as well as by a detailed summary of my day of talking through note-taking options with ChatGPT, after a summer of trying Obsidian and exploring some other tools.
The draft was generated by this prompt:
Please write a blog post in Alex's voice titled "Mission: PKM". Write the post in the form of a letter from a field operative who has been assigned the mission of selecting a notetaking and PKM system by her intelligence-agency handler. In this case, Alex's handler is Tim. Tim originally recruited Alex as an operative while she was at Oberlin and later used the YxYY retreat, which ran from 2013 to 2017 at the Ace Hotel in Palm Springs, as the cover to provide additional briefings, trainings and mission parameters; since then, Alex and Tim have communicated entirely in the form of comments on one another's Facebook posts, in a clever way of covering their true mission and relationship. Assume that this letter will be delivered to Tim through the Agency's encrypted messaging system, so you can refer to their previous meets and message exchanges without fear of blowing their cover. Tim's assignment to Alex for this letter was, "I would enjoy going down the rabbit hole of PKM, and reading the results of your 20 hours of analysis, should you choose to publish them somewhere." So provide an appropriate level of technical detail to Tim on Alex's thought process and decision-making. The final draft should be 1500 to 2500 words.
The initial draft made some major errors, so I provided corrections and context:
The style and structure are terrific, but you arrived at the wrong conclusion! I am going to give Obsidian another try, not Mem. So here are the corrections to your draft:
CONTEXT AND CONCLUSION: This process started from me deciding to give up on Obsidian after two months of giving it a serious try as an alternative to Evernote (which I used for 17 years). The "mission" I am briefing Tim on is the mission to assess potential alternatives, not just in terms of software, but in terms of note-taking approach. This work including clarifying my mission goals (which turned out to focus on 4 key criteria.)
The real thinking started from these prompts (which were background to the materials I gave you):
I think I just gave up on Obsidian and decided to recommit to Evernote. I cancelled Tana without really trying it. But now you remind me the reason I wanted to try obsidian was the ability to connect notes.
The truth is that I am a little muddled about my Evernote vs Coda vs Scrivener vs Zotero vs ChatGPT workflow. I use all three constantly, but I also have a lot of problems in my workflow. Mabye we shoudl try a voice coaching session where I talk to you about the problems I am trying to solve and we can puzzle this through. But I would like you to first internalize Tiago Forte's BASB and PARA methods. Do you know them?
[via ChatGPT’s voice interface, and slightly cleaned up here]: Okay, my problems. It bugs me, Skitch has kind of broken, and now I'm trying to use something called XSnapper, which I have through Setapp, which is kind of helping. We'll see if it actually works. So that's one problem. Another thing is I did want the sort of mind mapping ability in Obsidian to be able to connect my notes, like that aspect of Roam and Obsidian really interests me and seems like my kind of jam, but it just seems like it comes at too high a price in terms of the usability of Obsidian. I am pretty convinced I'm giving up on Obsidian, you know, among other reasons, like I need to write about tools that humans can use, and it's just like too hard.
Another problem I have is that because I do so much of my work in Coda now, and Coda search is just not fabulous, it's just hard for me to find things. Like I do my article writing in Coda, I have my notes in Zotero often because I'm highlighting articles, then I had to do so much of my writing and thinking in ChatGPT or in Claude, and then even in Scite, things get trapped in all those apps. Although I was experimenting with this workflow that would like export my ChatGPT chats to Obsidian, and the idea that that would become adjustable, but just like the sheer volume of my chat history in ChatGPT is so enormous, it becomes really unwieldy. And of course, it doesn't solve the problem of Claude or Scite.
So there's just like a search problem of like, I have to try and remember, did I create this in Evernote? Did I create this in Coda? Which Coda? Coda's search sucks, and I created this thing called the Super Coda ages ago, which is like a pretty good index of my actual Coda, individual Coda docs, but it doesn't like search across docs, and the Coda cross doc search is just really not fabulous. So I guess, yeah, I think those are my pain points.
It's just, and then also just like my Evernote notebook is perpetually, you know, I just sort of drop everything in Evernote, and I don't even bother to try and file it or organize it now. And so, you know, the point Tiago Forte makes about like the serendipity of having things in an organizable way, I just, basically, I can find things, sort of, although it gets harder and harder if I have to search across, you know, chatgpt’s history, and Scrivener, and Zotero, and Coda, and Evernote. And I now I'm going to have to import all my Obsidian notes over the past couple of months into Evernote. The most important problem is probably thinking through the relationship between Coda and Evernote. But I also, yeah, I would really like there to be some kind of like visualization tool, maybe the Tana thing where every bit of information is potentially its own node is a very cool idea. But to me, the deal breaker is that there's no mobile app. But Jesus Christ, Evernote is so expensive now, I wouldn't preclude trying Tana again. So there you go. That's my big brain dump.
CONTENT: Provide a section where I brief Tim on what I've learned about major PKM schools of thought: PARA/CODE, Zettelkasten, Atomic Notes, Commonplace journal, the "daily notes" practice. Describe them as if they are each a faction in an undercover intelligence war we are trying to comprehend.
STYLE: Don't open with the digression about the Ace. Tuck those references into the draft more subtly. (Like, "It sure would have been useful to have a better way to save my notes on YxYY", somewhere). I love the thing of referring to platforms as suspects I interrogated.
TIMELINE: Don't say this was weeks. The Obsidian "mission" was a two-month investigation. Then I hade a 24-hour 'operation' to interrogate alternate suspects. Now I'm going back undercover into the world of Obsidian nerds (including the obsidian discord server) to ensure we have reached the right conclusion for Mission:PKM.
Summarizing my PKM exploration
The summary of my PKM exploration was developed by giving generic Claude the transcript of my PKM exploration with GPT, with the following sequence of prompts:
Prompt 1. Please summarize the attached analysis of my PKM selection process. What steps did I take to think through my choices? What platforms did I consider? What were my pain points and goals? What made it hard for me to choose? How did my thinking evolve over time? What were my top accomplishments and insights? What remains unresolved? What will I try to do in Obsidian now that I have landed there?
Prompt 2. Excellent. Please provide a greater level of detail including excerpts in the prompter's voice. You are developing a 1500 to 2500 word briefing note that will be used as a first draft of the prompter's article about her exploration and the 4 notetaking problems she is trying to solve with a new platform.
Prompt 3. Please include a new section, "Evernote, Obsidian, or....??" Where we discuss why i had reservations about obsidian after my 2months of trying it, why I wasn't just prepared to go back to evernote, and the frustrations I expressed with each platfrom.
Then include 4 draft "what I learned" sections:
1. Note-taking religions: Recap of AI-generated notes on zettelkasten, commonplace book, PARA/CODE
Unified Search support [did it meet this criterion? yes/no/how/how not?]
AI-Driven Organization [did it meet this criterion? yes/no/how/how not?]
Zettelkasten-Style Linking [did it meet this criterion? yes/no/how/how not?]
Mind Mapping [did it meet this criterion? yes/no/how/how not?]
Temptations [what did user Alex say, if anything, about what made this platform appealing? What did GPT say about its best features/appeal/why to consider it? this can include criteria beyond the 4 key facets alex was looking for]
Reservations [what did user Alex say, if anything, about what made this platform irritating, problematic or a no-go? What did GPT say about its worst features/limitations/why it might not meet needs? this can include criteria beyond the 4 key facets alex was looking for]
Want to print your doc? This is not the way.
Try clicking the ⋯ next to your doc name or using a keyboard shortcut (
About this draft
