Work is Work

Uses existing mathematical models to build upper/lower bound estimates of how the cost to perform tasks increases as you add more people (absent novel interventions to change the nature of the work), and suggests types of interventions to avoid the whole thing grinding to a halt.

This is a very nice complement to the “erikbern/headcount goals and feature factories” article. The two use similar kinds of models, but at different levels. That article points out that the addition of enough task-related overhead will prevent a hypothetical 10x developer from performing any better than a 1x developer, basically meaning your mean performance hits the floor. This article points out that the addition of people, without a tremendous effort to make the nature of work far more impactful for everyone, causes exactly such overheads to accrue.