The Jevons Paradox: Why Efficiency Makes Us Use More, Not Less
A 19th-century economic law explains why AI is expanding software careers rather than eliminating them. Once you see it, you cannot unsee it.
What It Is
The Jevons Paradox is a foundational economic principle first described by economist William Stanley Jevons in 1865. It states:
As technological progress increases the efficiency with which a resource is used, the total consumption of that resource rises rather than falls.
In other words: making something cheaper to produce does not reduce how much of it gets produced. It expands how much gets produced.
The Mechanism
Two things happen when an innovation makes a resource more efficient:
Effective cost drops. You need less of the resource to produce the same unit of output, making it cheaper per unit.
Demand expands. Because it is cheaper, entirely new markets, use cases, and scales of production become financially viable. Projects that were previously too expensive to justify now get greenlit.
If demand for that resource is highly price-elastic, the explosion in new usage completely obliterates the initial efficiency savings. The result is a net increase in total resource consumption. This secondary surge is called the rebound effect.
Three Examples Across History
- 1865, Industrial Revolution: James Watt introduces a highly efficient steam engine. Instead of conserving coal, Britain's total coal consumption explodes because steam power becomes cheap enough to deploy across every major industry.
- Late 1900s, Digital Age: Computers make editing and formatting text highly efficient. Instead of the predicted paperless office, global paper consumption tripled between 1980 and 2000 because printing became completely frictionless.
- 2020s, Artificial Intelligence: Large language models drastically reduce the cost and time required to generate code. Instead of hiring fewer developers, firms scale up. Teams build massive systems and deploy energy-intensive models for billions of micro-tasks that previously would not have justified the compute cost, spiking data center power demand.
The Jevons Paradox is a critical warning for environmental policy: technological efficiency alone does not solve resource depletion. Without constraints or alternative cost mechanisms, optimization acts as an accelerant for growth, not a brake.
What This Means for Developers
The fear that AI will replace developers comes from a linear assumption: if AI makes coding 10 times faster, we will need 10 times fewer developers.
The Jevons Paradox shows why the opposite happens when demand is highly elastic.
The resource is developer hours. The efficiency gain is AI assistance, which reduces the time and cost required to generate syntactically correct code to near zero. The rebound effect is that software creation is now cheap, fast, and accessible, so the global demand for custom software is exploding. Projects that were too expensive, too niche, or too complex to justify now get built.
We are moving from an era of software scarcity to an era of software abundance. Organizations are not downsizing engineering teams. They are scaling the complexity, volume, and scope of what they build.
The Shift: From Syntax to Architecture
The paradox guarantees a surge in total software production, but it fundamentally changes what a developer does. Value has moved up the stack.
Skills that are losing value: memorizing syntax, writing boilerplate, configuring basic APIs, manual debugging. AI handles this instantly.
Skills gaining value: system architecture, database modeling, secure agentic workflows, business logic translation, data orchestration.
You no longer need to spend years mastering the mechanics of writing code before building a product. AI handles the operational work, letting a junior developer today operate with the systemic leverage of a senior developer from five years ago.
The New Bottlenecks
- Context and state management: keeping massive codebases coherent as AI generates more of them.
- Review and verification: code generation is cheap, but code maintenance is expensive. Someone has to own the output.
- Integration and deployment: connecting disparate APIs, environments, and infrastructure seamlessly.
- Business logic translation: understanding what needs to be built well enough to direct the tools that build it.
"The developers winning right now are acting as directors, architects, and systems thinkers rather than line-by-line writers. If you enter the field now with an architecture-first mindset, you skip years of syntax grinding and move straight to building complex, high-leverage systems."
Does it make sense to become a developer now? Yes. The Jevons Paradox explains exactly why. The career is not dying. It is shifting its bottleneck from syntax execution to systems thinking. The market does not need fewer developers. It needs developers who know how to direct an infinite supply of automated coding capacity.
The economic concepts your industry is not explaining to you.
Clear frameworks for professionals navigating AI's impact on their work. No hype, no hustle. One email a week.
SubscribeNo spam. Unsubscribe anytime.