The moment we hand more of our code to AI, the shape of our work changes. The easier it becomes to generate code, the more valuable it is to decide what that code should do, a counterintuitive case that lands exactly right in Software Design in the Age of AI. We are not moving into a world of less design. We are moving into a world where design is the only thing that separates a useful tool from a pile of plausible-looking fragments. That is not a small shift. It is the entire job description, rewritten.
For anyone who has spent years wrestling with spreadsheets, the parallel is immediate. Traditional spreadsheets reward patience with structure. You learn where the formulas live, which cells depend on which, and how to trace a broken reference back to its source. AI-native tools change the bargain. They can write the formula, suggest the layout, and even anticipate the next column you need. But they cannot know what question you are actually trying to answer. That act of definition, of choosing the right frame before the tool starts generating, is design. And it is precisely where the value now concentrates. If you are using AI to skip straight to output, you are missing the point. The output was never the hard part. The hard part is deciding what output matters.
What this means in practical terms is that our editorial stance is not about praising the technology or fearing it. It is about redirecting attention to the craft that remains. AI coding makes software design more important, and we would extend that logic directly to your workflow. When you open a new spreadsheet or a new script, resist the urge to ask the AI for the answer. Ask it to help you understand the problem first. Ask it to show you what you are assuming, what you are leaving out, what the shape of the solution should be before you commit to any code. That is a habit, not a feature. It is also the difference between using a tool and being used by it.
Our honest take is that this is a discipline problem, not a technical one. The tools will keep improving, and the barrier to entry will keep dropping. What will not improve on its own is your judgment about what to build. We would tell any reader who asks us directly: treat the AI as a brilliant intern who needs clear direction, not as an oracle. The moment you stop designing, you start debugging your own blind spots. The specific consequence to watch is this: as code generation becomes commoditized, the people who thrive will be the ones who can articulate the why behind the what. If you cannot explain the design of your data model or your spreadsheet logic in plain sentences, the AI will not save you. It will just make your mistakes faster. That is the open question we are sitting with, and it is the one worth taking seriously.