---
type: "synthesis"
articles: ["a075", "a084", "a101"]
tags: ["jevons-paradox", "induced-demand", "efficiency", "friction"]
id: "cross-efficiency-paradox"
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-seg-futures"
sourceUrl: "(unified vault: 11 sources)"
sourceTitle: "HBR — Futures / Macro — where it's all going"
---
# Why Efficiency Won't Save You

Three articles independently rebut the comforting assumption that "cheaper/better AI fixes its own problems." They are the same economic law wearing different hats.

**A101** applies Jevons to energy: [[concept-ai-jevons-paradox]] and [[claim-efficiency-increases-demand]] / [[contrarian-efficiency-increases-demand]] — [[entity-deepseek-d2]]'s cost collapse *expands* viable use cases, so total watts rise, not fall.

**A084** applies the identical logic to labor: [[concept-induced-demand]] — cheaper code expands the total surface of software that must be built, secured, and maintained, so demand for high-level engineering *surges* (the radiology parallel). It then argues the counter-intuitive remedy: [[concept-deliberate-inefficiency]] / [[contrarian-inefficiency-is-good]] — reintroduce friction (named sign-offs, pairing) to internalize the shared cost of training the next generation.

**A075** supplies the macro data point: despite the AI-investment boom, [[claim-post-covid-downshift]] shows global digital momentum *decelerated* (4.3%→2.4%) — efficiency and capital did not automatically translate into evolution.

**The unifying insight:** efficiency reallocates and expands demand rather than reducing aggregate load; and in systems dependent on human judgment or shared resources, *some* friction is protective, not wasteful. Both A101 and A084 attach the same honest caveat — hard caps, carbon pricing, or strong governance could flip the sign. See [[cross-relocating-scarcity]] and [[cross-judgment-accountability]].