I remember in the early 2000s reading a book called "My Job Went to India and All I Got Was This Lousy Book."

On the cover, the author, Chad Fowler, had an Indian guy holding a sign that read, "Will code for food."

It's hard to remember the book verbatim, but Fowler described how software development jobs were moving to India at the time, and the challenges people in the US were facing with that transition: time zones, language barriers, cultural differences, and communication.

We are in a similar situation with AI.

And if there's one thing AI is very good at, it's writing code. It can generate thousands of lines of code in seconds.

Generating code has never been easier.

But one of Fowler's arguments, in a book that's now almost 20 years old, was that outsourcing syntax writing to people who don't understand the architecture, the design, and sometimes the culture and language of the organization creates a very different set of challenges.

Writing software isn't just churning out syntax. It requires a mental model and an understanding of what you're actually building.

The same thing applies to AI agents.

If you simply prompt an AI agent to build something without giving it a clear architecture, specific requirements, and guardrails, the agent will probably produce something.

But who knows what you've actually built?

It's not surprising that people who vibe code entire projects often either don't finish them or abandon them because the code becomes difficult to maintain.

I think that's one of the biggest traps of vibe coding: maintainability.

So yes, AI will write more and more of the code from now on. And that code will become cheaper and cheaper to produce.

But what won't go away, and may actually become more valuable, is the person who knows how to build software structurally.

The person who understands the architecture.

The person who can see the system in their head before asking the AI to build it.

AI may replace the person writing the syntax.

I'm not convinced it will replace the person who knows what should be built in the first place.