xandreed

Series

Reading paths

Ordered, multi-part arcs. Standalone posts live on the home page.

Building a coding agent 8 parts
  1. An agent loop is a while loop with good manners
  2. Error messages are prompt engineering now
  3. A quality bar is a workspace artifact, not a prompt
  4. A check that already passes cannot measure the work
  5. The whole permission model is one folder
  6. Pay for the index, not the book: extending an agent with markdown on disk
  7. The model role, not the model id — and why 'code' earns its own tier
  8. Retry is not a policy
Effect, from zero 2 parts
  1. Effect from zero: the beginner's guide I wanted
  2. AI applications are Effect-shaped
How an agent remembers 3 parts
  1. The transcript is the memory
  2. The context window is a resource you engineer, not a buffer you fill
  3. SQLite is the easy part; durability is the feature
Observing an agent 2 parts
  1. Observability is built in; where it goes is a layer
  2. Name spans for humans, query them by attribute