Clayton Christensen

In the [YOU] on AI Field Guide

The cycle that began with [YOU] on AI insists on seeing the machine clearly—without the narcotic of hype or the paralysis of fear. Christensen is the cycle’s most rigorous analyst of the competitive mechanism by which clear sight is most urgently required: the structural dynamics that determine which organizations survive a technology transition and which are displaced regardless of how rationally they behave within their existing frameworks. His central paradox—that the incumbent does everything right and loses anyway—is the framework that prevents the false comfort of believing that quality and customer focus are sufficient protection against structural competitive change.

His framework explains the specific pattern the cycle documents in the software industry: the SaaSpocalypse, the trillion-dollar repricing of enterprise software companies, was not driven by AI-generated code that exceeded professional quality. It was driven by AI-generated code that was adequate for the three functions the median user actually needed, available at negligible cost, and deployed by people who had never been customers of the professional software market. The incumbents were correct that the code was architecturally inferior. The assessment was irrelevant to the competitive dynamic, which was not about their existing customers but about the billions of non-consumers they had never served.

Christensen’s distinction between sustaining and disruptive innovation is the diagnostic tool that prevents the most common analytical error in the AI discourse: treating all AI applications as a single phenomenon. The experienced developer using AI to do existing work faster is experiencing a sustaining innovation that benefits her employer without threatening her market position. The marketing manager building her own analytics dashboard through conversation is experiencing a disruptive innovation that bypasses the professional developer entirely—creating a new category of production that serves non-consumers at costs too low for professional developers to match. The triumphalist and the elegist, both correct within their frames, are observing different phenomena and drawing conclusions that are each valid but mutually inconsistent because they are watching sustaining and disruptive uses of the same technology and failing to distinguish them.

The jobs-to-be-done lens reveals the deepest insight the cycle draws from Christensen’s work: the job that AI was hired to do is not “write code.” It is close the gap between what I can imagine and what I can build—the imagination-to-artifact ratio collapsed to the width of a conversation. This job had been waiting, unfilled, for the entire history of computing, and the adoption speed of AI tools measured not product quality but the depth of the unmet need. Understanding this is understanding why the disruption is not bounded by the size of the software development market. It is as large as the population of people who have ever had an idea they could not realize.

Origin

Clayton Christensen was born in Salt Lake City in 1952 and trained as an economist at Brigham Young University before studying at Oxford as a Rhodes Scholar and earning his MBA and doctorate from Harvard Business School, where he spent most of his career. His intellectual formation was shaped by a specific puzzle: why did excellent, well-managed companies fail? The conventional answer—bad management, missed trends, insufficient innovation—did not fit the evidence. Many of the companies that failed had been the most admired in their industries and had responded to competitive threats in ways that analysts praised at the time. Christensen set out to find the mechanism of their failure.

His breakthrough came through a meticulous study of the disk drive industry, where the pattern emerged with unusual clarity. In successive generations of disk drives—from 14-inch to 8-inch to 5.25-inch to 3.5-inch—the leading firms of each generation failed to transition to the next, not because they lacked the technology but because the new drives initially served markets their best customers did not value. Each time, the rational resource allocation process directed investment away from the new format and toward improving the existing one. Each time, the new format’s improving trajectory eventually intersected with the mainstream market’s requirements, and by then the incumbent’s window for response had largely closed.

The framework Christensen developed from this research—and extended across industries as different as steel and excavators and department stores—was published as The Innovator’s Dilemma in 1997. The book was named the most influential management book of its era by several assessments, and Christensen spent the following two decades developing its implications in The Innovator’s Solution (2003), applying the framework to education, healthcare, and other sectors where established institutions displayed the same structural vulnerabilities as incumbent firms. He died of leukemia in January 2020, leaving the Christensen Institute to continue applying his framework to the AI transition he did not live to see.

Key Ideas

Disruptive Innovation. The mechanism by which incumbents are displaced not by superior competitors but by inferior ones. A disruptor enters at the low end, serving customers the incumbent has overlooked or abandoned, at performance levels the incumbent’s best customers would find inadequate. It improves along a trajectory initially invisible to the incumbent because the trajectory targets dimensions the incumbent does not measure. By the time the disruptor’s performance intersects with mainstream market requirements, the incumbent’s window for response has largely closed. The pattern is not a tendency or metaphor; it is a structural dynamic observable across industries as different as disk drives and universities. Disruptive innovation is not better. It is adequate—and adequacy, aimed at a large enough non-consuming population, is always a more powerful competitive position than excellence aimed at an already-served market.

Sustaining vs. Disruptive. The foundational distinction that prevents the most common analytical error. Sustaining innovations improve existing products for existing customers; incumbents almost always win the sustaining innovation competition. Disruptive innovations serve different customers or serve existing customers in contexts where the existing product is unavailable; incumbents almost always lose this competition because their resource allocation processes systematically direct investment away from opportunities their existing customers do not value. The same underlying technology—AI—can be sustaining or disruptive depending on the business model in which it is deployed. Treating them as a single phenomenon produces analysis that is simultaneously optimistic and pessimistic without being useful.

Jobs to Be Done. The reframing of customer behavior: people do not buy products, they hire products to do jobs. The job is a progress the customer is trying to make in a particular circumstance. The competitors for any job are not the products in the same category but all the candidates the customer might hire to make the same progress. The job AI was hired to do is not write code but close the imagination-to-artifact gap—and its competitors are not other development tools but non-consumption, the void where an idea lived and died without becoming real.

The Innovator’s Response. The three-element prescription for incumbents facing disruption: recognition that the threat is structural rather than cyclical; separation of a distinct organizational unit with its own cost structure, metrics, and cultural norms to pursue the disruptive opportunity; and willingness to cannibalize the existing business before an external disruptor does it. The response is demanding and most incumbents fail at the third element: the organizational immune system activates when the separate unit begins competing with the parent, and the pressure to absorb and integrate neutralizes the disruptive potential that separation was designed to protect.

The Value Network Shift. Disruptions do not merely redirect existing value; they shift the entire value network—the ecosystem that defines what is valued and how value is captured. The pre-AI value network in software rewarded execution: the ability to translate specifications into code was the foundational capability. The post-AI value network rewards judgment: the ability to decide what should be executed, for whom, and why. This is not a reallocation within the existing network. It is the creation of a new network with different participants, different metrics, and different power dynamics.