Evaluate search results by reading only the bolded/highlighted passages from progressive summarization — skip full-text scanning

Why This Is a Rule

Tiago Forte's progressive summarization creates layers of distillation within notes: Layer 1 (full text), Layer 2 (bold key passages), Layer 3 (highlight the most essential phrases within bold passages). When you search your system and get 15 results, reading the full text of each to evaluate relevance would take 30-60 minutes. Reading only the bold/highlighted passages — the pre-distilled essence — lets you evaluate the same 15 results in 5-10 minutes.

The progressive summarization layers function as a pre-computed relevance filter. The bold passages represent what past-you determined was most important about this note. When present-you is evaluating search results, the bold passages answer the question "Is this note relevant to what I'm looking for?" far faster than scanning the full text. If the bold passage matches your current need, open the note and read the full context. If not, skip to the next result.

This retrieval pattern is the payoff for the upfront investment of progressive summarization. Without it, every search result requires full-text evaluation — O(n × text_length). With it, each result requires only bold-passage evaluation — O(n × bold_length), which is typically 10-20% of full text. The time savings compound with every search across hundreds of notes.

When This Fires

• When evaluating search results in a note system that uses progressive summarization
• When search returns many results and you need to quickly identify the most relevant ones
• When retrieval from your notes system feels slow relative to the information you know is there
• Complements Title notes with the natural-language words you'd search for, not abstract labels or jargon — search terms beat category names (searchable titles) with the post-search evaluation technique

Common Failure Mode

Full-text scanning of every search result: opening each note and reading from top to bottom to determine relevance. This is the brute-force approach that progressive summarization was designed to eliminate. If you're reading full text to evaluate search results, either your notes aren't progressively summarized or you're not using the layers during retrieval.

The Protocol

(1) When search returns results, scan the list for title relevance first (Title notes with the natural-language words you'd search for, not abstract labels or jargon — search terms beat category names — titles should contain search terms). (2) For potentially relevant results, open the note and scan only bold or highlighted passages. These are the pre-distilled key points. (3) If the bold passages match your current need → read the full context around them. (4) If the bold passages don't match → skip to the next result. Don't read the full note hoping the relevant part is hiding in the un-highlighted text. (5) If you frequently find relevant content in non-highlighted passages, your progressive summarization needs updating — the layers aren't capturing the right material. Flag these notes for re-summarization.