Podcast: AI won’t replace developers. But it will change what makes a great one

As AI becomes more capable of handling specific engineering tasks, developers may spend less time writing code themselves. Instead, they may focus more on defining intent, designing solutions, setting guardrails, and evaluating whether AI-generated results actually solve the problem.

From writing code to directing AI

That does not make engineering knowledge less important. Quite the opposite. Developers still need to understand what good software looks like if they want to judge the output of an AI model.

What changes is where they spend their time.

Alexander Zwart expects skills such as defining requirements, setting security guardrails and critically assessing AI-generated outcomes to become increasingly important. Working with AI is also unlikely to be a simple one-prompt exercise. The real value comes from interaction: questioning the output, refining your instructions and using AI to sharpen your own thinking.

“I really believe in enhancing the human and having human-centric AI, rather than AI replacing us.”

Build, learn and adapt

At Rabobank, agentic software development has moved beyond the experimental stage. The focus is on further developing this approach, with security and risk considerations embedded throughout the process. Learning and adapting remain essential to that ongoing development.

That does not mean every AI initiative will succeed. Alexander Zwart argues that success is about learning quickly, recognizing what works and what does not work, and to act early on those insights.

His broader message is simple: stay curious. The technology is developing quickly, and nobody knows exactly what software engineering will look like five or ten years from now. What matters is the ability to keep learning and adapting.

Want to hear more about human-centric AI, agentic software development and the skills tomorrow’s engineers will need? Listen to the full conversation with Alexander Zwart in our podcast.