Generative AI organizes information with high accuracy and speed — very welcome. Doing it myself eats a lot of head resource, too heavy for spare moments at work.
Not only this project — I'm helped so much that almost any work is hard to advance without generative AI, specifically Claude Code.
Beyond analyzing this blog, I'm loading various data and trying to build a database about myself.
This time: analysis results when I loaded the blog, and thoughts from building a self-database.
My thinking patterns from analyzing this blog were:
＝＝＝＝＝＝＝＝＝＝＝＝＝
A way of thinking that consistently repeats across six years. Core: "discomfort/lack → put into words → update own design parameters." Not distributing frameworks for others — work logs of updating my own operating settings, with the blog published as a byproduct.
- Reframing conventional wisdom: catchphrase notions like "management studies are useful," "issue-driven," "keep it simple" — reconstruct from premises
- Concrete → abstract → caveat: elevate primary experience (diner debates, bouncers, haircuts) into universal frames, always adding "not all-purpose"
- Seeking the middle of dichotomies: analysis vs synthesis / short vs long / numbers vs values / logic vs conviction
- Modeling/formulas: failure impact, PDCA element breakdown, problem-solving essence — drop phenomena into formulas or classifications immediately
- (C1) Reframing conventional wisdom: reflex to inspect premises of catchphrase notions. Shows up at least once per article
- (C2) Speed of abstracting experience: fast elevation of small primary experience into universal frames + the "not all-purpose" caveat
- (C3) Responsiveness of modeling: speed of dropping phenomena into formulas/classifications. Peaked in 210130 "Essence of problem solving"
- (C4) Connecting daily life to macro analysis: figure shop → market cap, moving house → cognitive load — natural jumps
＝＝＝＝＝＝＝＝＝＝＝＝＝
Generative AI basically affirms everything, so some evaluation is probably excessive — don't overtrust.
I also had it produce reader personas.
=====================
Core commonality: seeking a writer who speaks for "discomfort I feel too but haven't put into words."
- Think for themselves at work but feel they don't get language for it
- Self-help efficiency thinking and anti-self-help rebellion both feel crude
- Bookmark while thinking "I wanted to say that"
- Purpose of reading isn't "get answers" but "confirm/expand my own thinking"
Articles that land: Does management studies not help business?, Muddy / grit, Don't keep it simple, To companies running some kind of interaction platform, A tip for writing interesting blog posts, A misunderstanding about others' evaluation.
What pushes them away: strong assertion tone, conclusion-first, walls of headings & bullets, buzz-chasing.
=====================
This also lands as "yeah, that sounds about right" — but I rarely get to put it into words accurately, so it's interesting.
That's it for blog analysis results for now.
From here: thoughts while building a self-database including blog analysis.
This blog isn't ultra-frequent, but it records thought at high purity and detail, so it's pretty good data — still, there are other sources I want to reference.
First, fact-based work data and achievements. Simply: like a resume — but it hadn't been updated since my last job change. Independent or not, project results and measures are hard to review later in detail; maybe I should have updated a career doc regularly. Still, I probably won't use it for job hunting, so a resume itself isn't needed — I want finer records of what I did. Resumes only dress up clear wins; this DB should include work that's bad for showing off but big experience for me, and things that didn't produce results.
Next: daily work-journal data. Roughly what I did and what I think, typed casually into Google Forms. I later learned it's so casual that as reference data for a self-DB it had little value. Rereading myself I think "ah, I was thinking that then," and nothing more. Ideally: write insights, hypotheses, decisions in as much detail as possible. Casual writing is easy, but after building this self-DB I stopped valuing producing low-value data, so I made it more businesslike — a business daily report. That should also raise reflective effect.
Lastly: things I summarize for myself the way I put thought into the blog. Not huge volume — several docs of thinking axes I consider especially important, like essences.
Summary: blog, work journal, career history, thinking docs — four.
I dump them into a Notion database.
I first used Notion about three years ago, found it awkward, but dumped lots of data and notes anyway — then paid "information hostage fees" as a subscription.
For AI-linked database use it regained value. Flexibility plus page hierarchy turned out quite convenient. Same use case could be Google Drive + Sheets with AI, but for this case Notion suited me better.
By the way, this project made me open Notion after a long time — I never looked at the old accumulated data, so if you're paying "information hostage fees," I recommend cutting the cord and canceling.
Back: I ended up storing in three layers.
One: raw data. Blogs and docs above, or links to raw data.
Two: observation layer. Summaries, grouping, tagging of raw data — structured thinking patterns.
Three: thinking-OS layer. Generalized personal thinking patterns. The thinking docs above belong here.
Through this, I can fairly say I have richer self-data than average — blogs and the habit of writing and accumulating thought often.
So I thought: structure those four role-split data sources nicely, find commonalities and intersections, and maybe get deep insights.
Expectation miss: however newly labeled the structure, insights only hit "yeah, that tracks" temperature.
Well thought through: I already regularly put thinking patterns into words and summarize them — restructuring that again won't yield much new discovery. Already verbalized.
But I got a reverse way of thinking: what data should I accumulate so future-me gets new insights?
I.e. what kind of daily language output would raise the whole database's value — a good prompt for that.
Technique talk like verbalization skill may exist, but dig deeper and what experiences you should seek also matters.
With the goal of expanding the self-database, producing data has started to feel fun.
That's all.
Tweet© 2026 ZUUHE