When AI coding assistants and agentic tools started taking over software development, the initial industry panic was predictable. Headlines warned that programming was dying, software engineering teams would shrink to zero, and writing code would soon be obsolete.
Yet here we are in 2026, and something very different is happening.
Atlassian CEO Mike Cannon-Brookes captured this reality best in a recent tech industry report: "As the cost of building technology goes down, the amount of technology we build goes up."
In economics, this phenomenon is known as the Jevons Paradox. When a resource becomes more efficient to produce, consumption doesn't decrease, it explodes. The same thing is happening to software engineering today.
From 0 to 2 Websites in a Few Days
I experienced this firsthand just a few weeks ago.
For 16 years, I worked in IT without launching a personal portfolio. Building a custom website from scratch always felt like a project that required dozens of hours I simply couldn't justify spending. But once I used AI agents to handle the initial build of desaputro.com, the barrier to entry collapsed.
Instead of stopping there, the low friction inspired me to keep going. Within days, I launched a second site—an e-commerce game top-up store called melogaming.com.
Before AI tools, launching two distinct digital platforms in a single week would have been unthinkable for a solo developer working alongside a full-time job. Lowering the cost of building didn't make me build less; it freed me to build more.
Lowering Friction Unlocks Backlogged Ideas
Every company and developer has a backlog of ideas that never see the light of day. We leave side projects unfinished, push back internal developer tools, and skip building custom automation scripts because the ROI on developer hours doesn't justify the effort.
When AI tools reduce the time required to build a feature from three days to three hours, those backlogged ideas suddenly become viable:
Internal Tools: Teams can build custom dashboards, data pipelines, and internal utility tools that were previously deprioritized.
Side Experiments: Solo developers can test digital products and e-commerce ideas rapidly without high upfront risk.
Refactoring Legacy Code: Upgrading legacy systems becomes feasible when AI agents handle monotonous syntax translations.
Instead of reducing developer demand, cheaper code creation expands the scope of what we can build.
The Role Escalation: From Typist to Product Builder
AI is not eliminating developers, but it is aggressively shifting what it means to be one.
When code generation becomes trivial, the bottleneck shifts upstream. The challenge is no longer writing the function; it is knowing what software needs to exist, how the system components fit together, and how to deliver a smooth user experience.
Developers who adapt aren't competing with AI to write syntax faster. They are leveraging AI to act as product builders, system architects, and technical directors.
Final Thoughts
The idea that AI would kill software engineering assumed that the world's demand for software was fixed. But software demand is endless. As long as there are business problems to solve, user experiences to refine, and digital ideas to explore, making code easier to write will only accelerate human creativity.
We aren't building less technology. We are just getting started.
Until then, see you at the top!
