Decision: CODEX Method Over PARA
Context
After 3 years of using PARA (Projects, Areas, Resources, Archives) for personal knowledge management, the system kept breaking every time I switched tools. Notion → Obsidian → back to Notion → back to Obsidian. Notes were scattered. AI agents couldn’t make sense of the structure. And PARA has no execution layer — it organizes files but doesn’t tell you (or your AI) what to do with them.
Needed: a framework that survives tool changes, works for both humans AND AI, and covers the full lifecycle from capture to execution.
Options Considered
-
Stay with PARA — Proven, widely adopted, huge community
- Pros: Simple 4-folder structure, lots of tutorials
- Cons: No AI support, no execution model, breaks on tool migration
-
GTD + Zettelkasten hybrid — Task management meets idea networks
- Pros: Good for both tasks and knowledge
- Cons: Complex, no AI layer, two systems to maintain
-
Build CODEX — New framework designed for the AI era
- Pros: AI-native, includes execution + metrics, tool-agnostic (plain markdown)
- Cons: Unproven, requires building from scratch, no community yet
Decision
Build CODEX. The existing frameworks were designed for humans working alone. With AI agents becoming co-pilots in daily life, we need a framework that gives AI context, structure, and permission to act.
Rationale
The key insight: PARA and GTD stop at “organize.” CODEX goes to “Execute” (AI takes action) and “eXamine” (feedback loop). Plain markdown means it survives any tool migration. The 4-quadrant, 100-domain numbering system gives both humans and AI a shared addressing scheme for all of life.
Consequences
- Built the One Mind Codex vault from scratch (this vault)
- Migrated all PARA-organized notes into the CODEX quadrant structure
- Started building AI agent integrations that read/write the codex directly
Review Date
2026-06-15 — Check if the system has held up through daily use