Generating code isn't the same as operating production systems.

AI can generate a working feature in minutes.

It can create APIs, write database queries, generate tests, and even suggest fixes when something fails.

But a working feature is not the same as a production-ready system.

Production introduces a different set of problems:
- What happens when traffic suddenly increases?
- What happens when a database query works with 100 records but struggles with 10 million?
- What happens when a background job gets retried thousands of times?
- What happens when an external AI service becomes slow or unavailable?
- what happens when the infrastructure bill grows faster than the customer base? 

These aren't just coding problems.

They're scaling and operational trade-offs.

AI can help engineers solve many of them, but someone still needs to understand the system well enough to identify the right problem and validate the solution.

This is where I think the role of engineering is changing.

We may write less code manually.

But we need to understand the systems we're building even better.

Because production doesn't care who wrote the code.

It cares whether the system can handle the load.