The 2,400-Year-Old Idea That Built SpaceX

“What is a rocket actually made of? What are the raw materials worth?”

When Elon Musk asked this question in 2002, everyone told him rockets were simply expensive — $65 million per launch was just the market price. But Musk decomposed the problem to first principles: raw materials (aluminum, titanium, copper, carbon fiber) accounted for only 2% of the typical rocket price. The “Idiot Index” — ratio of finished product cost to material cost — was approximately 50x.

The result? SpaceX reduced launch costs from $54,000/kg to $2,720/kg, with a target of $100/kg — a 99.8% reduction.

This is first principles thinking in action. And the idea is 2,400 years old.

From Aristotle to Musk

The concept originates from ancient Greek philosophy. Aristotle defined first principles (archai) in his Metaphysics as “the first basis from which a thing is known” — fundamental propositions that are “true, primary, and indemonstrable.”

René Descartes later developed the method of systematic doubt — doubting everything until reaching indubitable truths. His famous conclusion: “I think, therefore I am” (Cogito ergo sum).

Marcus Chown, a British physicist who studied under Richard Feynman at Caltech, popularized these ideas in his 2023 book. Feynman’s famous epigram captures the essence: “What I cannot create, I do not understand.”

The Core Method: Deconstruct & Reconstruct

First principles thinking is a two-phase process:

Phase 1: Deconstruct — Break down complex problems into their most fundamental, irreducible units. Like disassembling a machine to individual components, then identifying the physical laws behind them.

Key question: “What is the most basic truth here? What cannot be further simplified?”

Phase 2: Reconstruct — Based on these fundamental principles, redesign solutions from scratch — ignoring existing frameworks, experience, and conventions.

Key question: “If I were building this for the first time, knowing only the fundamental truths, what would I create?”

Three Cognitive Traps

The video identifies three traps that prevent people from using first principles thinking:

Trap 1: Herd Mentality — “Everyone does it this way, so it must be right.”

Asch’s conformity experiments (1951) showed that 75% of people conform at least once. When ONE ally breaks ranks, conformity drops 80%.

Trap 2: Path Dependency — “We’ve always done it this way.”

Brian Arthur (1989) proved mathematically that in increasing-returns markets, winners are path-dependent, not necessarily optimal. The QWERTY keyboard was designed in 1873 to SLOW typing — yet 99%+ still use it.

Trap 3: Surface Illusion — “This is too complex; there are no patterns.”

Kahneman’s research shows System 1 handles 96% of decisions automatically. We default to pattern-matching, not deep analysis.

The Three-Step Method

Step 1: Trace to Origins — Use the “5 Whys” method (from Toyota) to decompose problems until reaching fundamental principles.

Step 2: Reconstruct — Forget existing frameworks. Design from the first principle upward.

Step 3: Verify — Test with small samples, collect data, iterate.

Verified Case Studies

Tesla Batteries: In 2006, battery packs cost $600/kWh. Musk’s first principles analysis: raw materials on London Metal Exchange = $80/kWh. The gap was purely manufacturing inefficiency. By 2023, industry average reached $128/kWh.

Amazon Prime: Bezos asked: “What do customers truly want?” Answer: fast delivery, wide selection, low prices. Prime (2005) was initially criticized but became Amazon’s core profit driver.

WeChat: Zhang Xiaolong built WeChat from a single first principle: connecting people. From messaging to Moments to Mini Programs, every feature iterated from this foundation.

Meituan: Wang Xing deconstructed dining: users need convenient, fast, good food — not the dine-in experience. Result: $60B+ market cap, 700M+ users.

When First Principles Fail

First principles is not a universal solution:

  • Aristotle was wrong for 2,000 years — He used first principles to conclude heavy objects fall faster than light ones. Galileo empirically disproved it.
  • Philip Anderson’s “More is Different” (1972) — Complex systems have emergent properties not computable from parts alone.
  • The Turkey Problem (Taleb) — A turkey fed for 1,000 days concludes caregivers are benevolent. On day 1,001 (Thanksgiving), its model catastrophically fails.

8 Actionable Recommendations

  1. Practice the “5 Whys” daily — Ask “why” five times to reach root cause
  2. Identify your cognitive traps — Name the trap to escape it
  3. Seek disconfirming evidence — Search for what contradicts your beliefs
  4. Build one ally relationship — Asch showed one dissenter reduces conformity by 80%
  5. Apply the 1% rule — 1.01^365 = 37.78x improvement in one year
  6. Question “best practices” — Ask what first principle they serve
  7. Embrace “beginner’s mind” — Pretend you’ve never seen the problem before
  8. Accept the limits of FP — Complex adaptive systems require emergent thinking

What first principle would you start from if you were rebuilding your work today?

第一性原理思維：穿透表象，直達本質

建立 SpaceX 的 2,400 年古老思想

「火箭究竟是由什麼組成的？原材料值多少錢？」

2002年，當 Elon Musk 問這個問題時，所有人都告訴他火箭就是很貴——每次發射 6,500 萬美元只是市場價格。但 Musk 將問題分解到第一性原理：原材料（鋁、鈦、銅、碳纖維）只佔典型火箭價格的 2%。「傻瓜指數」——成品成本與材料成本的比率——大約是 50 倍。

結果？SpaceX 將發射成本從每公斤 54,000 美元降低到 2,720 美元，目標是 100 美元/公斤——降低了 99.8%。

這就是第一性原理思維的實際應用。而這個想法已經有 2,400 年的歷史。

從亞里士多德到馬斯克

這個概念源自古希臘哲學。亞里士多德在《形上學》中將第一性原理（archai）定義為「事物被認知的第一個基礎」——「真實、基本且不可證明」的基本命題。

笛卡爾後來發展了系統性懷疑的方法——懷疑一切，直到達到不可懷疑的真理。他著名的結論：「我思故我在」（Cogito ergo sum）。

英國物理學家 Marcus Chown 曾在加州理工學院跟隨費曼學習，他在 2023 年的書中普及了這些思想。費曼著名的格言捕捉了本質：「我無法創造的，我就不理解。」

核心方法：拆解與重構

第一性原理思維是一個兩階段過程：

第一階段：拆解 — 將複雜問題分解為最基本、不可約的單位。就像把機器拆卸到個別組件，然後識別背後的物理定律。

關鍵問題：「這裡最基本的真相是什麼？什麼不能再簡化？」

第二階段：重構 — 基於這些基本原理，從零開始重新設計解決方案——忽略現有框架、經驗和慣例。

關鍵問題：「如果我從頭開始建造，只知道基本真相，我會創造什麼？」

三個認知陷阱

影片識別了三個阻止人們使用第一性原理思維的陷阱：

陷阱一：從眾心理 — 「每個人都這樣做，所以一定是對的。」

Asch 的從眾實驗（1951年）顯示，75% 的人至少從眾一次。當一個盟友站出來反對時，從眾率下降 80%。

陷阱二：路徑依賴 — 「我們一直都是這樣做的。」

Brian Arthur（1989年）在數學上證明，在報酬遞增市場中，贏家是路徑依賴的，不一定是最優的。QWERTY 鍵盤在 1873 年設計時是為了放慢打字速度——但 99% 以上的人仍在使用它。

陷阱三：表象迷惑 — 「這太複雜了；沒有規律可循。」

Kahneman 的研究顯示，系統一處理 96% 的決策。我們默認使用模式匹配，而非深度分析。

三步方法

步驟一：溯源 — 使用「5個為什麼」方法（來自豐田）分解問題，直到達到基本原理。

步驟二：重構 — 忘記現有框架。從第一性原理向上設計。

步驟三：驗證 — 小樣本測試，收集數據，迭代。

驗證案例

特斯拉電池： 2006年，電池組成本為 600 美元/千瓦時。Musk 的第一性原理分析：倫敦金屬交易所原材料 = 80 美元/千瓦時。差距純粹是製造效率問題。到 2023 年，行業平均達到 128 美元/千瓦時。

Amazon Prime： Bezos 問：「客戶真正想要什麼？」答案：快速配送、廣泛選擇、低價格。Prime（2005年）最初受到批評，但成為 Amazon 的核心利潤驅動力。

微信： 張小龍從一個第一性原理建立微信：連接人。從即時通訊到朋友圈到小程序，每個功能都從這個基礎迭代。

美團： 王興解構餐飲：用戶需要方便、快速、好的食物——不是堂食體驗。結果：市值超過 600 億美元，7 億以上用戶。

第一性原理何時失敗

第一性原理不是萬能解決方案：

  • 亞里士多德錯了 2,000 年 — 他用第一性原理得出重物比輕物下落快的結論。伽利略用實驗證偽了它。
  • Philip Anderson 的「多即不同」（1972年） — 複雜系統具有不能從部分計算的湧現屬性。
  • 火雞問題（Taleb） — 一隻火雞被餵養了 1,000 天，得出照顧者是仁慈的結論。在第 1,001 天（感恩節），它的模型災難性地失敗了。

8 個可執行建議

  1. 每天練習「5個為什麼」 — 問五次「為什麼」以達到根本原因
  2. 識別你的認知陷阱 — 命名陷阱以逃脫它
  3. 尋求反證 — 尋找與你信念相矛盾的证据
  4. 建立一個盟友關係 — Asch 顯示一個反對者將從眾率降低 80%
  5. 應用 1% 法則 — 1.01^365 = 一年內提升 37.78 倍
  6. 質疑「最佳實踐」 — 問它們服務於什麼第一性原理
  7. 擁抱「初學者心態」 — 假設你從未見過這個問題
  8. 接受第一性原理的局限 — 複雜適應系統需要湧現思維

如果今天你要重建你的工作，你會從哪個第一性原理開始？

References:

  • Aristotle, “Metaphysics” (384-322 BC)
  • Marcus Chown, “The One Thing You Need to Know” (2023)
  • Elon Musk, CNBC Interview on Battery Costs
  • Asch, S. (1951). “Effects of group pressure upon the modification and distortion of judgments”
  • Brian Arthur, W. (1989). “Competing Technologies, Increasing Returns, and Lock-In”
  • Kahneman, D. (2011). “Thinking, Fast and Slow”
  • Philip Anderson (1972). “More is Different”
  • Bilibili Video: BV1dyXqBaETY by 認知進化的Eleven