A necessary question for engineering leaders

A necessary question for engineering leaders: what happens when AI increases output but leaves your engineers exhausted?

This is a real problem. So is the uncomfortable fact that there is no pre-AI version of software engineering to return to.

Leaders need to hold both truths at once. The work is changing, and people who found meaning in building software themselves may genuinely lose part of what made the profession rewarding. Treating that as simple resistance to change misses both the human and organisational risk.

Three leadership implications follow.

First, AI does not make engineering judgment, mentorship or capability development optional. Senior engineers are still needed to challenge outputs and develop juniors. If AI adoption removes the path by which juniors become seniors, the organisation is consuming its future capability.

Second, engineering was never only about writing code. Architecture, testing, review, documentation and operational decisions have always mattered. But it would be disingenuous to pretend that reviewing a growing volume of machine-generated output is the same craft, or creates the same satisfaction.

Third, resilience requires independent competence. When AI produces a plausible but faulty solution, the team must still understand the system well enough to detect the failure, diagnose it and recover. “Ask the AI again” is not an engineering fallback.

The question is not whether to adopt AI. It is whether the operating model still develops engineers who can exercise judgment when AI is wrong.

Are we increasing delivery capacity, or quietly eroding the human capability that makes that capacity safe?