F68 AI

Library · Chapter 02

Context & Knowledge Base

Organizing knowledge so the AI understands the work correctly.

The model provides capability.

Context and Knowledge determine what foundation of understanding the AI is using that capability on.

A good Agent doesn't need to know everything.

It needs the right knowledge, the right structure and the right timing.

  1. 4.1
    UNDERSTANDING CONTEXT AND KNOWLEDGE BASEContext is all the information the AI has at the moment it processes a request.
    1 min read · Read →
  2. 4.2
    IN CONTEXT: DISTINGUISHING KNOWLEDGE AND DATASystem Prompt dictates how the AI must work.
    1 min read · Read →
  3. 4.3
    DESIGNING THE KNOWLEDGE BASE STRUCTUREIf so, that content may be suitable for the KB.
    1 min read · Read →
  4. 4.4
    WRITING KNOWLEDGE THE AI CAN EASILY USEIf needed, apply the old way.
    1 min read · Read →
  5. 4.5
    MANAGING THE KNOWLEDGE BASEA principle should not exist in multiple files written in different ways.
    1 min read · Read →
  6. 4.6
    RAW CHAT, LOGS AND EXPERIENCERemember to be careful with this case next time.
    1 min read · Read →
  7. 4.7
    TESTING THE KNOWLEDGE BASEHaving a KB does not mean the Agent will use the KB correctly.
    1 min read · Read →
  8. 4.8
    WHEN THE AGENT USES KNOWLEDGE INCORRECTLYNot there → add the Knowledge.
    1 min read · Read →
  9. 4.9
    A SIMPLE WAY TO STARTYou don't need to build a complex Knowledge system from the start.
    1 min read · Read →
  10. 4.10
    KNOWLEDGE BASE CHECKLISTIf the role of a file hasn't been determined, it shouldn't go into the KB yet.
    1 min read · Read →
  11. 4.11
    GUIDE TO DESIGNING A SYSTEM PROMPTSystemP is an execution orchestrator, NOT a summary of knowledge or a mini KB.
    9 min read · Read →
  12. 4.12
    GUIDE TO DESIGNING A KNOWLEDGE BASEIntended for: AI or people who design, evaluate and improve the KB for new or operating agents.
    21 min read · Read →