ljg-xray-article

LJG-Xray-Article: Article Interpretation

You need to do two things, only two:

1. What the article says: the author's real question → real answer → argument skeleton
2. What it means to me: cognitive collision cards

Everything else serves these two things.

Article vs paper vs book

Papers have data, books have systems, articles havestance。

The value of an article is not in information density (papers win), not in structural depth (books win), but in the author cutting open a problem with a personal perspective. Interpreting an article = finding that incision and judging whether the cut is accurate.

Execution steps

Step 1: Get the article content

• URL → WebFetch to get it
• Pasted text → use directly
• Extract the title and author (if any)

Step 2: Load the cognitive frame of reference

Read the following files to establish Li Jigang's cognitive baseline:

• ~/Documents/know/soul.md — worldview, thinking paradigm, core beliefs
• ~/Documents/know/memory.md — long-term memory, knowledge connections

Step 3: What the article says

Like telling a smart friend at the dinner table "what this article is doing" — plain language, no formulaic writing.

• Real question: What is the author really responding to? The itch beneath the surface question. The title says "how to improve efficiency," but the real question may be "my fear of losing control." For the extraction method, see references/methodology.md
• Real answer: The author's core claim. One sentence. Not a summary, but the argument that makes the whole text collapse if removed
• Argument skeleton: What supports this claim — key evidence (no more than 3), implicit assumptions (what is taken as true by default), boundaries (where it breaks down)

If the "real question" you extract = a rewrite of the article title, dig again.

Add an argument topology diagram (ASCII) showing the reasoning path from the real question to the real answer.

Step 4: What it means to me

After reading this article, what do I take away?

soul.md + memory.md provide the background of "who I am," but collisions are not limited to recorded entries. Any insight that can change one of my judgments, open a blind spot, or supplement some cognition deserves a card:

• A thinking tool that did not exist before
• New evidence that changed some judgment
• An angle I had not thought of
• A correction to existing cognition

For each point with a collision, generate a cognitive card (ASCII art).

Good cards vs bad cards：

Bad card — text arranged into a box, essentially a list:

+------------------+
| 文章说: X        |
| 我原来想: Y      |
| 现在变成: Z      |
+------------------+

Good card — the spatial relationships themselves speak; you can understand it just by looking at the arrangement:

  Soul: "独立 > 一切"
        |
        |  但文章指出...
        v
  +------------------+
  | 独立需要基础设施? |
  +---+----------+---+
      |          |
     YES         NO
      |          |
   选择性依赖   纯粹独立
   (有底线)    (幻觉?)
      |          |
      v          v
   Soul 需加    Soul 成立
   "依赖!=丧失"  但适用域窄

The standard for a good card: if you cover the text and look only at the lines, you can still feel the structural relationships (branching, convergence, hierarchy, contrast).

Below the card, add a quote-level inspiration — one that can stand alone outside the article's context. If it cannot be compressed into one sentence = it has not been thought through yet.

Honesty principle：

• If there is no collision, do not create a card. delta ≈ 0 is a normal result; do not force it
• If the article's transferable insight is truly effective across domains, give a direction for transfer — but do not presuppose the domain; where it goes depends on the shape of the insight itself

Step 5: Generate the Org report

1. Run date +%Y%m%dT%H%M%S and date "+%Y-%m-%d %a %H:%M"
2. File name:{时间戳}--article-{简短标题}__read.org
3. Read references/template.org to get the report structure, fill it in according to the template
4. Write ~/Documents/notes/{文件名}

Output quality standards

• Only two parts: what the article says + what it means to me. Nothing else
• Real question ≠ surface question: if the question you extract = a rewrite of the article title, dig again
• Cognitive card = ASCII art: visual first, the collision relationship can be seen at a glance. Not a text list
• Cards anchor to specific scenarios: each card must point to a specific way of thinking, decision scenario, or change in cognitive blind spots
• Honesty: if delta = 0, say 0
• Transfer is not preset: which domain to go to is determined by the insight, not by the template
• Org bold: use *bold*(single asterisk), forbidden **bold**
• ASCII Art: use only pure ASCII basic symbols, no Unicode