The AI era is increasing the value of engineering judgment
Generating code is becoming the easy part. The hard part is trusting what gets generated—and deciding what is worth building.
The AI era isn’t reducing the value of engineers. It’s increasing the value of engineering judgment.
Our job as developers is no longer primarily writing code. It’s building systems that can produce software with confidence.
Generating code is becoming the easy part. The hard part is trusting what gets generated. That means we must obsess over software quality like never before:
- How do we test as fast as we ship?
- How do we know our verification strategy actually works?
- How do we produce evidence that every change is safe to merge?
This shift also changes what we should expect from QA. We no longer want QA engineers spending most of their time manually checking buttons. We want them focused on what the product fundamentally delivers for users:
- Does this feature actually make someone’s life easier?
- Does the workflow make sense?
- Is the product enjoyable to use?
We now have armies of AI agents that can run regression suites, execute end-to-end tests, verify functionality, and even critique UX and user journeys.
Human judgment remains our scarcest and most valuable resource. Let’s spend it where it matters most: empathy, product intuition, understanding customer problems, questioning assumptions, and making better trade-offs.
I don’t think we humans are less valuable now; the leverage has simply moved.