AI Didn't Kill Software Engineering. It Made It More Valuable Than Ever.

We’ve Been Asking the Wrong Question

For years the industry asked:

“Will AI replace software developers?”

Now we know that isn’t the right question.

The better question is:

What becomes valuable when writing code is no longer expensive?

That single shift changes software engineering.

For decades every major methodology—from Waterfall to Agile, DDD and
Clean Architecture—evolved in a world where writing software was
expensive. Documentation aged quickly because updating it was costly.
Specifications were abandoned because implementation consumed weeks.

Generative AI changed that assumption.

Writing code is now almost free.

Value moved elsewhere.

Code Is Cheap. Judgment Isn’t.

Modern LLMs generate hundreds of lines of code in seconds.

They do not decide:

• what should be built;
• which business rules matter;
• which edge cases deserve attention;
• what architectural trade-offs should survive for years.

Those remain engineering decisions.

The Bottleneck Has Moved

Typing is no longer the bottleneck.

Thinking is.

Specifications.

Architecture.

Acceptance criteria.

Code review.

Engineering has become less about producing code and more about
producing clarity.

Why Vibe Coding Doesn’t Scale

Chats are a poor long-term source of truth.

Every prompt expands the context window.

Every correction burns more tokens.

Eventually the model reasons about the conversation rather than the
software.

That is the hidden cost of vibe coding.

Specifications Become the Project Memory

That realization led me to write Spec Driven Development.

Instead of relying on conversations, every feature starts with a
specification, evolves into a plan, becomes executable tasks and only
then turns into code.

The outcome is straightforward:

• fewer ambiguities;
• fewer tokens;
• less rework;
• more predictable software.

https://books.kodel.com.br/en/books/sdd/

But Specifications Need Architecture

Great specifications cannot rescue a chaotic codebase.

Architecture exists to answer one question:

Where does this belong?

That question inspired FOCUS Architecture.

Rather than multiplying layers, it organizes software around four
responsibilities:

• View
• Orchestrator
• Use Cases
• Repository

Its purpose isn’t novelty.

Its purpose is lowering the cost of change.

In the AI era that also means smaller context windows, fewer tokens and
fewer unintended modifications.

https://books.kodel.com.br/en/books/focus/

Standing on Giants

Neither book claims to reinvent software engineering.

Both synthesize decades of work from Martin Fowler, Kent Beck, Eric
Evans, Bertrand Meyer, Robert C. Martin, Andrew Hunt, David Thomas and
many others, reinterpreting timeless ideas for AI-assisted development.

Two Books. One Goal.

Spec Driven Development structures engineering thinking.

FOCUS Architecture structures engineering systems.

One answers what should be built.

The other answers where it belongs.

Together they help human engineers and AI agents build software that
remains understandable years after it ships.

Final Thoughts

AI made code inexpensive.

Engineering became even more valuable.

The future won’t belong to teams with the best AI.

It will belong to teams with the best engineering discipline.