The Jevons Paradox: Why AI Is Good News for Access Developers

AI is going to put software developers out of work.

At least that's the prevailing narrative.

AI is better, faster, and cheaper than humans at producing code. Therefore, there's no need for programmers anymore. The math sounds simple enough. But as H. L. Mencken reminds us,

"There is always a well-known solution to every human problem—neat, plausible, and wrong."

AI's impact on the software development job market will be complex. And the notion that AI will replace developers entirely is, too, a prediction that's "neat, plausible, and wrong."

People of 2026, let me introduce you to 19th-century economist William Stanley Jevons.

William Stanley Jevons and The Coal Question

In 1865, Jevons published The Coal Question. At the time, Britain's industrial dominance ran on coal, and there was growing concern that the country's coal supply would not last forever.

Steam engines had been getting dramatically more efficient ever since James Watt's improvements in the late 18th century. The common assumption was that this was good news for the coal supply. If each new engine could do the same amount of work with less coal, then demand for coal would plummet. The world would simply no longer need as much of it.

The Paradox

Jevons made the famous counterintuitive observation that the opposite was true.

In his words:

"It is wholly a confusion of ideas to suppose that the economical use of fuel is equivalent to a diminished consumption. The very contrary is the truth."

As the efficiency of coal-powered engines went up, total demand for coal went up even faster. The reason is that the marginal cost of using coal dropped to a point where it now had more uses and more applications. Jobs that had never been worth doing with a steam engine suddenly were. Each engine burned less coal, but there were far more engines.

This is what we now call the Jevons paradox: making a resource more efficient to use can increase, rather than decrease, the total amount of it that gets used.

Software's Coal Moment

Now swap "coal" for "software development."

AI, and agentic AI software development in particular, is driving down the cost of producing software. The conventional wisdom says that if software gets cheaper to build, we'll need fewer people to build it.

I think Jevons would disagree. The way I think this is going to work is that there will be an overall increase in demand for software developers, precisely because the marginal cost of producing software is going to plummet.

Creative Destruction for the Big Platforms

That doesn't mean every corner of the software industry comes out ahead.

Where I think this is going to have a creative destruction-type effect on the software developer labor force is within the large platforms.

As the marginal cost of creating custom applications goes down, so does the amount of effort required to create and maintain large platforms. Something like Salesforce, Microsoft Dynamics, or any large ERP system would need fewer developers at the actual product owner company to maintain those flagship products.

Pure code-writing jobs are going away (and not coming back).

The bigger shift is that the need for those flagship products is also going to go down, because it will be so much cheaper to create fully custom software. Businesses have long bent their processes to fit off-the-shelf platforms because building something custom simply cost too much. As that changes, a lot of those businesses are going to ask why they're paying for a platform designed for everyone when they could have software designed for them.

The Rise of Small Custom Software

That brings us to where the real growth is going to happen.

Demand for fully custom software is going to go up.

Project sizes, in terms of development effort, are going to go down to the point where a larger number of smaller businesses will be able to make developing custom software economically viable at their smaller sizes. A custom system that only made sense for a company with hundreds of employees will start to make sense for a company with a few dozen, or even a handful.

That's the steam engine moment for software. Each project requires less effort, but there are going to be far more projects.

Why This Is Good News for Access Developers

This is especially good news for Access developers.

The key attribute that will make software developers valuable in the age of agentic AI software development is the ability to design systems that solve the business need. When an AI agent can write the code, the hard part is no longer writing the code. The hard part is knowing what the software needs to do in the first place.

That's a strength of most Access developers.

Translating what the business needs into software that fills that need is the job. These are outcome-based applications, if you will, where the goal is to achieve some business outcome with software. That's something that Access developers already do all day, every day.

And notice who the new customers are going to be: smaller businesses that could never justify custom software before. Those are exactly the kinds of organizations Access developers have been serving for decades.

Final Thoughts

Jevons was right about coal. Making it more efficient to use didn't make it less important; it made it essential to far more of the economy.

I believe the same thing is about to happen with software.

And when it does, the developers who understand the business need, rather than just the syntax, are going to be in higher demand than ever. That describes a lot of Access developers.

AI attributions: this post started with a stream of consciousness spoken into Wispr Flow, transcribed to text, and fed to Opus 5.5 with instructions to keep my phrasing but make the whole thing coherent. It then went through several iterations of cleanup and human rewrites of a few entire paragraphs. The result is a collaboration between man and machine with the goal of bringing to bear the best of both.