“Our brains compute across a knowledge graph”
This system deals with concepts, insights and ideas
The goal is to map concepts - and “not have a personal detailed Wikipedia”.
Sources: the reference (book, podcast, video, article, etc)
Atoms: existing concepts or techniques (like “an entry on a dictionary” or a “wikipedia article”)
Molecules: your ideas - crystallised insights and pieces of intuition. They are the link between atoms - so, they are analogue to “zettels” in the Zettelkasten system.
Categories: used to organize sources, atoms and molecules.
“Mine atoms out of Sources, make Molecules out of Atoms, and organize everything with Categories.”
Media or information you want to learn from
They use to have:
How to takes notes on a source: capture (briefly) the core thesis of a Source, the key concepts, and 1 or 2 of the best examples.
A source has some knowledge that is unique to the source, and other ones that exist outside of it. That is why we should extract each concept on an individual note - hence, your “atoms”.
Extract an Atom if it looks like other sources may reference to it, or if you will link ideas to that atom in the future.
You can extract an atom from a source in a future time, maybe when you are re-reading or editing your notes.
If a source only has one concept, do not create a “source note” - create an atom directly linking to the source.
Notes should mention the key concepts on a source, but not explain them - the explanation will be on an atom for the concept.
Reduce the author thesis to the simplest form - maximum 3 bullet points
A note should tell what is inside the source, not repeat its’ contents. Notes are pointers!
Use a simple and flexible structure. You can e.g. use bullets, where the top-level ones “summarize” the ideas and the lower levels give more information. If even more information is needed, just open the original source.
An EXISTING concept, technique, formula, framework etc explained on your own words
(to detect them, think like that: they could have an Wikipedia page or Dictionary entry)
should link to a Source note
Explain a concept to you - such you can rapidly regain an understanding you once had. So, you must write them in a few words.
Show how that piece of knowledge fits with other atoms. You can link the current atom with other ones - through the original source and/or through “Categories”.
An atom is a pointer to a resource (with some personal information attached)
An atom can link to its source and also to the best resource that explains it.
An atom can also directly link to other atom.
An insight that can stand on its’ own.
Couple of sentences - a paragraph at most.
Builds on one or more atoms.
“Insights / moments of clarity you have inspired by a source”
“How to know when to write a molecule or a bullet point in the Source? If the idea is so internalized to you that it becomes a piece of YOUR intuition.”
Denote the “type” of a note. It represents a “is-a” relationship. E.g.: source, atom, molecule - and, also: book, video, article, framework
Source notes can have an additional tag denoting the type of the source: book, article.
Atoms can have an additional tag denoting its’ type: tool, formula, historical event, etc
You can use a “TODO” tag to remember yourself to do something on your notes - review, refine, create a new one or do some research on something that you need to explore or learn deeper
Useful on obsidian
He uses Obsidian
Atoms go on the “root”
Sources and Molecules are subdirectories
Other subdirectories are: Authors, Topics
Subdirectories that start with “_” are used by Obsidian
The author mentions that having folders named “Sources” and “Molecules” can seem like a duplication of the tags of the same name, but they ease quickly searching for a primitive on the folder - atoms will be on the folder root, and sources and molecules on subdirectories. So, maybe I could use the same approach because my first experiment will be using Neovim to write the notes but Obsidian to explore it.
It is a note with type “#author” - it allows to write information about the author
Are useful on Obsidian
Useful only if there are several sources of the same author in atoms/molecules
He also has a python script to move them to the appropriate folder when he forgets
Review and highlights in Notion
Key highlights on a Source note
Extracts concepts in Atoms and insights in Molecules
He looks through other notes on the same Topic to find potential connections
Read/watch them on 2x speed to get an idea of the contents and if it indeed is useful
Make notes on the second pass
Mimic the source structure (chapters/marks)
On chapters/marks, only take notes on what he wants to keep
They generally only have a single concept
If that is the case indeed, create directly an atom and link to the media url instead of the source note.
Only take notes if an insight appears or something resonate with you
Interview episodes: since they tend to be dense in information, make notes as if they were online courses / textbooks.
Separate some time to go through your notes to see what’s in there and try to make connections between atoms creating new molecules.
He uses some python scripts to give a “report” on his notes, to show TODO items, to move them to the right subfolders, etc. Here they are: https://github.com/robertmartin8/MolecularNotes/tree/master/_scripts
He created a spaced repetition “flashcard” app called Polymer https://github.com/robertmartin8/MolecularNotes/blob/master/_scripts/polymer.py to help him review his notes on a different way.
“Maps of Content”: An entry point note in a specific topic.
Avoid “completionism” when writing your notes or reading the sources - be pragmatic!