Are We All Prompt Engineers? How AI Changed What It Means to Build Software

Writing code used to be the bottleneck. It isn’t anymore. This talk is not about ChatGPT or prompt tricks. It’s about what happens to software engineering when AI can generate most of the code — and where the real value of an engineer moves to. Drawing on real experience shipping production systems with AI-assisted workflows, we’ll trace how the engineer’s role has evolved from writing every line by hand, to the framework era, to today’s AI-native era. We’ll look honestly at both sides: the genuine productivity gains (more time for design, judgment, and decisions) and the hidden cost almost nobody talks about — skill erosion, and the danger of confidently shipping code you no longer fully understand. Through a personal story of a backend system that AI kept getting subtly, dangerously wrong, we’ll explore why judgment, architecture, and systems thinking matter more now, not less — and why “prompt engineer” is the wrong mental model for the future. The conclusion: we are not becoming prompt engineers. We are becoming AI-native engineers — people who direct intelligence, human and artificial, toward problems worth solving, and stay accountable for the result. You’ll leave with a clear, practical sense of which skills to invest in next.