Key terms and definitions used across essays, posts, and systems. Terms are automatically linked throughout the site.
A higher-order software engineering discipline where AI agents do the implementation and the engineer owns architecture, correctness, and the reliability of the agent loop itself — designing the systems, standards, and verification layers that govern what software agents build and whether humans can understand and own the result.
A system whose internal workings are hidden or incomprehensible to its operators — you can see inputs and outputs, but not how or why decisions are made inside.
The total amount of mental effort required to understand or operate a system. Lower cognitive load means easier comprehension and fewer errors.
The maximum amount of text (measured in tokens) that a large language model can process in a single session — effectively the AI's working memory limit.
A measure of the number of independent paths through a piece of code. Higher values mean more branches, more test cases required, and more cognitive effort to understand. A proxy for how hard a function is to test and safely change.
The degree to which a system's structure and naming reflect the real-world domain it models, making it readable by anyone who understands the business — not just engineers.
A software design approach where the system's structure and language are modeled around the real-world business domain it serves, rather than technical abstractions.
A data architecture pattern where every change is stored as an immutable event in sequence, rather than overwriting current state — enabling full auditability and time-travel through system history.
The degree to which you — or anyone working with you — can actually see and understand a complex system well enough to trust it, improve it, and change it without fear.
A holistic approach to analysis that focuses on how components interrelate and work together over time within larger systems.
The ability to understand not just a system's current state, but how and why it reached that state — through auditable history of events and decisions over time.
A shared vocabulary used consistently between code and business stakeholders, so that terms in the system mean the same thing to everyone — from engineers to product managers.
A software structure where each feature is organized as a self-contained slice through all layers of the stack (UI to data), rather than grouping code by technical layer.