The knowledge graph got too complicated over the last two years — so much so that I stopped sticking to it. I removed some categories to simplify.
Fragments could potentially go next, but I still capture so many small things. The question is whether there needs to be a private vs. public layer between them. Maybe they'll both just be logs in the end.
Fragments → Logs → Journals → Artifacts → Spaces
Flow:
| Level | Name | Density | Primary Function | Example Sources | Children / Inputs |
|---|---|---|---|---|---|
| 0 | Fragments | 🔹 Lightest | Capture small sparks — quotes, links, screenshots, single thoughts, observations | Web snippets, books, podcasts, conversations | None (atomic) |
| 1 | Logs | 🔹 Low-medium | Track your own activity or process. Documentation of builds, configs, debugging, studio notes | "Got vector search working," "MCP server setup" | Fragments |
| 2 | Journals | 🔸 Medium | Reflective or integrative writing — lessons, patterns, synthesis over time | "What I learned about sensor APIs," "Why physical AI needs ground truth" | Fragments, Logs |
| 3 | Artifacts | 🔸 High | Discrete outputs — code, essays, prototypes, research docs, designs. Has shape and edges. | Essays, repos, demos, slide decks, reports | Journals, Logs, Fragments |
| 4 | Spaces | 🔸 Meta-level | Ongoing domains of practice that hold multiple artifacts. Context, not container. | "gndclouds platform," "sustainable making," "design engineering practice" | Artifacts, Journals |