038 Deep Cc09554b

# Focusing Mode

Deep, narrow attention on specific signal.

## When to Use

- Signal identified as important
- Need full understanding of one thing
- Detail matters for decision
- Surface-level insufficient
- Ready to invest significant attention

## Mental Model

**Microscope:** Sacrifice breadth for depth. See what scanning cannot reveal.

## Focus Approaches

### Exhaustive
Leave no stone unturned.

**Use when:** High-stakes decision depends on this, need to be certain, will be asked to defend.

**Process:**
1. Identify all available sources
2. Review each source thoroughly
3. Cross-reference findings
4. Document gaps
5. Continue until no new information

### Targeted
Answer specific questions.

**Use when:** Have specific questions to answer, don't need complete picture, time-constrained.

**Process:**
1. Define questions upfront
2. Identify sources likely to answer
3. Seek answers directly
4. Stop when questions answered
5. Note remaining unknowns

### Iterative
Deepen progressively until returns diminish.

**Use when:** Uncertain how deep to go, exploring unfamiliar territory, want to optimize effort.

**Process:**
1. Start with overview
2. Identify most important aspect
3. Go deeper on that
4. Assess: enough or more needed?
5. Repeat until diminishing returns

## Execution

### Target Definition

Before focusing, define:

**Signal:** What specific thing are we examining?

**Why important:** What makes this worth deep attention?

**Questions:**
- Must-answer: [Questions that must be resolved]
- Nice-to-answer: [Questions that would be helpful]

**Boundaries:** What's in scope? What's explicitly out of scope?

### Source Prioritization

**Primary sources:** Original source of information (company announcements, official documents, research data). Highest reliability.

**Secondary sources:** Analysis of primary sources (analyst reports, news articles). High reliability.

**Tertiary sources:** Aggregation of secondary sources (Wikipedia, roundups). Medium reliability.

### Process

1. **Prepare** — Review target and questions, gather sources, prepare notes structure
2. **Primary sources** — Review each thoroughly, extract answers to questions, note what's still unknown
3. **Secondary sources** — Fill gaps from primary, cross-reference claims, add context
4. **Tertiary sources** — Only if gaps remain, verify any claims found, note lower reliability
5. **Synthesize** — Consolidate findings, answer original questions, document confidence levels

### Depth Signals

**Go deeper if:**
- Key question still unanswered
- Finding contradicts expectation
- Critical detail missing
- Source suggests more exists

**Stop if:**
- All must-answer questions addressed
- Diminishing new information
- Time budget exhausted
- Remaining unknowns not critical

## Output Format

```markdown
## Focusing: [Signal]

**Target:** [What we examined]
**Why important:** [Rationale for deep dive]
**Approach:** [Exhaustive/Targeted/Iterative]
**Time spent:** [Duration]

### Questions & Findings

**Q: [Must-answer question 1]**
A: [Answer]
- Confidence: [X%]
- Sources: [What supported this]
- Notes: [Caveats or nuances]

**Q: [Must-answer question 2]**
A: [Answer]
- Confidence: [X%]
- Sources: [What supported this]

### Understanding

**Summary:** [Consolidated understanding in 2-3 sentences]

**Key insights:**
- [Insight 1]: [Evidence]
- [Insight 2]: [Evidence]

**Nuances:**
- [Important subtlety or caveat]

**Still unknown:**
- [Gap 1]
- [Gap 2]

### Implications

**For thinking:** [How this affects reasoning about the broader problem]

**For action:** [What this suggests doing]

**Suggested thinking mode:** [Which rsn-reasoning-problems mode to use next]
```

## Quality Gates

| Gate | Requirement |
|------|-------------|
| Questions defined | Before focusing starts |
| Confidence stated | Every finding has confidence level |
| Multiple sources | Key claims from multiple sources |
| Unknowns documented | What we still don't know |
| Implications extracted | What this means for next steps |

## Anti-Patterns

| Avoid | Problem | Do Instead |
|-------|---------|------------|
| Endless research | Never actionable | Define questions upfront |
| Single source | Bias risk | Require multiple sources |
| No confidence levels | False certainty | State confidence explicitly |
| Ignored unknowns | Blind spots | Document what's not known |
| No implications | Insight without action | Extract meaning for decisions |