There has been a longstanding misunderstanding about software engineering. Some would say that the actual writing of the code takes all the time. Typing has never been the bottleneck. Sure, you might need to learn the syntax of a new language, or understand some of the available APIs (or how to look them up), but the effort that takes the time wasn’t the actual writing of the code.
Instead, it was AND is understanding the problem, figuring out the best way to tackle that problem and then choosing the right tools, frameworks and approaches. Writing the code was just the last mile. Some engineers would start writing the code first & they’d often have to redo work, or on code reviews we’d find all sorts of issues. The work is in understanding the problem BEFORE you write a single line of code.
AI coding tools don’t change that dynamic; they simply remove the typing from the engineer’s workload. So when organizations claim that AI is making their engineers 50-70% more productive, I’m wondering how much those organizations misunderstand what engineers do.
It’s the equivalent of claiming that a chef can be replaced by a tool that automatically heats food. We don’t value chefs for their ability to heat food; rather, we value them for knowing which ingredients to choose, how to prepare them, and the effect that will have on the final dish.
Why AI Tools Don’t Replace Software Engineers
