Building Knowledge Bases / Second Brains — Historical Track Audit

This is historical evidence, preserved on 2026-09-09 from source checkpoint 718c026. The matching Astro checkpoint is 4bde108. The original outline below preserves planning inputs, not a second curriculum inventory or a commitment to build every named idea.

The current Track Guide and shared Track Map carry the approved composition: Context Engineering as signature, Context Cleanup as short entry, Build the Brief Before the Work as longer practice, three developing collection routes, and optional context experiments.

Next Artifact Pass — Evidence and Unresolved Questions

• Context Engineering remains Sketch / Planned, owned by Renaissance Solutions, primary to AI Literacy, with no related Track metadata. Its Knowledge Bases signature role is a Track composition decision, not a reassignment or status promotion.
• Context Cleanup and Build the Brief Before the Work are Beta, owned by Mike, and primary to Context Engineering. Both have canonical Activities, Notebooks, learner guides, and facilitator guides. Learning-design, accessibility, and pilot review remain open. Neither creates a maintained knowledge base.
• The module estimates 8–12 minutes for short practice and 20–30 for recommended practice, while the canonical Labs estimate 15–20 and 30–40 respectively. The Track uses Lab estimates, not invented pilot timing. Reconcile the module estimates in a subsequent artifact pass.
• Context Engineering metadata names Conversation Engineering as a dependency, while its learner page says preparation is not required. Both remain unchanged. The Track preserves the stated dependency while explaining the beginner entry; reconcile this discrepancy at the module level.
• Context Cleanup has a complete paper-first route with an optional chatbot restatement check. Build the Brief Before the Work says a chat tool is optional for comparison, but its working steps ask learners to run the chatbot, change a variable, and compare outputs. Its completion record includes both results. Do not treat sorting alone as completing that comparison.
• The signature module allows supplied examples. A facilitator needs to prepare an appropriate output comparison set for no-tool delivery of the longer Lab or optional experiments; this Track does not manufacture an existing reviewed packet or rewrite their completion criteria.
• How Context Constrains the Continuation and Contradictions in Context are Beta, primary to Context Engineering, with Notebooks and learner/facilitator materials. Their Activities use live comparison steps; required-document collections remain Coming Soon. The first Lab retains learning-design, accessibility, and pilot review; the second also needs safety review. Their estimates are 25–40 and 30–45 minutes.
• All four Labs remain related to AI Literacy only. Their Knowledge Bases placement is a candidate cross-Track relationship, not reciprocal metadata mutation. Optional experiment ownership stays with the signature module.
• The old Context Engineering Basics idea is represented by the existing Context Engineering source, not a new module. The old pile-of-documents versus curated-packet idea is served in part by the existing entry practice; it is not listed as another missing Lab.
• What Is a Knowledge Base?, Building a Second Brain, Source Quality and Provenance, Knowledge Maps and Relationships, Retrieval and Verification, Memory Types, and Maintaining Stored Memories remain proposed directions without source pages. Some may become sections or practice choices rather than separate modules.
• The three developing Lab ideas preserve the old starter-folder, five-document map, and stored-memory review directions. They remain unbuilt proposals, not settled package titles. The memory exercise should begin with fictional examples, not learner account access or private records.
• First collection gap: a small, reviewed source set, purpose, naming guidance, source records, index, guided Activity, standalone Notebook, and facilitator guide. Follow with a return visit that checks retrieval and updates rather than claiming one context brief is a maintained knowledge base.
• Memory scope must distinguish saved notes, project files, conversation context, and product memory features. Retention and deletion claims require tool-specific verification; removing a local file does not establish what a connected service retains.
• Shared support uses Conversation Engineering and Before You Paste as Sketch / Planned and Reality Check for AI as Draft / Drafting. The old “Trust, but check” dependency is addressed through verification support, not a new mandatory sequence or changes to the Seed / Planned Trust, But Check the Wires module.
• Obsidian is one possible tool, not a required installation or the default. Manual indexing and lookup come before optional automated retrieval. No product-specific tutorial, connected collection, or retrieval system is created.
• The Track remains Draft / Drafting and Unassigned. Draft links stay visible; demonstrations remain Coming Soon. Preserve the existing Additional Reading section. All module and Lab sources, statuses, ownership, dependencies, and review contracts are unchanged.

Original Track Guide Snapshot

Building Knowledge Bases / Second Brains — Original Outline

Status: Working track outline, not finished curriculum.

This track teaches learners how to organize knowledge so people and AI tools can use it. It starts with ordinary human needs: notes, documents, policies, memories, source material, and relationships. AI makes this more powerful, but it also makes organization, source quality, and verification more important.

The invitation is not to build a perfect archive. The invitation is to make knowledge easier to find, trust, connect, and use. A good knowledge base helps the human think more clearly before it helps the AI answer.

Audience: Learners with notes, documents, family history, business knowledge, school material, policies, or creative archives.

Purpose: Teach practical knowledge organization, context quality, and retrieval habits.

Likely Modules

• What Is a Knowledge Base?
• Building a Second Brain
• Memory Types
• Context Engineering Basics
• Source Quality and Provenance
• Maintaining Stored Memories
• Knowledge Maps and Relationships
• Retrieval and Verification

Starter Lab Ideas

• Build a small second-brain starter folder.
• Map how five documents relate to one another.
• Compare a pile of documents to a curated context packet.
• Review stored chatbot memories and decide what to keep, change, or delete.

Additional Reading

No additional reading is listed yet.

Dependencies

• Prompting basics.
• Trust, but check.
• Before You Paste.
• Basic file and source organization.

MVP Holes

• Decide whether Obsidian is an example, a default, or one option among many.
• Define the privacy boundary for personal memory and family-history examples.
• Decide how much retrieval technology belongs in the first version.
• Create a reusable “memory types” explainer.

---

Source: ops/track-audits/building-knowledge-bases-second-brains.md