Select a work node and trace only its audited conceptual dependencies.
The driving question
What changes when ideas, information, networks, organizations, and tasks become productive assets?
Digitization changed the economy less by accelerating calculation than by altering the costs of creating, copying, coordinating, and connecting. Its returns depended on standards, organization, skills, and assets that conventional accounts often failed to see.
Audited knowledge dependency network
Fixed coordinates. Twelve audited dependencies. Distance has no meaning.
Build the digital firm
A copy costs almost nothing. Coordination does not.
45potential direct pairs n(n−1)/2
45 is topology—not economic value.
Failure → repair
Potential connections explode, while governance, organization, lock-in, and complementary capital remain scarce.
Edges are disclosed editorial dependencies; coordinates and distance have no metric meaning.
The canon · Ten landmark works
Read the contribution. Then read the boundary.
Chronological, not ranked. Influence records intellectual reach—not endorsement or empirical validation.
011985
Michael L. Katz and Carl Shapiro
Network Externalities, Competition, and Compatibility
American Economic Review 75(3), 424–440
What it made visible
Network value depends on compatible participation, linking expectations, installed base, and competitive advantage.
Why it mattered
It made compatibility and network scale central economic variables and supplied the basic model for standards wars and software ecosystems.
The limit
AI networks also depend on endogenous quality, data feedback, compute, inference cost, synthetic agents, and negative externalities; user count alone is insufficient.
Competing Technologies, Increasing Returns, and Lock-In by Historical Events
The Economic Journal 99(394), 116–131
What it made visible
Early stochastic adoption can tip increasing-returns markets into persistent, potentially inferior equilibria.
Why it mattered
It gave economics a formal account of path dependence, tipping, irreversibility, and the welfare ambiguity of standards competition.
The limit
AI technologies co-evolve through learning and updates; lock-in may reside in data, workflows, evaluation infrastructure, compute, and skills rather than installed products alone.
Journal of Political Economy 98(5, Part 2), S71–S102
What it made visible
Profit-seeking R&D produces non-rival, partially excludable ideas under monopolistic competition.
Why it mattered
It supplied the microeconomic architecture of endogenous growth and linked innovation to human capital, market incentives, and institutions.
The limit
AI makes idea generation partly machine-based and forces distinctions among generation, correctness, validation, ownership, diffusion, and implementation.
The Productivity Paradox of Information Technology
Communications of the ACM 36(12), 66–77
What it made visible
Technical capability and measured productivity can diverge through mismeasurement, lags, redistribution, and implementation failure.
Why it mattered
It reframed IT as a technology whose return depends on complementary investment, organizational redesign, learning, and measurement.
The limit
AI adds probabilistic quality, verification work, model risk, hidden cognitive services, and consumer surplus that installed-capital measures capture poorly.
Information Rules: A Strategic Guide to the Network Economy
Harvard Business School Press
What it made visible
High first-copy cost, cheap replication, versioning, switching costs, and network effects shape information-good strategy.
Why it mattered
It unified industrial-organization insights into the practical economics of Internet-era pricing, standards, rights, and lock-in.
The limit
Generative systems produce personalized outputs at non-zero inference cost, update continually, take actions, and introduce provenance, liability, and data-rights problems.
Timothy F. Bresnahan, Erik Brynjolfsson, and Lorin M. Hitt
Information Technology, Workplace Organization, and the Demand for Skilled Labor: Firm-Level Evidence
The Quarterly Journal of Economics 117(1), 339–376
What it made visible
IT, decentralized organization, new products, and skilled labor form a complementary investment system.
Why it mattered
Firm-level evidence explained why identical hardware investments yield different productivity outcomes and can raise demand for skills.
The limit
AI may substitute for skilled cognitive tasks and move the scarce complement toward judgment, verification, data stewardship, workflow ownership, and authority to act.
Journal of the European Economic Association 1(4), 990–1029
What it made visible
Platforms choose asymmetric prices and rules to internalize cross-group participation externalities.
Why it mattered
It showed that price structure, not only the total price level, determines platform participation, welfare, competition, and governance.
The limit
AI adds model, data, compute, tool, and agent sides; synthetic participation, ranking, defaults, and vertical integration can matter more than money prices.
David H. Autor, Frank Levy, and Richard J. Murnane
The Skill Content of Recent Technological Change: An Empirical Exploration
The Quarterly Journal of Economics 118(4), 1279–1333
What it made visible
Technology changes tasks within jobs, substituting for routine rules while complementing non-routine problem solving and communication.
Why it mattered
It shifted labor economics from occupations to task bundles and supplied a mechanism for job redesign and changing skill demand.
The limit
Foundation models reach nominally non-routine language and prediction tasks, requiring reliability, context, verification, tacit knowledge, liability, and demand creation to enter the model.
Measuring Capital and Technology: An Expanded Framework
Chapter in Measuring Capital in the New Economy, University of Chicago Press/NBER
What it made visible
Persistent spending on software, data, R&D, design, brands, skills, and organization is investment even without physical form.
Why it mattered
It revealed intangible investment at roughly tangible-investment scale and changed how economists interpret saving, output, productivity, and the digital firm.
The limit
AI blurs training investment, inference expense, purchased services, internal capital, ownership, depreciation, valuation, and quality adjustment.
These ten works are ordered by first publication. Selection weighs paradigm effect, conceptual durability, downstream reach, cross-generational influence, and non-redundancy. Every entry names a contribution and a boundary.
