closed, and its last coal-fired power plant shut down in 2024.
Also read: how OpenAI aims to squeeze an entire research lab into a single data center within two years.