The oldest trick, with a new reason
Split responsibilities into layers so humans and agents can reason about each layer independently.
The best way to solve a big problem is to turn it into a sum of small ones.
The why we covered in Module 1: a small piece of problem is the only things you — or an agent — can reason about reliably and verify. One approach is to decompose the responsibilities in layers, so you can think about each layer independently from one another. Each layer should have clear responsibilities and clear dependency flow from one another. This should help reduce cognitive load for you or your agent, and concentrate the core of the complexity into a single layer.