Skip to content

jcksamuels/connected-thinking

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Connected Thinking

Obsidian vault + Claude Code harness. Built for ADHD brains.

Run claude . for a 10-minute onboarding that personalizes the vault to how you think. Notes link themselves like associative memory after that.


Why most note systems fail ADHD brains

What systems assume ADHD reality
You'll remember where you filed it Out of sight, gone
You'll apply a taxonomy consistently Decision fatigue burns working memory
You'll review on a schedule Time blindness kills voluntary routines
You can hold structure in your head Working memory caps at ~4 items
You'll stick with one system Interest shifts; rigid systems get abandoned

Result: a graveyard of half-built vaults, retried every few months.


What this fixes

  • Memory is associative, not hierarchical. Six broad folders. Real structure lives in [[wikilinks]], #tags, and Dataview queries. Notes find each other.
  • Working memory is tiny. Templates pre-decide structure so you only think about content.
  • Spaced retrieval beats willpower. Spaced Repetition plugin schedules reviews from passages in your own notes.
  • Visual + verbal beats either alone. Graph view, Canvas, and Bases give spatial encodings of the same notes.

How notes connect themselves

The vault behaves like an associative network. Notes don't sit in folders waiting to be retrieved by hand — they surface each other based on what they link to, the way related thoughts trigger each other in a brain.

The mechanism

  • Every [[wikilink]] you write creates a directed edge in the graph.

  • Subject, Definition, and Resource templates ship with an embedded Dataview query:

    list
    from [[]] and #definition
    

    [[]] resolves to the current note. Translation: list every note that links to this one and is tagged #definition.

What this means in practice

Write a Definition note about "Stoicism", tag it #definition, and link it back to a Subject note "Philosophy". The Philosophy note's embedded query auto-lists Stoicism. Forever after. No filing. No indexing. No maintenance.

Why it compounds

  • You write each connection once. The system uses it many times.
  • The graph gets more useful with size, not less — opposite of folder hierarchies, where size buries things.
  • Open any note in your active field and the relevant definitions, resources, and reflections you've ever linked to it are already on the page.

This mirrors Hebbian learning at the neural level: cells that fire together wire together. Notes you've connected once stay connected.


ADHD-specific wins

Challenge What handles it
Working memory limits Frontmatter + templates
Blank-page paralysis Pre-built fields
Object permanence Smart Connections (semantic search)
Hyperfocus tangents 0. Notes/ inbox, retrieve by topic later
Time blindness Calendar + daily notes
Decision fatigue 6 folders, not 60
Task switching One vault for everything
Restarting after a gap Conventional defaults — easy to resume

Also good for

Autism — predictable structure, pattern-friendly graph view, low-stim theme, explicit frontmatter.

Dyslexia — semantic search ignores spelling, SF Pro typography, spatial navigation via Canvas and Graph.

Tool, not treatment. Talk to a clinician if you're struggling.


What ships

Layer What
Folders 0. Notes5. Experiments
Templates Subject, Resource, Definition, Book Analysis
Theme Cupertino — calm, low-stim
Plugins Dataview, Templater, Calendar, Smart Connections, Style Settings, Spaced Repetition
AI search Smart Connections — local embeddings, optional Ollama chat
AI authoring Claude Code with 5 Obsidian-aware skills + 2 custom (onboarding, book-analyst)

Vault ships empty. Notes are yours.


vs other systems

This PARA Zettelkasten Notion
Built for ADHD Yes No No No
Pre-built templates Yes No No Yes
Local AI search Yes No No No
AI agent for editing Yes No No Chat only
Files on your disk Yes Yes Yes No
Setup time 10 min DIY DIY Locked-in

Setup

  1. git clone https://github.com/jcksamuels/connected-thinking ~/Documents/connected-thinking

  2. Obsidian → Open folder as vaultconnected-thinking/brAIn/

  3. Settings → Community plugins → turn off Restricted mode, then install all six (links below). They auto-enable once installed because the vault config already lists them as enabled.

  4. Settings → Templates and Templater → folder = Templates

  5. Settings → Appearance → Theme → Cupertino

  6. First-run onboarding (recommended)

    claude .

    On the first session, Claude reads CLAUDE.md, sees no Profile note exists, and runs a 10-minute conversation to learn how you think, what you're working on, and what you want the vault to capture. It then writes your Profile, Goals, and starter Subject notes — the seed graph from which everything else grows. Re-run any time by saying "onboard me" or /onboard.


Daily use

Folders

  • 0. Notes/ — inbox
  • 1. Education/ — subjects, definitions, resources
  • 2. Life/ — goals, inspiration, reading
  • 3. Ventures/ — work, businesses
  • 4. Archives/ — old
  • 5. Experiments/ — sandbox

AI

  • Smart Connections — find notes by meaning when you forgot the title
  • Claude Code — draft, reorganize, summarize across multiple notes

Both optional.


Customize

Rename folders, edit templates. Nothing breaks if Templates/ keeps its name.


Gitignored

.obsidian/plugins/, workspace.json, .smart-env/, .trash/, .DS_Store, per-machine Claude Code state. Configure Smart Connections (Ollama endpoint, model) on first run.


References

Cowan 2001 (working memory). Clark & Chalmers 1998 (extended mind). Ebbinghaus 1885 (forgetting curve). Paivio 1971 (dual coding).


License

MIT.

About

Obsidian vault + Claude Code harness built for ADHD brains. Pre-configured templates, local AI search, and a folder model grounded in cognitive science. A plug-and-play second brain for neurodivergent thinkers.

Topics

Resources

Stars

Watchers

Forks

Packages

 
 
 

Contributors