AI has fundamentally changed software development — but not in the way many people think.
Writing code is no longer the biggest bottleneck. Understanding the product, making good architectural decisions, and defining the right problems to solve have become far more valuable than producing another hundred lines of code.
AI is exceptional at accelerating implementation and helping teams validate ideas quickly. That's why building prototypes or internal tools has never been easier. But there's also a downside: many people now assume that if AI can generate code, it can build production software on its own. In reality, the cost of generating code has dropped dramatically, while the cost of fixing poorly thought-out systems months later has become even higher.
That's why I believe in an engineering-first approach to AI. We use it every day, and I expect AI to become as essential — and as invisible — as electricity. The competitive advantage won't come from simply using AI. It will come from knowing where to trust it, where to challenge it, and how to combine it with solid engineering judgment.
