How Efficiency in Software Development Could Fuel Increased Demand

For the past two years, I have been leading a team in building an AI platform, gaining firsthand insight into the evolving role of AI in software development. While AI-driven automation has improved efficiency and streamlined workflows, it has also introduced new challenges and unexpected consequences. One of the most striking realizations I’ve had is that efficiency gains can paradoxically increase demand—a phenomenon known as the Jevons Paradox.

What is the Jevons Paradox?

First observed by economist William Stanley Jevons in the 19th century, this paradox suggests that efficiency improvements often lead to greater, not lesser, resource usage. Historically, this has been seen in:

• Coal and Steam Engines: More efficient steam engines led to a boom in coal consumption as industrial activity surged.
• Lighting Innovations: Cheaper, more efficient lighting (from gas lamps to LEDs) resulted in increased overall light usage.
• Automobile Fuel Efficiency: More fuel-efficient cars encouraged more driving, counteracting fuel savings.

But what happens when we apply it to AI in software development? Could making programming more efficient trigger an explosion in demand for software and computational resources, ultimately increasing the demand for software engineers? My experience suggests that the answer is yes.

AI's Role in Software Development

AI is reshaping software development by automating repetitive tasks, accelerating workflows, and reducing the time required to build applications. Here are a few real-world examples:

• Automated Coding: AI-assisted coding tools help developers generate boilerplate code, debug issues, and streamline routine coding tasks.
• Low-Code & No-Code Platforms: AI-powered platforms enable non-programmers to create applications, reducing barriers to software development.
• Automated Testing & Deployment: AI-driven DevOps tools enhance software deployment and maintenance, making release cycles faster and more efficient.
• Automated Documentation: AI can generate documentation from code, aiding developers in understanding how to solve problems or reverse-engineer legacy systems.
• Automated Generation for Development Specifications: AI can generate user stories, acceptance criteria, or technical detailing, further streamlining the development process.

These efficiencies significantly reduce development costs and time-to-market. But does this mean a net decrease in demand for software development? In my experience, it does not.

It Could Lower Demand—But the Reality Says Otherwise

AI's efficiency gains should theoretically lead to a lower overall demand for software engineers. However, indications suggest otherwise. Instead of reducing software production, AI’s ability to streamline development could trigger an explosion in demand.

We know that:

• AI automates coding tasks, making development faster and more accessible.
• Businesses can now build and deploy software with fewer resources.

Reduced costs lead to increased demand:

• More software creation: Cheaper and easier development encourages more applications to be built, as ROI considerations shift in many cases.
• Expanded project scope: Developers can create more ambitious, complex applications. We are seeing only the new tools, not the new problems and consumer behaviors that will emerge.
• Greater accessibility: Non-developers are empowered to build software, further expanding the ecosystem.

Increased demand for computing resources:

• AI-driven coding could lead to more frequent software updates, requiring greater cloud storage and computing power, which in turn demands more engineers to work on foundational components.
• More applications mean higher energy consumption in data centers, consequently requiring more engineers to manage these expanding capabilities.

To support my theory that this is a plausible scenario, I present four pieces of evidence that suggest this alternative is viable and, in my view, serve as leading indicators of what I believe will be the reality in the coming years.

1. The Bureau of Labor Statistics supports this perspective, projecting a 17% increase in software engineering jobs (approximately 328,000 new roles) by 2033—significantly faster than the average job growth rate.
2. The Pragmatic Engineer recently released an article exploring the state of the software engineering job market. While it indicates a current slowdown, this is not attributed to AI. There is short-term job growth despite significant advancements in AI agents and tools for software engineers.
3. Throughout the history of software engineering, we have seen several transformational events that brought radical productivity gains, such as the rise of modern programming languages (C, C++, Java, etc.), the introduction and adoption of agile methodologies, and even the internet. Each of these advancements enabled engineers to achieve new levels of productivity by providing better tools.
4. From my own experience, I’ve been fortunate to work at CI&T with a platform that accelerates software development—CI&T Flow. It has been widely adopted across the company, reaching an 85% adoption rate. In most of our client engagements, where we achieve real productivity gains, the conversation isn’t about reducing team sizes but about leveraging newfound productivity to drive business growth.

The data supports the hypothesis that making custom software creation cheaper, faster, and easier will lead to a significant increase in demand for solutions—even when considering the new demands and experiences that will inevitably emerge.

Preparing for the Shift

While AI-driven efficiency in software development presents both opportunities and challenges, individuals and organizations can take proactive steps to navigate this transformation. Here’s how I am approaching this unique moment in time—perhaps it may serve as inspiration for others:

As an Individual:

• Leverage AI as a Tool, Not a Replacement
  • AI-powered tools like GitHub Copilot, ChatGPT, and CI&T Flow should enhance your skills, not replace your problem-solving ability.
  • Focus on understanding the logic and architecture behind AI-generated code rather than relying on it blindly.
• Prioritize Code Quality Over Quantity
  • AI makes it easier to produce software, but that doesn’t mean all software should be produced.
  • Adopt best practices for writing efficient, maintainable, and secure code to prevent unnecessary complexity and waste.
• Be Mindful of Computational Costs
  • AI-driven development increases cloud computing usage.
  • Optimize code for performance and efficiency to reduce unnecessary server load and energy consumption.

• Develop Expertise in High-Value Areas
  • AI excels at repetitive tasks but lacks creativity and critical problem-solving abilities.
  • Focus on high-level skills like system architecture, software security, and AI governance to stay relevant.

As a Leader in My Organization

• Strategic AI Adoption
  • Ensure that AI is used to address meaningful problems rather than being treated as a mere novelty.
  • Establish guidelines for where and when AI should be integrated into development workflows.
• Drive Sustainable Development Practices
  • Encourage teams to optimize AI models and coding practices to minimize unnecessary resource consumption, ensuring efficiency and sustainability.
• Purposefully Reinvest Productivity Gains
  • AI will boost productivity—leaders must establish clear strategies to capture and reinvest these gains to enhance company growth.
  • Create a space that allows engineers to focus on higher-order problem-solving and innovation.
• Invest in Upskilling & Education
  • Provide opportunities for engineers to expand their expertise in AI integration, cloud optimization, and advanced system design.
  • Encourage learning programs on AI ethics, security, and sustainability.
• Advocate for Thoughtful AI Governance
  • Actively participate in AI governance discussions to shape responsible AI policies and ethical frameworks.
  • Ensure that company policies align with industry best practices for responsible AI usage.

The Future is in Our Hands

AI in software development is no longer science fiction; it’s a reality driving significant productivity gains and fundamentally reshaping how work is done. We are in the midst of this transformation, and there are clear paths ahead for how we can leverage the technology (at least for now).

---