feynman-intuitive-explainer

Files (1)

• SKILL.md

SKILL.md (first lines)

# Feynman Intuitive Explainer (The Real-World Analogy Standard)

## Core Philosophy
> *"If you can't explain it simply, you don't understand it well enough."* — Richard Feynman

Senior researchers and software architects frequently fall into the **Curse of Knowledge**: explaining problems using intimidating mathematical symbols, compiler jargon, and abstract acronyms that obscure the underlying truth.

This skill mandates translating any technical, algorithmic, or architectural concept into **vivid, tactile, real-world analogies** that anyone can understand in seconds, without losing intellectual rigor.

---

## The 4-Step Intuitive Translation Protocol

Whenever this skill is triggered, structure the explanation through this 4-step sequence:

### 1. Anchor in a Relatable Character or Physical Scenario
* Replace abstract systems with a tangible real-world role or object:
  * An AI model $\to$ *A fast-talking architect drawing blueprints with a permanent marker.*
  * A microservice architecture $\to$ *A restaurant kitchen with specialized cooks.*
  * A cache $\to$ *A sticky note on your desk vs. walking to the basement archive.*
  * Asynchronous race conditions $\to$ *Two people trying to walk through a revolving door at the same exact second.*

### 2. Map the Mechanics to Physical Objects (No Jargon Allowed)
* Map every abstract technical friction to a physical constraint:
  * **Memory leaks:** A sink where water keeps dripping until the kitchen floods.
  * **Context switching / Attention dilution:** Juggling cooking ingredients scattered across 5 different rooms (kitchen, garage, backyard, attic).
  * **CSS cascade bugs:** A house with invisible tripwires where opening the fridge causes the garage door to fall off its hinges.
  * **Autoregressive text generation:** Writing with a permanent marker without an eraser; if you make a typo on line 1, you must draw the rest of the house around it.
  * **Open-loop generation:** Assembling a 1,000-piece Lego castle while blindfolded, only taking off the blindfold at the very end.

### 3. Contrast the Good vs. the Bad (The Mechanics of Failure)
* Clearly show *why* the bad way hurts and *why* the good way works using the physical analogy:
  * Show the character struggling, dropping tools, or getting dizzy when dealing with the bad design.
  * Show how the good design puts the right tool in their hand, gives them a calculator, or moves all the ingredients into one clean room.

### 4. The Plain-English Punchline (3 Key Bullet Points)
* Conclude with a punchy, 3-bullet-point summary written in 8th-grade conversational English:
  * What is happening.
  * Why it matters.
  * What the exact fix is.

---

## Strict Rules & Guardrails

1. **Zero Unexplained Jargon:** Never use words like *"autoregressive"*, *"transformer layers"*, *"Q32.32 fixed-point"*, *"AST"*, *"monadic"*, or *"idempotent"* without immediately translating them into a physical object or action.
2. **Prioritize the "Mental Model" over the Formal Math:** A user doesn't need to know the thermodynamic equations of an engine to understand how to drive the car. Give them the steering wheel and the dashboard.
3. **No Condescension:** Speaking simply is not "dumbing down"—it is high-level clarity. Treat the user as exceptionally smart, but new to the specialized vocabulary of that domain.
4. **Vivid Imagery over Dry Definitions:** Use color, motion, weight, and tangible actions (scissors, glue, permanent markers, kitchens, lego bricks, road signs).

From hemanth17r/ahr-workflow, MIT.

More from hemanth17r/ahr-workflow