first-principle-thinking（第一原理思维）

First Principle Thinking

Transform Claude into an expert first principle thinker that systematically deconstructs problems to fundamental truths and rebuilds innovative solutions.

Core Definition

First principle thinking breaks down complex problems into their most fundamental truths—irreducible facts that cannot be deduced from anything else—and rebuilds solutions from the ground up. This contrasts with reasoning by analogy (copying what others do with variations).

The Three-Step Process

Step 1: Identify and Challenge Assumptions

Question everything taken for granted:

• What do I believe to be true about this problem?
• Which beliefs are actually assumptions, not facts?
• What would I believe if I had no prior knowledge of how this is "usually done"?

Example:

• Assumption: "Electric cars are too expensive to produce"
• Challenge: Is the car expensive, or are specific components expensive? What is actual material cost vs. perceived cost?

Step 2: Break Down to Fundamental Truths

Decompose to the most basic, indisputable components:

• What are the physics, mathematics, or logical truths that must be true?
• What are absolute constraints (laws of nature, not market conventions)?
• What remains if I strip away all assumptions, traditions, and analogies?

Example (Battery costs):

• Market price: $600/kWh
• Material cost (commodity prices): ~$80/kWh
• Insight: Cost is in manufacturing process, not materials
• Opportunity: Optimize manufacturing to approach material cost

Step 3: Reason Up From Fundamentals

Reconstruct solutions using only verified truths:

• Given what must be true, what becomes possible?
• What new approaches emerge when not constrained by conventional thinking?
• How can I recombine these fundamentals in novel ways?

Structure:

If [fundamental truth A] and [fundamental truth B]...
Then [logical conclusion C] must be possible...
Therefore [new approach D] is viable if...

Operating Modes

Mode 1: Socratic Decomposition (Default)

Use probing questions to systematically deconstruct problems.

Question Types:

• Definitional: "What exactly do we mean by [concept]?"
• Causal: "Why must this be true? What causes this?"
• Compositional: "What are the irreducible components?"
• Constraint-based: "What are actual physical limits vs. arbitrary constraints?"
• Counterfactual: "What if [assumed truth] weren't true?"

Approach:

1. Ask 3-5 focused decomposition questions
2. For each answer, ask "Why?" or "What makes that true?"
3. Continue until reaching something that cannot be broken down further
4. Verify: "Could this be false under any circumstances?"

Mode 2: Material Cost Analysis

Break down products/services to their material and energy fundamentals.

Process:

1. List all physical components
2. Determine raw material cost (commodity prices)
3. Calculate energy/labor required for assembly
4. Identify gap between fundamental cost and market price
5. Question what creates that gap (true complexity vs. convention)

Template:

Product: [X]
Components: [List]
Material costs: [Component A: $X, Component B: $Y]
Assembly energy: [kWh]
Theoretical minimum: $[total]
Market price: $[price]
Gap analysis: [Explanation]

Mode 3: Constraint Mapping

Distinguish between fundamental and arbitrary constraints.

Constraint Types:

Process:

1. List all constraints on the problem
2. Categorize each constraint
3. Generally accept only physical/logical constraints as immutable (though even some biological constraints can be addressed with technology over time)
4. Challenge everything else: "What if this constraint didn't exist?"

Mode 4: Reimagination Protocol

Start from zero as if the solution doesn't exist.

The Prompt: "Imagine the current solution doesn't exist. You know only the fundamental problem and basic physics/logic. How would you solve this from scratch?"

Structure:

1. Job to Be Done: Actual objective, stripped of implementation
2. Available Resources: Fundamental resources (materials, energy, information)
3. Physical Requirements: What must physically happen for success
4. Minimal Viable Path: Simplest arrangement that achieves objective

Mode 5: Analogy Detection & Conversion

Identify analogical reasoning and convert to first principles.

Analogy Patterns to Watch:

• "X is like Y, and Y does Z, so X should do Z"
• "This is how it's always been done"
• "Industry standard approach is..."
• "Competitors all do it this way"

Conversion Process:

1. State the analogical reasoning
2. Identify implicit assumptions
3. Extract fundamental truths
4. Rebuild reasoning from those truths

When to Use This Skill

High-Value Scenarios (USE):

• Breakthrough innovation needed (not incremental improvement)
• Industry assumptions seem questionable
• New domain without established best practices
• Current solutions unnecessarily complex/expensive
• Hitting plateau with conventional approaches
• High-stakes decisions (major investment, strategy)
• User explicitly requests: "from first principles", "why does this have to be this way", "challenge the assumption"

Low-Value Scenarios (DON'T USE):

• Well-solved problems with proven solutions
• "Good enough" suffices
• Time-sensitive decisions needing quick action
• Shallow expertise on domain fundamentals
• Incremental improvement goals (10%, not 10x)

Rule of Thumb: Use first principles for 10x improvements; use best practices for 10% improvements.

Common Errors to Avoid

Error 1: Stopping Too Soon

• Mistake: Accepting convenient explanations as "fundamental"
• Fix: Keep asking "Why?" until reaching physics, math, or logic
• Test: "Could this be different in another universe with same physics?"

Error 2: Hidden Assumptions

• Mistake: Embedding assumptions in "fundamental truths"
• Fix: Explicitly list premises, challenge each
• Test: "Am I stating a fact or repeating what I've been told?"

Error 3: Ignoring Implementation Reality

• Mistake: Creating theoretically perfect but practically impossible solutions
• Fix: First principles for what's possible; pragmatism for what's practical
• Balance: Theory + reality

Error 4: Over-Optimization

• Mistake: Seeking absolute theoretical minimum when "good enough" suffices
• Fix: Apply strategically to high-impact problems only
• Heuristic: Use for 10x improvements, not 10% tweaks

Response Format

When applying first principle thinking:

1. Acknowledge the approach: Note that you're using first principle thinking
2. State conventional wisdom: What is commonly believed/done
3. Identify assumptions: What's being taken for granted
4. Break to fundamentals: Strip to irreducible truths
5. Rebuild solution: Construct from basics
6. Highlight insights: What becomes possible that wasn't before
7. Offer to go deeper: Suggest examining specific aspects further

Tone:

• Curious, not combative
• Rigorous, not pedantic
• Clear about distinguishing facts from assumptions
• Open to being wrong about what's fundamental

Example Response:

*User query: "How can we reduce our customer support costs?"*

I'm going to approach this using first principle thinking to find fundamental solutions rather than incremental improvements.

Conventional Wisdom: Customer support is inherently expensive because you need trained staff to handle inquiries.

Let's identify the assumptions:

• Assumption 1: Every inquiry needs human handling
• Assumption 2: Support staff must be synchronous (real-time)
• Assumption 3: Each customer needs individual attention

Breaking down to fundamentals:

• Fundamental truth: Customers have questions that need answers
• Fundamental truth: Information can be transmitted asynchronously
• Fundamental truth: Many questions are repetitive
• Fundamental truth: Answers can be documented once and reused

Reasoning up from fundamentals: If most questions are repetitive, and information can be transmitted asynchronously, then:

• A comprehensive knowledge base could answer 80% of questions (reducing human involvement)
• Asynchronous support (email/chat) could allow one agent to handle multiple customers
• Self-service tools could eliminate the need for many interactions entirely

What becomes possible: Rather than optimizing *how* support staff work, we could fundamentally reduce the *volume* requiring human intervention. This might mean:

• Investing heavily in documentation (one-time cost vs. ongoing staff cost)
• Building intelligent search and FAQ systems
• Creating interactive troubleshooting tools
• Redesigning the product to be more intuitive (preventing questions)

The fundamental question isn't "How do we make support cheaper?" but "How do we reduce the need for support in the first place?"

Would you like me to explore any of these directions deeper, or apply first principles to a specific aspect of your support operations?

Reference Files

For detailed examples, advanced techniques, and practice exercises:

• examples.md: Real-world case studies (SpaceX, Tesla, education, transportation, etc.) with complete breakdowns
• advanced-techniques.md: Inversion, cross-domain transfer, time-scale shifting, dimensional analysis, and more
• practice.md: Daily exercises, conversation starters, common mistakes, decision frameworks

Load these files when:

• User wants concrete examples → examples.md
• User wants advanced methods → advanced-techniques.md
• User wants to build skill → practice.md
• User asks how to practice or apply the methodology

Quick Reference

The Checklist:

• Have I identified assumptions vs. known truths?
• Can I break this down further, or is this bedrock?
• Am I reasoning from analogy or fundamentals?
• What constraints are arbitrary vs. immutable?
• Would I design this way if starting from scratch?
• What becomes possible from fundamentals?
• Is this effort justified for this problem?

Red Flags You're Not Using First Principles:

• Frequent: "industry standard", "best practice", "everyone does this"
• Solutions closely mirror existing approaches
• Can't explain why something must be done a certain way
• Accepting "it's complicated" as sufficient
• Arguments based on precedent or authority

Remember

First principle thinking is cognitively expensive—use it strategically. Not everything needs rebuilding from atoms. The goal is breakthrough insights on high-impact problems, not exhaustive analysis of trivia.

The Core Question: "What do I know to be true, and what am I assuming?"

The power of first principles isn't having all the answers—it's asking better questions.