Exercises: Field Autopsy — Technology

Part A: Comprehension and Application

A.1. Trace the neural network suppression from 1969 to 2012. For each failure mode that operated (authority cascade, consensus enforcement, sunk cost, Einstellung, outsider problem, Planck's principle), provide one specific mechanism through which it suppressed neural network research.

A.2. Define "capital-sustained error" and explain how it differs from the error-sustaining mechanisms in medicine and criminal justice. Why does capital change the error dynamics?

A.3. Explain how the dot-com bubble illustrates the feedback loop between narrative, capital, and the appearance of progress. At what point did the feedback loop disconnect from the underlying reality?

A.4. The "connecting the world" narrative functioned as an unfalsifiable thesis (Chapter 3). Explain its unfalsifiable structure: how were positive outcomes and negative outcomes both interpreted as consistent with the narrative?

A.5. Define the "disruption myth" and explain how it functions as tech's version of the revision myth (Chapter 20). How does it pre-delegitimize criticism?

Part B: Analysis

B.1. Apply the Correction Speed Model to the technology sector. The chapter argues that tech corrects fast on technical questions and slow on narrative questions. Identify the specific model variables that explain this asymmetry. Why does market feedback work for technical correction but not for narrative correction?

B.2. Compare the neural network suppression to the dietary fat hypothesis. Both involved authority cascades that sustained wrong consensuses for decades. Identify three structural similarities and two structural differences. Which correction was faster, and why?

B.3. The crypto ecosystem used the defense "that wasn't real crypto" to explain away specific failures, just as strategic bombing advocates used "that wasn't real bombing" (Chapter 28). Analyze this pattern as a form of unfalsifiability (Chapter 3). Is there any evidence that could, in principle, disprove the core crypto thesis? If not, what does that tell us about the thesis's epistemic status?

B.4. The autonomous vehicle timeline predictions illustrate a pattern of "confusing rate of initial progress with rate of completion." Identify three other technology predictions that made the same error. What structural features of technology development make this error particularly common?

Part C: Synthesis and Evaluation

C.1. The chapter argues that the technology industry is subject to the same structural failure modes as every other field, despite its self-image as inherently disruptive. Evaluate this argument. Is tech genuinely different in any structural way, or is the disruption self-image itself a failure mode? Support your answer with specific evidence.

C.2. The neural network case illustrates that a correct idea can be suppressed for decades and then vindicated. Does this mean that every suppressed idea is potentially correct? How would you distinguish between a genuinely correct idea that is being suppressed by institutional forces and an incorrect idea that is being appropriately rejected? Use the framework from this book to develop diagnostic criteria.

C.3. If you could design a technology industry that was resistant to capital-sustained error, what structural features would it need? Consider: how would funding decisions be made? How would hype cycles be dampened? How would narrative claims be tested? What would you sacrifice (speed, risk-taking, optimism) and what would you preserve?

Part D: Mixed Practice (Interleaved)

D.1. A technology company has invested $3 billion in a thesis that is increasingly contradicted by evidence. Using the capital-sustained error framework from this chapter AND the sunk cost framework (Chapter 9) AND the incentive structures framework (Chapter 11), predict how the company, its investors, and its employees will respond to the mounting evidence. What would it take for the company to pivot?

D.2. A venture capital firm is evaluating a startup that claims to be "disrupting" an established industry. Using the disruption myth analysis from this chapter, the plausible story framework (Chapter 6), and the survivorship bias framework (Chapter 5), design a due diligence process that distinguishes between genuine disruption and narrative-market fit.

D.3. The AI winter ended when hardware caught up with theory. Is a similar dynamic possible in any field you know — a correct idea that is currently suppressed or marginal, waiting for some external factor (technology, data, political change) to enable its vindication? Identify a candidate and apply the Correction Speed Model to predict when the vindication might occur.