Conversation

Anyone ever try long-term self-summarizing documents on ambiguous things? Like a mind map that you keep adding to over months, with occasional refractors? I don’t mean linearly cumulative things like spreadsheets but things where adding items takes thought
18
74
Share themes and best practices if you do something like this. Looking for alternative paradigms to notebooks where there is no progressive summarization unless you adopt the discipline to do it.
3
15
My old phd/postdoc notes were models of discipline. I’d summarize every working session at the end, do periodic meso-summaries every few weeks, and maintain a running table of contents with 1-line summaries. In a 2020-21 notebook I just finished there’s only dates.
2
19
But this was 1998-2006 when I had like 10x more meta-cognitive executive attention management energy and also maintained a functional GTD. All that’s fallen apart now. My information empire/second brain is now like late stage Trantor. Brink of collapse.
1
26
Without a cumulative meta-process of progressive summarization, notes are basically private shitposting and a cognitive random walk. The 2020-21 (paper) notebook I just finished is like 10% of the review utility as one of my 3-ring binders from say 2002 when I was great at this.
1
19
I think it’s just regular aging (34y old brain vs 46) but Waldenponders would say Twitter rotted my brain since I joined in 2007 🤣 Whatever the cause, gotta build a senior-citizen second brain system. I though Roam would do the trick and it dies help, bus no magic cure.
2
16
After publishes his book, I’ll write a derivative ebook called “second brains for senior citizens sliding into senility” Subtitle: how to run your life on 10,000 prefrontal cortex neurons or less
2
21
The big problem with evolving a rolled-up state as opposed to periodically rolling up a state-log (aka chrono notebook), is that refactorings are like earthquakes. Imagine rebasing the root of a mind map with 1000 nodes? I think code version control has a similar problem.
1
4
There may be a high floor on the complexity here. Nobody’s yet figured out how to automate something like a fork —> refactor —> merge on even trivial code. Mergeconflict solving I suspect is isomorphic to summary-state notes refractors. Both probably np-hard if well-posed at all.
Replying to
Every lifestyle/workflow problem has to be re-solved from scratch for every life stage. The way people usually solve things for the 45-60 life stage is some combination of dropping activities and adding minions. Minions are expensive though and the robot butlers currently suck.
2
11
Hmm, maybe I can set up 2 foundations at opposite ends of my brain to prepare for dark age when main part of brain, the prefrontal cortex empire, finally falls. First foundation will be a notes app. Second foundation will be a small ML model trained on my twitter 🤔
1
12