Week 9 Study Guide: Making Silicon Valley

Big idea

1. Silicon Valley is a place, but it is also a story about how technology works, a system for producing wealth and social order, and a way of shaping the future. The story makes all of this look natural.
2. The standard story has four layers: electronics, Stanford, silicon, computing. Its heroes are engineers and entrepreneurs, while the military and venture capital do the work underneath.
3. Because the story makes the Valley look inevitable, the winners of one wave get to ride the next: chips, PCs, the internet, platforms, now AI.
4. Innovation is an ideology. Most of life depends on maintenance, and most work is upkeep.
5. The AI business sells prediction as a financial asset. It depends on extracting data, minerals, and cheap labor.
6. AI runs on hidden human labor controlled through time discipline. It is the old factory logic with better sensors.
7. The military-industrial complex never left. The Valley now embraces it openly in the name of defending democracy.

Always ask who tells this story, who gets to be the agent of progress, and who is left out.

Nicholas, "Silicon Valley and the Emergence of Investment Styles"

Core idea: Three things built the Valley: universities, military spending, and a distinctive culture and legal climate. The cluster created demand for venture capital, and the Valley shaped how VCs invest.

• Three investment styles. Arthur Rock bet on people, Tom Perkins on technology, Don Valentine on market size.
• Stanford under Terman. Stanford owned its researchers' patents, built the Stanford Industrial Park, and encouraged students to start companies while they kept teaching. At MIT the same thing could get you fired.
• The military as early adopter. Until 1967 the military bought over half of all Bay Area integrated circuits. It paid for the learning curve, and chip prices fell from $31.60 to $2.33 in five years, which is what made a consumer market possible.
• Cost-plus contracts built the manufacturing base. When defense spending fell in the 1960s, Valley firms switched product lines while Boston's Route 128 did not adapt, and the industry moved west.
• Culture and law. The culture encouraged risk, accepted failure, and drew immigrants. California's Section 16600 voids non-compete agreements, so engineers could leave and start rivals. Fairchild exists because of it.

Vinsel & Russell, "Hail the Maintainers"

Core idea: Innovation is an overvalued ideology. Maintenance, repair, and infrastructure matter more for most people's lives.

• When faith in progress collapsed in the late 1960s, innovation took its place as a smaller word with no moral content.
• In the 1970s and 80s, innovation policy became the answer to Japan and Silicon Valley became the model every region tried to copy. Schumpeter's creative destruction was revived and Christensen made disruption a mantra.
• Innovative regions produce inequality. The Google bus protests in San Francisco showed this is a feature rather than a bug.
• Three corrections. Technology is bigger than innovation, since most of the objects around us are old and still in use. Infrastructure failures are symptoms of a broken politics. And innovators are about 1% of the workforce, while most work is upkeep done by maintainers.
• Gender and class. Cowan's More Work for Mother found that labor-saving appliances raised standards and created more housework. Low-wage service work is maintenance too.
• Caveats. Innovation is value-agnostic, since crack cocaine was a highly innovative product. The real move is from means to ends: who benefits, and to what end?

Fang, "Silicon Valley Goes to War"

Core idea: The connect-the-world ethos hid the Valley's Cold War roots. It now openly embraces the military-industrial complex, using a China arms race story that is self-serving and contradicts the democratic values it claims to defend.

• The shift. Y Combinator now funds missile startups, defense VC grew tenfold in five years, and Palantir's Karp argues the Valley wasted its engineers on delivery apps.
• Act 1, Cold War. Stanford was the cold war university, early chips went into missiles and satellites, and Lockheed was Santa Clara County's largest employer before the PC.
• Act 2, counterculture. After Vietnam, activists repurposed military tools like the computer for personal expression, which produced the Homebrew Computer Club. This anti-military culture peaked with Project Maven in 2018, when 4,000 Google employees forced the company to drop a Pentagon drone contract.
• Act 3, the return. Military work never stopped, as Palantir and Project Nimbus show. With the consumer internet saturated, the growing defense budget is the next market.
• The politics. Defense startups want to undo the 1993 Last Supper that consolidated the industry into five giant contractors. The arms race story, that whoever reaches AGI first wins, justifies huge data center spending even though AGI is not clearly China's priority. OpenAI, Anthropic, and Google have all dropped their limits on military work.
• The democracy paradox. The Valley claims to defend democracy from autocrats while building deportation surveillance and supplying wars that human rights groups condemn. Illiberal means undermine the democratic end.

Crawford, "Labor"

Core idea: Workplace AI is a return to old industrial logics of surveillance, time discipline, and deskilling, now tracked down to the micro-movement. The real question is how humans are treated like robots.

• Amazon. Time clocks hang everywhere, workers get 15 minutes off-task per shift, and robots glide behind fences while workers wear braces and buy painkillers from vending machines. Nobody can opt out of working with the machines.
• The lineage. Smith's division of labor, Marx's worker as an appendage of the machine, and Thompson's time discipline, which turned time into a currency and a site of struggle.
• The panopticon began as a workplace. Samuel Bentham designed an inspection house for a factory before his brother Jeremy turned it into a prison.
• Babbage saw the factory as a computer in which value came from designing the process and labor was only a source of error. Chicago's disassembly line realized this, and when The Jungle exposed it, the result was food safety law and nothing for workers. Power accepts regulation at the margins.
• Taylor and Ford live on at McDonald's, where algorithmic scheduling sets shifts by the second.
• Ghost work. AI depends on hidden crowdworkers who label data and moderate content, often below minimum wage. Some are paid to pretend to be AI, which Crawford calls Potemkin AI. Self-checkout is fauxtomation: the labor moves to the customer.
• Mechanical Turk. The 1770 original hid a chess master inside a machine. Amazon named its crowdwork platform after it and Bezos calls it "artificial artificial intelligence." AI is "neither artificial nor intelligent."
• Privatizing time. Google runs its own planetary clock, TrueTime. Controlling time has always been a way of controlling political order, which is why workers at Fordlandia smashed the time clocks and why going slow was called sabotage.
• The rate. Amazon workers in Minnesota walked out, and Amazon would discuss anything except the productivity rate, which it called "our business model." The slogan we are all tech workers builds solidarity across janitors and engineers.

Sloane, "The Business of AI"

Core idea: AI sells prediction as an asset, and its value is mostly speculative. The business rests on three ideas: data is infinite, what gets automated depends on social status, and prediction can replace human judgment.

• Financialization. Everyone agrees AI is big business and almost nobody can explain how it makes money. OpenAI was valued at 30 times its revenue while losing money. Predictions are hard to price, so AI is sold as subscriptions to infrastructure.
• Data is not infinite. Data was called the new oil, but oil needs refining, so ask who works the rigs. Participating in society means producing data, and refining it falls on underpaid workers in the Majority World. AI is also physically extractive through data centers and rare-earth chips.
• Status decides automation. Low skill measures credentials rather than expertise. London cabbies spend years learning the Knowledge, and Uber devalued it overnight. Humans still beat robots at picking strawberries, so farms are rebuilt indoors to suit the machines. In reverse, once AI can code, coding loses status and the job becomes fixing AI output.
• Discretion. AI predicts first and humans interpret. Recruiters lost their precise Boolean search to black-box semantic search, so they resist by gaming the system, which adds work and lets bias back in. A dermatologist uses the risk score to learn and to persuade patients, and ignores it when he is sure. Recruiting is feminized and treated as automatable, while high-status medicine gets carefully designed AI.

Lecture 22: Making Silicon Valley 1

Core idea: The standard story is about materials, technologies, and entrepreneurs. Underneath are networks, the military, and capital, and the story naturalizes the Valley so it looks inevitable.

• Before all this, the Santa Clara Valley was the Valley of Heart's Delight, orchards on unceded Ohlone land.
• Electronics. Early radio in San Jose, vacuum tubes, and ship-to-shore wireless that served US naval power.
• Stanford. Terman, HP, radar tubes in WWII, SLAC. Stanford put a liberal-arts sheen on defense research and tied itself to local bankers and land developers.
• Silicon. The transistor at Bell Labs, Shockley moving west, the traitorous eight founding Fairchild, the integrated circuit, then Intel. Eisenhower's farewell address in 1961 warned against the military-industrial complex.
• Computing. Mainframes, then PCs, the internet, and smartphones.
• Keys to the story: a regional network effect, free movement of labor, the military-industrial complex, West Coast ambition, a chart that goes up and to the right, and venture capital.
• Who gets seen as the agent of history? The Ohlone, the Navy, and Stanford's land deals rarely appear.

Lecture 23: Making Silicon Valley 2

Core idea: Add venture capital and success narratives to the story. Platforms then turned data into a business model and a form of capital.

• The Californian Ideology fuses hippie counterculture with free-market libertarianism through faith in technology.
• The garage myth says founders came from nothing. The repeated pattern points to what it hides: Stanford, networks, and capital.
• Disruptive innovation and the lean startup became an operating manual for VCs and founders, and a culture. Move fast and break things slides into fake it till you make it, as SBF and Elizabeth Holmes showed.
• The dot-com bust did not end the story. Survivors rode the next wave of smartphones and social media.
• The platform economy means multi-sided markets, an enclosed space for data collection, and network effects that tend toward monopoly.
• The data imperative: everything can and should be rendered into data and accumulated.

Lecture 24: Making Silicon Valley 3

Core idea: Platforms led to gig labor and the cloud, then to the AI economy, then back to the military-industrial complex and into politics.

• Platforms run a cybernetic nudge loop: a user decides, the decision becomes data, and the choice architecture is adjusted to steer the next decision.
• The gig economy and microlabor feed AI, from Mechanical Turk to data annotation shops.
• Scaling laws showed that models improve predictably with more compute, data, and parameters. Scale became the strategy and capital the moat, so Nvidia and compute are now strategic resources.
• What is AI worth? The factors of production are data, compute, expertise, labor, and energy. The strategies of capture are frontier models, hyperscalers, conglomerates, and open source. Ask who controls each.
• Defense tech is back, and Starlink in Ukraine shows a private company acting as a sovereign power.
• Disruptive innovation as politics. Andreessen's firm backed Trump, tech CEOs lined up at the inauguration, and Thiel had long argued that freedom and democracy are incompatible.

How it connects

• The military is in every layer, from Navy radio and radar to missile chips and Lockheed, and now Palantir, Anduril, and OpenAI's Pentagon contracts. The anti-military decades were the exception.
• Capital drives the narrative. Investment styles became disruption as an ethos, then scaling-law bets, then valuations detached from revenue. Innovation is the buzzword that hides it.
• Labor is the hidden layer. Maintainers, ghost workers, the rate, devalued cabbies and recruiters, and data annotators are left out of the Valley's story. As in Week 1, these artifacts have politics.