The highest-leverage engineering work often produces no visible output.
It changes how future decisions are made.
Two senior engineers spend an hour debating architecture. At the end, they decide not to build a new service.
No pull request appears. No deployment occurs. No release notes are written.
From most dashboards, nothing happened.
In reality, they may have saved years of operational complexity.
Engineering dashboards celebrate activity: deployments, commits, tickets, incidents.
They rarely measure systems never created, risky migrations wisely avoided, abstractions rejected, or dependencies kept out of production.
Those decisions disappear into history even though they may create enormous value.
Early-career engineers often build. Experienced engineers increasingly simplify.
They ask whether a service can disappear, whether two systems can become one, whether a workflow can be removed.
The result is less software and more capability.
When AI can generate endless implementations, discernment becomes scarce.
Which ideas deserve to exist? Which should never be built?
That is engineering judgment.
Instead of asking how much software we produced, ask:
> How much unnecessary engineering did we permanently eliminate?
The answer may reveal the real leverage in the organization.