In the system prompting guide I shared, the key point is this: given how powerful current LLMs and agentic systems are, their effectiveness hinges on how well you encode context, nuance, and directional intent. A well-constructed prompt can even shape behaviors like temperature implicitly. In that sense, prompting is no longer just natural language input, it is an engineered architecture.

In 1999, Donella Meadows published a short paper that would become one of the most cited texts in systems thinking: Leverage Points: Places to Intervene in a System.

The central argument was deceptively simple. Every system has places where a small change produces large effects. The problem — the structural problem that makes this insight difficult to use — is that these high-leverage points are almost always the opposite of where practitioners look.

People focus on numbers: budgets, headcounts, parameters. Numbers are visible and adjustable. Adjusting them feels like an intervention. But adjusting numbers almost never changes a system’s behavior in any fundamental way.