I actually never got into note-taking before I found Obsidian. I used Google Drive all throughout college and up to 2023, so any knowledge I had written down was sequestered by an ad-hoc folder structure that was mostly chronological by year, or in my physical notebooks by subject. It was also limited by what I felt was worth writing down enough that it merited a Doc, or what I could write in one session before getting distracted and never touching it again. And mentally I limited myself by always wanting to write something down "right", all spell-checked and grammatically correct and sensically organized, which led to often not writing anything down at all. Now I dump the words down and come back to it later when I want to "garden."

My brother showed me stuff like Trilium circa 2017, but I hadn't the thinking process I do now to even know where to start racking my brain for stuff to write down.

When I read Chernow's biography of Alexander Hamilton, I was in awe of his ability to write so damn much. I never thought I'd be able to do that. Turns out the secret is just three things: have stuff you're passionate about, be able to recall it, and write so damn much about it. When your thinking process is based around "how would I phrase/word/frame this to write it down," if you have the right process and organization to be able to access it later, it's even easier than talking. For some they can keep everything they know in their head. I'm not one of them. but I can write everything down, and in writing it down I end up remembering it better. My professors were right all along.

And if one looks at the actual written letters people like Hamilton [1] and co. wrote, or even back to Isaaci Newtoni [2][3], They're riddled with spelling errors and strange latinesque grammar and informal formalities. Yet they're revered for their ability to write. Because it really is the thought that countts [4], not the words.

(Very little of these comments is new thoughts I'm having now. Most of it is thoughts I had and documented when I was super into PKM in 2023 and since, and now comes back up as those neurons fire again and I consider the new idea of "should AI be my PKM?" after reading the post.)

---

Yeah, the graphs are cool for a little bit. But only post-facto, once one has an amount of data points where it might become useful. If the AI is doing the organization then any personal significance is lost. Or rather it was never there to begin with.

Wikilinks is the feature I use most often, outside of my folder organization (PARA). Now when I have a thought, it goes down a chain of "do I have a note already this can go into," to "no, it should be a new note. are there any notes I should wikilink in this one, or link this one in?"

I think I made a good decision early on when I was inspired by the Emacs documentation to add a basic "Related: " line before the first header (and after the YAML). There I dump any wikilink I think might possibly be something I want to reference, or find this note via a backlink, without having to think about where to put it in the body.

E.g.

{YAML header}

Related: [[Artemis]], [[Artemis II]], [[NASA Engineering]], [[Space MOC]], [[NASA CAPCOM]],

# Artemis II Mission Timeline and Notes

{body and rest of note, my own record of things that happened as I watch the stream}

---

> AI tools certainly help with creating a mass of notes.

Agreed. Presuming the implication is it creates a mass of notes, but of generic information stated generically. I'm really proud of my 4100 notes, because I know (aside from a few catagories like web clipping) even if they're a mess, they're my mess. I definitely could have gone the last three years without having found Obsidian, but I wouldn't have as clear a record of them as I do now. Or the rest of my life, as I slowly add stuff about my past, or migrate old writing into it. I also definitely repeat myself by saying the same information in different places, but in different ways. It's not 'efficient' information density-wise, but it is designed for a human to read and see the human behind the writing.

I also believe I think clearer, as often when I'm recalling information I'm actually recalling my note in my head on that subject. I write so much that in conversation "I was thinking the other day" is analogous to "I wrote down in my notes".

I might be crazy but I would put my vault in my will as something to be passed on, because there's so much me in them. My yearly journals in /02 Areas/Journals/ are the most obvious ones, but I have a /02 Areas/Writing/ folder that's just notes I consider "writing", whic is distincy from the contents of /03 Resources/ folder that's the "general knowledge" knowledgebase.

---

Anyway, I guess my tl;dr is that AI can never write about thinking as well as a human can, and in my opinion it's the thinking that important, not the writing. the writing or the words is merely a tool in thinking. Karpathy mistakes the words to be the goal, rather than the thinking that caused the words.

---

One last thing: I just re-read the HN guidelines out of curiosity, and I noticed they recently added "Don't post generated comments or AI-edited comments."

I could copy and paste almost anything from my vault into an HN comment without violating this rule. Anybody creating a PKM with this sytem could not. They would have to rewrite it in their own words. So one might as well just right it themself in the first place if they ever think they might want to reuse it in a place like HN.

---

[1] https://outhistory.org/exhibits/show/rev/hamilton-laurens-le...

[2] https://www.newtonproject.ox.ac.uk/view/texts/diplomatic/MIN...

[3] A while back, while in a Newton phase, I decided arbitrarily to refer to him as "Isaaci Newtoni," as that's how he called himself. I reinforced that by using that name for him in my notes. Now I call him that instinctually, not consciously.

[4] Intentional.

LLM Wiki – example of an "idea file"