Institutional Capacity Gap

In the AI Story

The capacity gap is invisible precisely because it is ambient. In functioning technology ecosystems, the capacity is so pervasive it becomes like water to a fish: the professional norms of code review, the mentoring relationships that transmit tacit knowledge, the quality standards that distinguish shipping code from production code, the market mechanisms that convert output to value — all of this is the medium through which technology operates, not a separate thing to be invested in. Practitioners inside these ecosystems often cannot see the capacity because they have never operated without it.

The capacity gap becomes visible at the edges of the functioning ecosystem — in regions where institutions are weaker, where mentoring networks are sparser, where quality standards are local rather than global, where market mechanisms do not connect builders to users. In these contexts, the same tools produce different outcomes, and the difference is not in the tools but in the capacity gap between the two contexts. This is what Toyama documented in school after school, clinic after clinic, extension service after extension service: the capacity gap explained the outcome gap with a consistency that no technology-focused explanation could match.

For AI, the capacity gap has specific dimensions. It includes educational capacity (can the user evaluate AI output?), institutional capacity (are there quality standards for AI-augmented work?), market capacity (can AI-produced output reach paying users?), and cultural capacity (do the professional norms support sustained engagement rather than compulsive use?). Each of these dimensions is produced by investments that take years and depend on functioning institutions that cannot be substituted by more technology.

The political economy of the capacity gap is unforgiving. Closing the access gap generates revenue for the technology industry. Closing the capacity gap generates expense for governments, nonprofits, and communities that are already under-resourced. The industry's incentives drive investment toward the access gap. The outcomes require investment in the capacity gap. The alignment fails, and the failure reproduces itself across every technology transition.

Origin

Toyama developed the distinction across his Microsoft Research India years and articulated it formally in Geek Heresy. The framework builds on and sharpens analyses from development economics — particularly the new institutional economics of Douglass North and the capability approach of Amartya Sen — by bringing field evidence from technology deployments that could not be explained within either framework alone.

Key Ideas

Two gaps, not one. The access gap and the capacity gap are distinct, and attention to the first without investment in the second produces formal participation without substantive outcomes.

The capacity gap is ambient. It is invisible inside functioning ecosystems because it is the medium of function. It becomes visible only at the edges.

Multiple dimensions. Educational, institutional, market, and cultural capacity each matter, and all must be present for technology to produce outcomes.

Resistant to technological substitution. The capacity gap cannot be closed by distributing more technology, because the capacity is what technology depends on, not what it produces.

Political-economic asymmetry. The incentives of the technology industry drive investment toward the access gap; the outcomes require investment in the capacity gap. The misalignment is structural.