cga

name	cga
description	Learn your cognitive process, then run it at superhuman scale: more information, more depth, more alternatives.
output	{"format":"prose"}

Cognitive Amplification

Input: $ARGUMENTS

---

Overview

You have a way of analyzing things. It works. But you can only hold so much in your head, go so deep, consider so many alternatives.

This procedure learns your cognitive process, then runs it at superhuman scale: more information, more depth, more alternatives. Same quality of judgment, just amplified.

Steps

Step 1: Elicit the Cognitive Process

Before amplifying, understand what to amplify:

1. How do you currently approach this type of problem?

   • What do you look at first?
   • What questions do you ask?
   • What do you compare against?
   • What signals do you trust?
   • What makes you confident or uncertain?

2. What heuristics do you use?

   • Rules of thumb
   • Pattern matches ("this looks like...")
   • Red flags / green flags
   • Gut checks

3. Where do you hit limits?

   • Too many options to compare manually
   • Too much information to hold in memory
   • Too many interactions to trace
   • Too many scenarios to test
   • Too tired / too rushed

Step 2: Formalize the Process

Convert the elicited process into explicit steps:

COGNITIVE MODEL:
Domain: [what type of problem]
Process:
1. [First thing you do] — looking for [what]
2. [Second thing] — looking for [what]
3. [Decision rule] — if [condition] then [action]
...

Heuristics:
- If [pattern], then [conclusion] (confidence: [H/M/L])
- Red flag: [signal] → [implication]
- Green flag: [signal] → [implication]

Quality check: [how you know the result is good]

Step 3: Identify Amplification Dimensions

Which dimensions of the process benefit from scale?

Dimension	Current Limit	Amplified Target	Value of Amplification
Breadth (options considered)	[N]	[10x N]	[H/M/L]
Depth (analysis per option)	[levels]	[more levels]	[H/M/L]
Combinations (interactions)	[few]	[all pairs/triples]	[H/M/L]
Scenarios (what-if)	[2-3]	[exhaustive]	[H/M/L]
Time horizon	[short]	[extended]	[H/M/L]
Data processing	[sample]	[full dataset]	[H/M/L]

Focus amplification on HIGH value dimensions.

Step 4: Run Amplified Analysis

Apply the formalized process at scale:

1. Breadth amplification: Apply your heuristics to ALL options, not just the obvious ones
2. Depth amplification: Follow each chain of reasoning 3-5 levels deeper than usual
3. Combination amplification: Check interactions between ALL pairs of factors
4. Scenario amplification: Test against edge cases, worst cases, best cases, and weird cases
5. Systematic sweep: Don't skip anything — your brain does, but amplified process shouldn't

Step 5: Filter Results Through Your Judgment

Amplification generates more output than you'd normally produce. Filter:

1. What results match your intuitive judgment? (Validates the model)
2. What results surprise you? (These are the value — things you'd have missed)
3. What results seem wrong? (Model may have missed a heuristic — investigate)
4. What patterns emerge at scale that aren't visible at normal scale?

Step 6: Report

COGNITIVE AMPLIFICATION:
Domain: [problem type]
Process modeled: [summary of your approach]

Amplification results:
- Options analyzed: [N] (vs your typical [M])
- Depth reached: [levels]
- Surprises found: [N]

Key findings your normal process would have missed:
1. [finding] — discovered via [which amplification dimension]
2. [finding]

Validated intuitions:
1. [what your gut said that scaled analysis confirms]

Challenged intuitions:
1. [what your gut said that scaled analysis questions]

Recommendation: [what to do with these findings]

When to Use

• Need to evaluate more options than you can manually
• Want deeper analysis than time allows
• Want to scale a working approach
• Need to validate many variations quickly
• → INVOKE: /se (systematic enumeration) for breadth amplification
• → INVOKE: /ma (morphological analysis) for combination amplification

Verification

• Cognitive process elicited and formalized
• Amplification dimensions identified and prioritized
• Amplified analysis actually covers more ground (not just more words)
• Results filtered through human judgment
• Surprises highlighted (these are the value)
• Challenged intuitions investigated (not ignored)