The sustaining-disruptive distinction has particular force in the AI transition because both dynamics are operating simultaneously. Sustaining use of AI appears in Segal's Trivandrum observations: experienced engineers using Claude Code to amplify their existing capabilities, performing the same jobs faster, with less friction, at measurable productivity multiples. Sustaining use helps incumbents — the firm whose experienced professionals become twenty times more productive benefits directly without facing a threat to its market position, organizational structure, or competitive advantage.
Disruptive use appears elsewhere: the marketing manager building a custom analytics dashboard without requesting engineering resources, the teacher building an interactive lesson without knowing what an API is. These users are not doing existing work faster. They are doing work that did not previously exist in their professional repertoire, serving themselves in a domain where they were previously non-consumers. Disruptive use threatens incumbents by expanding the population of producers and reducing dependence on professional service providers.
The divergent public reactions to AI that You On AI catalogs — the triumphalists celebrating productivity gains, the elegists mourning the loss of craft — are not disagreements about the same phenomenon. They are observations of different phenomena. The triumphalist is typically engaged in sustaining use; her productivity has genuinely increased. The elegist is typically observing disruptive use; the market for his specific expertise is genuinely declining. Both are right; neither sees the whole picture.
The strategic error most incumbent software firms are making in 2026 is investing heavily in sustaining uses of AI while dismissing disruptive uses. Michael B. Horn of the Christensen Institute has put the point directly: the business model, not the technology, determines whether AI sustains or disrupts. The same model deployed within an incumbent's existing business model sustains. Deployed within a new business model that serves non-consumers at dramatically lower cost, it disrupts.
The distinction was introduced in The Innovator's Dilemma (1997) and elaborated in The Innovator's Solution (2003). It has been refined continuously by Christensen, his collaborators, and the Christensen Institute over three decades of application across industries.
Categorical, not continuous. Sustaining and disruptive are two distinct competitive dynamics with different causes, trajectories, and outcomes.
Incumbents win sustaining, lose disruptive. The same organizational capabilities that produce victory in sustaining competitions produce defeat in disruptive ones.
Business model determines category. The same technology can sustain or disrupt depending on the business model in which it is deployed.
Both can operate simultaneously. A single technology can be a sustaining innovation for one population and a disruptive innovation for another.
Some analysts, including Saneel Radia and Jason Cohen, have argued that AI may reverse the innovator's dilemma by favoring incumbents through data moats and proprietary customer relationships. This argument conflates sustaining and disruptive uses in exactly the way the framework warns against: the incumbent's data advantage is a sustaining advantage that protects against sustaining competition but does not protect against disruptive competition serving non-consumers.