AI Got Better While I Was Away. Software Didn't.

In a reflective piece on the state of modern programming, the author argues that while AI tools have significantly improved their ability to generate code, the quality of software engineering has not kept pace. The article contends that the industry remains plagued by overengineering, unnecessary complexity, and a focus on implementation speed rather than problem-solving. The author suggests that coding itself is becoming a commodity, shifting the value of a developer toward judgment, architectural taste, and the ability to simplify systems. Rather than celebrating the sheer volume of code AI can produce, the author warns that developers must become better at filtering these outputs to avoid creating massive piles of mediocre, complex software. Ultimately, the piece emphasizes that the core challenge of software development remains making sound decisions under uncertainty, rather than simply typing out code faster.