Cognification

In the [YOU] on AI Field Guide

The [YOU] on AI cycle is, at one level, a field guide to living through the early phase of cognification: the period when the cognitive tools arrive at individual desks before the institutional frameworks required to govern them have formed. The cycle asks how individuals should position themselves within a transition they did not initiate and cannot stop, and Kelly’s framework provides the historical context that makes the question tractable. Cognification has been coming since Kelly proposed the verb in 2016. The tools are now here. The institutions are not yet built.

The Law of Amplification—Kentaro Toyama’s principle that technology amplifies existing capacity—applies to cognification as it applied to electrification. Electrification did not automatically produce broad prosperity; it produced enormous gains for the owners of the machines before labor institutions were built to share the gains more broadly. Cognification will follow the same pattern unless the institutional work is done faster—and the speed of cognification, unlike the speed of electrification, is measured in years rather than decades.

The cycle’s specific claim that the builder in Trivandrum achieved a twenty-fold productivity multiplier in a week is, in Kelly’s framework, a data point in the cognification transition—the moment when a specific group of experienced professionals first felt the full force of distributed cognitive capability. What the cycle does not show, and what Kelly’s framework demands asking, is what the Trivandrum room would have produced if its twenty people had been twenty less-experienced people with weaker institutional support. The cognification is real. Its distribution is uneven. The institutions that determine how unevenly it distributes have not yet been built.

Origin

Kelly proposed the verb “cognify” in a 2016 interview to describe the emerging second industrial revolution: “The first saw us put the power of muscle into objects in the form of energy. Next, we will cognify anything that is electric.” The formulation was characteristically precise: not “automate” (which implies replacement) and not “augment” (which implies supplementation), but “cognify”—to add cognition to, to extend cognitive capability into, to distribute intelligence the way electricity was distributed. The verb marked the conceptual shift from AI as a specific tool to AI as an infrastructure.

Kelly developed the concept across The Inevitable (2016), where cognification appears as one of twelve forces that would characterize the next thirty years, and in subsequent essays, interviews, and keynotes. At a 2025 CEIBS keynote in Shanghai, he argued that cognification would reshape civilization as thoroughly as electrification did, and over a comparable timescale—not in a quarter, not in a year, but over decades, with real costs during the transition and genuine expansion afterward.

Key Ideas

The structural parallel to electrification. Electrification distributed physical force; cognification distributes cognitive capability. Both eliminate bottlenecks: electrification eliminated the bottleneck of human muscle in physical production; cognification eliminates the bottleneck of human expertise in cognitive production. Both are infrastructural: neither produces value directly but enables all the value produced by systems that depend on them. Both require institutional construction to distribute gains broadly: the eight-hour day, the weekend, public education, and labor law were not produced by electrification itself but by the deliberate institutional work that directed electrification’s gains toward broader benefit.

The elimination of translation cost. Cognification is most visible where it eliminates the translation cost between human intention and artifact. The designer in São Paulo had the intention and the understanding of user needs; she lacked the engineering expertise required to translate intention into working code. Cognification eliminated that bottleneck. The eleven-month timeline collapsed to an afternoon. What the tool did was not think for her but translate for her—take the dashboard she could see in her mind and render it in code she could not write. As translation costs approach zero across more domains, the class of people who can express their intentions as artifacts expands dramatically.

AI as infrastructure, not tool. Kelly’s most consequential move is to frame AI not as a powerful tool but as an infrastructure—comparable not to a hammer but to electricity. Infrastructure is not owned by its users; it is accessed by them. Infrastructure does not serve one purpose; it enables every purpose. Infrastructure requires governance as well as technology: the electrical grid required regulation, safety standards, and pricing rules that took decades to establish. The technium’s cognification will require analogous governance, and the delay between infrastructure arrival and governance construction is where the transition costs accumulate.