🐙practice(tesla, betterplace)

[[09-25|25-09-25]] \subsection{Hook: Tale of Two Ventures}\label{sec:intro-hook} In 2008, two ventures set out to reinvent the automobile with electric vehicles: one would rise to become a trillion-dollar icon, the other vanish into obscurity, though at the outset both looked equally promising with elite teams, proven founders, and massive funding ($60 million and $200 million, respectively). Tesla soon faced a 10-week supply chain delay that stranded $17,000 batteries as defective inventory, yet instead of paralysis they pivoted overnight, embracing what became “production hell” by localizing production, relaxing rigid specifications, and revising promises from “250 miles” to “200+,” thereby turning constraint into survival. Better Place, by contrast, seemed blessed at first as its Denmark pilot reached 15\% adoption and executives declared victory, doubling down on “3-minute battery swaps,” securing 100,000 Renault commitments, and building 38 identical stations, only to discover that while Danish pilots parked on open ground accessible to robots, Israeli customers parked underground where the system failed—success in Copenhagen collapsing in Tel Aviv. The contrast reflected management philosophy: Tesla’s VP Dunlay emphasized “perfect is the enemy of good; our gates are always open,” while Better Place’s CEO Agassi dismissed rivals like GM’s Volt as “stupid” and surrounded himself with loyalists, even hiring his girlfriend’s friends. In practice, Tesla’s $60 million cultivated learning velocity, whereas Better Place’s $850 million fueled rigidity, with capital amplifying opposite instincts—constraint forcing rapid experimentation, abundance funding fatal certainty. Ultimately, the divergence traces to $\tau$, the precision of initial promises: high precision locks beliefs and resists updating, while low precision enables adaptation; Tesla survived by staying vague, and Better Place perished by defending specifics. **1. Evidence from Better Place FastCo article:** **Supporting OIL thesis:** - "Better Place formally committed to order 100,000 Renault cars between 2011 and 2016" while actual sales: 1,500 total. High τ locked them into impossible commitments. - Battery swap stations "projected to cost no more than $500,000 each; they cost $2 million." When precision exceeded reality by 4x, they couldn't adapt infrastructure design. **Criticizing OIL thesis:** - "There was profligacy, marketing problems, hiring problems, problems with every conceivable part" suggests execution failures, not precision issues. - "Agassi brought Chotoveli to meetings and hired some of her friends to top positions"—cronyism and narcissism, not τ, killed the company. **2. Tesla vs Better Place Comparison Table:** |**Dimension**|**Tesla (Low τ)**|**Better Place (High τ)**| |---|---|---| |**Initial Capital**|$60M by 2007|$200M seed (3.3x Tesla)| |**Promise Precision**|"200+ miles"|"3-minute swaps, anytime, anywhere"| |**Infrastructure Cost**|Avoided commitment|$500K → $2M per station (4x overrun)| |**Supply Chain Response**|10-week pipeline killed → pivoted to local|Locked into Denmark design → couldn't adapt for Israel| |**Customer Promises**|"Emotional relationship" forgave defects|"Half the price" → Same price as Toyota Corolla| |**Market Entry**|100 Roadsters → Learn → Model S|100,000 cars committed → 1,500 delivered| |**Burn Rate**|$78M loss on $60M capital|$500K/day ($180M/year)| |**Final Status**|$1 trillion market cap|$850M burned, bankrupt| **Key insight:** Tesla's vague promises ("gates always open") enabled iteration through crisis. Better Place's precise commitments ("100,000 cars," "3-minute swaps") created contracts and infrastructure that couldn't bend when apartment dwellers couldn't access underground swaps. [[09-19|25-09-19]] 비아오텍 사람들에게 내 이론 소개할땐 [[🐙practice(moderna, theranos)]] 가 직방 ## 2. Tesla vs Better Place: Two Experiments, Two Meanings ### Better Place: Experiment as Demonstration ``` Purpose: "Demonstrate a known fact" - The 5-minute battery swap was THE answer - Experiments were theatrical performances for investors - Each test was designed to confirm, not discover - 700 employees running one giant demonstration Result: When reality diverged, they had no pivot mechanism ``` ### Tesla: Experiment as Discovery ``` Purpose: "Make a discovery" - Each car model was a hypothesis test - Roadster: "Can we make electric sexy?" - Model S: "Can we make it practical?" - Supercharger: "What if charging were free?" Result: 15+ pivots, each experiment revealed new paths ``` ## 3. The Cycle: Ignorance → Promise → Experiment ignorance promise experiment ## 현지의 종합 표: The Common Knowledge Trap | Dimension | Tesla (Success) | Better Place (Failure) | Key Insight | | :------------------------ | :----------------------------------------- | :---------------------------- | :------------------------------------------ | | **Epistemological Path** | 무지→약속→실험 (Ignorance→Promise→Experiment) | 약속→실험 (Promise→Experiment) | Missing ignorance stage = $850M loss | | **Initial τ (Precision)** | τ ≈ 0.1 (very loose) | τ ≈ ∞ (completely rigid) | Start with low precision, earn it gradually | | **Promise Nature** | "Someday, somehow" | "5 minutes exactly" | Strategic ambiguity preserves options | | **Experiment Type** | Discovery-oriented | Demonstration-oriented | Experiments should reveal, not confirm | | **Pivots** | 15+ strategic adaptations | 0 pivots | Adaptation signals learning | | **Initial Team** | 270 employees | 700 employees | Smaller teams = lower integration cost (i) | | **Complexity (c)** | 5 core subsystems | 15 interdependent systems | Deliberately simplify complexity | | **Integration Cost (i)** | Low (lean structure) | High (heavy organization) | i×c must stay manageable | | **Common Knowledge** | Strategically delayed | Prematurely crystallized | Prevent ignorance from becoming public | | **DNA Metaphor** | Loose→Gradually tightening | Tight from start→Snapped | Dynamic tension management | | **Optimal τ Formula** | τ* = V/(i×c) - 1 ≈ 2 | τ* = V/(i×c) - 1 ≈ 0.05 | Actual τ must match optimal τ | | **Information Strategy** | Selective absorption | Complete information pursuit | "통합비용과 복잡도 관리 후 정보 흡수" | | **Result** | $1 trillion value | $850 million loss | 1000x+ difference | ### 핵심 공식과 원칙 __τ_ = V/(i×c) - 1_* - V = Venture value potential - i = Information integration cost - c = Environmental complexity - τ = Optimal precision level **창업가의 원칙:** "통합비용(i)과 복잡도(c)를 먼저 낮춘 후에야 정밀도(τ)를 높여라" **Four Principles:** 1. Embrace private ignorance 2. Maintain public ambiguity 3. Delay consensus formation 4. Preserve pivot options --- ## PRHC Evidence Matrix: 실증적 약속 설계의 해부 | **단계** | **Tesla Motors (2006-2008)** | **Better Place (2007-2010)** | **이론적 함의** | | ---------------------------------- | -------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------ | -------------------------------------- | | **P** _Parameterize_ (매개변수화) | **2006.7.19**: "250 miles on a single charge" • 공식 보도자료 언어 • "up to" 한정어 포함 • EPA Highway 기준 명시 | **2007.10**: "3-minute battery swap" • Agassi 개인 약속 • "exactly" 강조 • 주유소 비교 반복 | 초기 약속의 주체(기업 vs CEO)와 정밀도가 미래 유연성 결정 | | **R** _Regularize_ (정규화) | **2007.초**: "more than 200 miles" • Martin Eberhard CEO 공식 수정 • 하한선만 명시 (상한 개방) • "original goal" 프레이밍으로 실패 아닌 조정 | **2008.5**: "$500,000 per station" • "주유소의 절반 가격" 고정 • "25마일마다 설치" 공식화 • 100,000대 목표 명시 | 제약 조건을 수치화하되, 범위(range)로 표현하면 조정 공간 확보 | | **H** _Hierarchize_ (계층화) | **2007.5.10**: Andrew Simpson 블로그 • "personally driven 200 miles" • 개인 경험을 증거로 제시 • 다층적 검증 구조 구축 | **2008-2009**: 다중 버전 난립 • "3분" → "5분" → "59.1초" • 각 데모마다 다른 수치 • 일관성 없는 메시지 계층 | 약속의 계층 구조가 명확해야 신뢰 유지; 모순된 계층은 신뢰 붕괴 | | **C** _Calibrate_ (조정) | **2007.12**: EPA 221마일 재조정 • 변속기 이슈 투명 공개 • 2008 실제: 244마일 달성 • 약속 하한(200+) 초과 성공 | **2010.도쿄**: "59.1초 평균" • 과도한 정밀성 주장 • 2012: 단일 모델만 지원 • 2013.5: 파산 | 보수적 약속 → 초과 달성 가능 과대 약속 → 조정 불가능 | ![[communcation(tesla, betterplace) 2025-09-08-20.svg]] %%[[communcation(tesla, betterplace) 2025-09-08-20.md|🖋 Edit in Excalidraw]]%%