Topics

Atomic reference pages, organized into clusters by area. Each cluster is a folder under topics/, and the folder note inside it (the file named for the folder) serves as its map of content. Clusters that share a theme nest under a parent folder, so the AI clusters sit together under topics/ai/.

The clusters span the three pillars of the vault, sometimes individually and sometimes across multiple. Pillar membership is carried by tags on each page, not by where the page lives.

Clusters

• AI: building software on top of language models, across prompt design, context curation, and agentic systems
• Caching: keeping fast copies of data, and the read, write, eviction, and invalidation strategies that keep a copy worth serving
• Databases: how a data store keeps records correct and reachable, from ACID guarantees down to the index structures underneath
• Distributed Systems: reasoning about a service that runs across many machines on a network that drops and delays messages
• Notation: how notation encodes structure through precedence and grouping, and where implicit precedence turns into a reading risk
• Python: the language as a system to reason about, from how the interpreter runs your code to how environments and dependencies stay isolated

More clusters land as the vault grows.