Context Type 1-2: Instructions and Examples

~12 min read

The book's CPU/RAM mental model for context engineering, then the first 2 of its 6 context types: Instructions (who/why/how) and Examples (behavioral demonstrations).

Context Type 1-2: Instructions and Examples is a Pro topic

Sign in, then upgrade to Pro or Power to unlock this topic and the full AI Engineering curriculum.

Key points

  • The book's mental model: if the LLM is a CPU, the context window is the RAM — context engineering means deliberately programming that limited working memory
  • Production agents need structure across a multi-dimensional context layer, not just one instruction embedded in a prompt line
  • Context Type 1, Instructions, covers the who (persona/role), why (goal/motivation), and how (steps/tone/format/constraints) of the agent
  • Context Type 2, Examples, shows good AND bad behavioral demonstrations — because models learn patterns much better than plain rules
  • Examples formalizes few-shot learning as one of six distinct context types, not folded generically into 'the prompt'