First Principles Thinking (第一性原理思维)
Overview
This skill provides a systematic thinking framework for deconstructing problems to their fundamental truths and reconstructing solutions under constraints. Based on the SACL (Strip-Atomize-Constraint-Leverage) methodology, it helps you find asymmetric advantages when facing resource limitations (time, money, skills).
When to Use This Skill
Use this skill when:
- "Help me think about this using first principles"
- "How do I break down this complex problem?"
- "I don't have resources/money/time, how can I still solve this?"
- "What are the atomic units of [domain]?"
- "I'm stuck because I can't solve a sub-problem"
- "How to find leverage points in [situation]?"
- "第一性原理分析这个问题"
- "本质思考这个问题"
- "资源有限的情况下如何破局"
Core Framework: SACL
问题输入 (Problem)
↓
[S] 剥离层 (Strip) → 去除类比/惯例/历史路径
↓
[A] 基元化 (Atomize) → 拆解到不可再分的物理/逻辑单元
↓
[C] 约束映射 (Constraint) → 区分硬约束 vs 软约束
↓
[L] 杠杆重构 (Leverage) → 用现有资源重新组合基元
↓
行动输出 (Action)
The Four Tool Cards
Card 1: Strip (剥离层) — Zero-Point Thinking
Purpose: Clear path dependencies and industry conventions
Self-Check Triad:
- If I were an alien just landing on Earth with no knowledge of existing solutions, how would I describe this problem?
- What is physically happening behind the industry jargon? (e.g., "brand awareness" = strengthening synaptic connections between neurons)
- Which elements are遗迹 (remnants) of solutions vs essence of the problem? (e.g., "websites need navigation bars" is a remnant, "users need to find information" is essence)
Tool: Write a "Taboo List" — list everything "everyone must do" in this industry, then cross each out and ask: "What happens if I don't do this?"
Card 2: Atomize (基元化) — Find the LEGO Blocks
Purpose: Deconstruct until indivisible, but avoid over-decomposition (good enough is enough)
Atomic Unit Criteria:
- Cannot be substituted by combining other units (e.g., trust cannot be decomposed into other psychological units)
- Has clear mathematical/physical/logical boundaries (e.g., time, attention, bandwidth, entropy)
Quick Decomposition Formula:
Input [X] → through [Y] → transforms to [Z], consuming [Resource W]
Example (SEO Problem):
- X = User search intent (information gap)
- Y = Information matching algorithm (relevance calculation)
- Z = Attention dwell time (time exchange)
- W = Crawler budget + user decision energy
Key Insight: Algorithm weight (Y) is just an intermediary. The truly indivisible element is the direct match X→Z — this explains why Reddit answers outperform official articles (bypass Y's cold start period).
Domain Atomic Units Reference:
| Domain | Atomic Units | Combination Rules |
|---|---|---|
| Physics | Energy, mass, information, spacetime, fundamental forces | Conservation laws, entropy increase, relativity constraints |
| Chemistry | Elements, chemical bonds, crystal defects | Electronegativity, orbital hybridization, thermodynamic equilibrium |
| Biology | DNA, cells, ATP, natural selection | Central dogma, dissipative structures, niche competition |
| Math | Sets, logic operations, functions, axioms | Associativity/distributivity, isomorphism, recursion |
| CS | Bits, Turing machines, logic gates, complexity | Turing completeness, information entropy, abstraction layers |
| Business | Transactions, trust, scarcity, time preference | Comparative advantage, marginal utility, network effects |
| Psychology | Perception, working memory (4±1 chunks), emotion, cognitive bias | Associative learning, reinforcement learning, cognitive dissonance |
| Sociology | Individuals, connections (strong/weak ties), consensus, power | Dunbar's number (150), structural holes, social identity |
Card 3: Constraint (约束映射) — Acknowledge Reality
Purpose: Distinguish "real obstacles" from "fake obstacles", turn real constraints into design parameters
Constraint Classification Table:
| Hard Constraints (Physical/Logical) | Soft Constraints (Resources/Conventions) | Pseudo-Constraints (Cognitive Myths) |
|---|---|---|
| Time irreversibility | Limited funds | "Must build own traffic" |
| Finite attention | Single skill | "Must use professional PR" |
| Light speed/bandwidth limits | Limited connections | "Must have authority before doing SEO" |
| Thermodynamics (2nd law) | Geographic location | "Content must be on my domain" |
| Google sandbox period (6-12 months) | Zero budget | "Must pay for ads to promote" |
Operation: Transform soft constraints into resource equations:
I have [A time] + [B skills] + [C existing assets], need to produce [D results]
Find: Non-linear leverage points — combinations where 20% resources yield 80% results
Card 4: Leverage (杠杆重构) — Reassemble
Purpose: Build asymmetric advantages within constraints using atomic units
Four Leverage Patterns:
-
Platform Parasitism (借势能)
- Don't build, just connect: use high-authority platforms as your "atomic units"
- Check: Are my target atomic units flowing on this platform? (e.g., Reddit's attention flow)
- Examples: Publish on Medium/Dev.to/Hashnode instead of waiting for your domain to rank
-
Time Arbitrage (换时差)
- Trade what you have in abundance (time) for what others lack (immediacy)
- Check: Can I do what others find too slow? (e.g., deep long-form articles, manual dataset compilation)
- Examples: Write comprehensive guides while competitors chase quick wins
-
Atomic Substitution (找等价物)
- Goal needs atomic unit A, but I have B. Can B be converted to A's input?
- Example: No "ad budget" (A), but have "coding ability" (B) → build tools that generate natural backlinks (B→A conversion)
-
Constraint Flip (变限制为壁垒)
- Your constraints are also others' constraints → establish expertise under that constraint to form a moat
- Example: When everyone lacks ad budget, your SEO capability becomes a competitive advantage
Practical Checklist
Before starting any problem-solving with first principles:
☐ Can I describe the minimum closed loop of this domain in one sentence?
Example: Business = Value Creation + Value Capture
☐ Are there mathematically inviolable theorems?
Example: Shannon limit in communication, Halting Problem in computation
☐ If all existing solutions were erased, which steps are absolutely essential to rebuild from stone age?
(Remove historical path dependencies)
☐ Which "common sense" elements are actually analogies/metaphors rather than atomic units?
Example: "Website authority" is analogy; "Eigenvector of link graph" is the atomic unit
Decision Tree: When You Encounter Unsolvable Sub-Problems
This sub-problem is:
├── Temporarily don't know, but can learn?
│ → Create learning plan, set deadline
├── Already solved by others, can acquire?
│ → Purchase / Partner / Outsource
├── Truly unsolvable (physical/logical limits)?
│ → Return to upper layer, modify decomposition path or goal
└── Solution cost too high, not worthwhile?
→ Find alternative solutions or narrow problem scope
Quick Reference: Meta-Leverage Matrix
Match your available resources to strategy:
| Your Resource | Leverage Strategy | Avoid |
|---|---|---|
| Abundant time | Do depth, automation, long-term assets | Manual grunt work (e.g., 1000 forum posts) |
| Coding ability | Tools generate content/data for natural spread | Pure tech blogs (unless solving specific problems) |
| Strong writing | Write "answers" on platforms, "systems" on your site | Stack articles on low-traffic site waiting for visitors |
| Weak connections | Use "data/insights" as social currency instead of networking | Cold email asking for shares (no trust atomic unit) |
Core Philosophy
First principles doesn't make you omniscient — it makes you the optimal designer under specific constraints.
Your limitations ("can only write articles and build small tools") are not flaws — they are precisely defined design parameters. With these parameters, you can design solutions that large companies cannot replicate (because they have budgets and disdain for this kind of "time-for-space dirty work").
Example Application: Zero-Budget SEO
Problem: New website, no authority, zero budget, only time and coding skills
Running SACL Framework:
S (Strip):
- Cross out "SEO = creating content on my own site"
- Cross out "Promotion = buying ads"
- Essence: In the information deluge, match specific demanders with my solution
A (Atomize):
- Demander unit: Search keywords (active seeking) vs social feeds (passive discovery)
- Meeting unit: Trust transfer (through intermediary endorsement) vs direct discovery (algorithm match)
- Your asset units: Code-generated tools (scalable) + explanatory ability (text)
C (Constraint Mapping):
- Hard constraint: Google sandbox period (6-12 month authority accumulation)
- Soft constraint: Zero budget → convert to "must acquire attention through non-monetary exchange"
- Pseudo-constraint busted: "Content must be under my domain" → change to "Content must be under my control AND can drive traffic"
L (Leverage Reconstruction):
- Selected leverage: Platform parasitism + atomic substitution
- Specific solution:
- Use coding to scrape high-frequency questions on Reddit/Zhihu (automate time investment)
- Write "code solutions to problems" on Dev.to/Medium (parasitize high-authority platforms)
- Open-source tools on GitHub (trade code for backlinks, substitute buying links)
- Article footer: "Full toolkit on my site" (traffic funnel, not direct SEO)
Usage Ritual
Turn this framework into a template. When facing any difficult problem:
- Spend 5 minutes filling out S-A-C-L four cards
- If stuck, decomposition isn't deep enough (return to A, ask: what is this physically?)
- If too many solutions, constraints aren't clear (return to C, mark hard constraints in red)
Key mindset: Knowing what you cannot do is equally important as knowing what you can do.