In 1865, economist William Stanley Jevons made a counterintuitive observation that would bear his name for generations: as steam engines became more efficient at burning coal, Britain's coal consumption didn't decrease — it exploded. More efficient coal meant cheaper power, which meant more applications for steam engines, which meant more coal burned overall. This phenomenon, known as the Jevons Paradox, is now playing out in an unexpected arena: software development.
AI coding assistants have fundamentally altered the economics of building software. What once required a team of developers working for weeks can now be accomplished by a single developer in days. GitHub Copilot, Claude, ChatGPT and other AI tools have compressed the timeline from concept to working code by an order of magnitude. The natural assumption would be that companies need fewer developers.
At veeone, we believe the opposite will happen: that we're about to see an explosion in the volume of software being created, and that this will increase rather than decrease the demand for experienced developers.
The build vs. buy equation has flipped
For decades, enterprise software decisions followed a predictable pattern. Engineering time was expensive and scarce, so companies defaulted to off-the-shelf solutions even when they were imperfect fits. The calculus was simple: could we justify pulling developers off high-priority work to build a custom internal tool? Usually, the answer was no. Better to pay for Salesforce, Jira, or whatever generic solution existed and train people to work around its limitations.
AI-assisted development has inverted this logic. When a developer can prototype a perfectly-tailored solution in a day, the ongoing friction cost of working around a generic tool's limitations suddenly exceeds the cost of building something custom. That weird data transformation your team does manually every week? Previously not worth building automation for. Now it's a two-hour project. That internal dashboard that would make three teams more efficient? Used to be a month of work. Now it's achievable in a week.
We believe the result won't be companies building the same amount of software with fewer people. Instead, they'll build exponentially more software with the same teams — or discover they need even larger teams to capitalise on this new efficiency.
The new bottleneck: judgment at scale
If anything can be built quickly, the critical question shifts from "can we build this?" to "should we build this?" This is a harder problem than it appears. When creation was expensive, the high barrier to entry naturally filtered out marginal ideas. Now that barrier has collapsed, companies will need new decision-making frameworks.
We're already seeing organisations grapple with governance questions they never had to answer before. Who approves new internal tools? What standards must they meet? How do you prevent fifteen different teams from building fifteen variations of the same thing because nobody knew the others existed?
Technical debt presents another challenge. AI makes creation cheap, but maintenance costs compound the same way they always have. Every tool needs updates, security patches, and eventually needs to be sunset when requirements change. When you have five mission-critical systems, you can dedicate resources to maintaining them properly. When you have fifty micro-tools, maintenance overhead becomes unwieldy.
The companies handling this transition best are those treating AI-assisted development not as a way to reduce their engineering teams, but as a force multiplier that reveals how much latent demand for software has always existed.
What this means for agencies and development partners
For software development agencies like veeone, the Jevons Paradox presents what might seem like a threat at first glance: if clients can build more with internal teams, why hire external help? But we believe the data will show the opposite is happening.
The explosion in software creation is revealing gaps in most organisations' capabilities. Yes, their developers can now build more, but they often lack the experience to build it right. The agencies that will thrive in this environment won't compete on speed — AI has commoditised fast implementation. They'll compete on judgment, architecture, and the kind of hard-won experience that helps companies avoid building themselves into a corner. When a client can prototype in a day, the value shifts to ensuring that prototype can evolve into production-grade software without accruing crippling technical debt.
The paradox points forward
The Jevons Paradox teaches us that making something more efficient doesn't reduce its usage — it expands the contexts in which it makes sense to use it. This has profound implications for how we think about the future of software development.
We're not heading toward a world with fewer developers. We believe we're heading toward a world where software becomes the default solution to more classes of problems. Problems that were previously "not worth solving with software" will become "let's build something for this." The bottleneck will shift from development capacity to judgment, architecture, and the strategic thinking required to navigate an environment of abundance.
The companies and developers who thrive will be those who recognise that AI hasn't made software development less valuable — it's made the ability to make good decisions about software development more valuable than ever. The question isn't whether you can build something. It's whether you should, how it fits into your broader systems, and how you'll maintain it over time.
At veeone, we're positioning ourselves for this shift — we've aligned our expertise around not just creating software quickly, but helping clients navigate the strategic and architectural challenges of an AI-assisted development world. Get in touch to talk through your internal apps challenges.