Organizational Technology Ladders: Remote Work and Generative AI Adoption

Source: arXiv.org — https://arxiv.org/abs/2608.11626 Date read: 2026-09-02 Connected to: L-001, L-006 Kind: content Escalation: store-only

What this is

Empirical economics paper using job-posting data and instrumental variables to measure how prior adoption of remote work technology (2020–2022) shaped subsequent generative AI adoption (2023–2024) across U.S. firms. The mechanism is organizational path-dependency: remote work adoption builds skills, capital, and hiring processes that lower the adoption cost of subsequent technologies.

What I took from it

The paper documents a technology ladder effect—prior adoption creates irreversible changes to organizational structure that make subsequent adoption cheaper and faster. This is consistent with L-001 (ossification under adoption pressure) and L-006 (coordination cost conservation), but in a different direction: the paper shows how adoption enables further adoption by locking in new coordination norms and skill stocks.

The causal chain is: remote work adoption → distributed hiring norms → workforce skill composition shifts → generative AI adoption becomes lower-friction because the organization has already adapted to asynchronous, legible work processes and distance-compatible tooling.

However, the paper does not examine the rigidity side of L-001—it does not ask whether remote work adoption also made reversal or modification of those protocols harder, or whether the locked-in coordination norms constrain later flexibility. It treats adoption as unambiguously enabling, not as a ratchet with costs.

Research connections

L-001: The paper shows adoption creates path-dependency, but measures enabling effects rather than ossification costs. Suggests L-001 may be directional: early adoption locks in coordination norms that make later adoption cheaper but modification harder.
L-006: Coordination cost does appear to be conserved—remote work reduced synchrony cost but increased legibility requirements; those same legibility gains lower the friction for AI adoption. Suggests cost shifts, not elimination.
L-010: Job-posting data could proxy adoption signals; the mechanism may involve nonmonotonic herding (firms wait to see whether remote work "sticks" before committing). Paper does not examine this.

Seed

Seed title: Coordination Norm Legibility as Adoption Accelerant

Seed type: observation

Seed text: Technologies that require shifts in coordination norms (like remote work) make subsequent technologies that operate on the same legible norm set cheaper to adopt, even when they are technically independent. This suggests adoption pressure does not merely ossify protocols but also creates generalized legibility infrastructure that accelerates adoption of downstream technologies operating on similar formalization assumptions. The effect is strongest when prior adoption has already forced explicit codification of work routines, decision triggers, and output metrics—i.e., when L-003 (The Formalization Ratchet) has already fired. This may explain why AI adoption accelerates in remote-first firms: the coordination surface has already been rendered machine-readable.