“The Law of Accelerating Returns” is an essay the inventor and futurist Ray Kurzweil published on March 7, 2001, on his own site KurzweilAI.net. It is the clearest single statement of the framework underlying his later book “The Singularity Is Near,” and it has become one of the most-cited and most-argued-over technology predictions of the century.
Kurzweil’s central move is to contrast two ways of forecasting. Most people, he argues, hold an “intuitive linear” view and assume the current rate of progress continues. The historical record, he claims, instead shows exponential - even doubly exponential - acceleration, because each technological paradigm speeds the arrival of the next. From this he draws his most quoted line: “It isn’t the case that we’ll experience 100 years of progress in the 21st century. Rather we’ll witness on the order of twenty thousand years of progress” measured at today’s rate.
The essay attaches dates to the curve. Kurzweil projects that around 2023 a 1,000-dollar computer will match the raw computational capacity of a single human brain, that by the late 2030s the same money buys far more, and that around 2049 it equals the capacity of the entire human race combined. He frames the “Singularity” - a point of change “so rapid and so profound that it represents a rupture in the fabric of human history” - as occurring near the middle of the century, a date he elsewhere fixes at 2045, with machines convincingly passing the Turing test by roughly 2029.
The essay is a primary source for the exponential, optimistic school of AI prediction. Its specific dated claims have given decades of commentators a concrete scorecard to grade - some milestones arriving roughly on time, others not - which is exactly why it remains a touchstone in arguments over how fast AI is really moving.