The Challenge of the Future

1. The History of Technology: Stagnation Outside Computers

From the late 17th century to the late 1960s, technological progress was relentless. Industrialization marked a deep shift in how humans get rich: from capturing value taken from others to creating new value through trade and invention. Of the roughly 100 billion people who have ever lived, most spent their lives in essentially static societies — the last few centuries are the great exception, and the 1960s were the peak of confidence that the next 50 years would bring unprecedented progress.

• Median wages have been flat since 1973; people run an Alice-in-Wonderland race, working harder just to stay in place.
• Per capita incomes still rise, but at ever-slower rates.
• Computing is the happy exception: Moore's Law, Kryder's Law, and their cousins have largely held.

Because computing is the one place where the machinery of progress still works — and the engine of Silicon Valley — computer science is the natural staging ground for restarting progress everywhere else.

2. Two Kinds of Progress: Globalization vs. Technology

Horizontal (extensive) progress means taking things that work and replicating them; its one-word name is globalization. China is the paradigm: its 50-year path is largely to become what the developed world already is, perhaps skipping a few steps. Even the phrase 'developed nation' smuggles in pessimism — it implies the frontier is finished and everyone else just needs to catch up. Vertical (intensive) progress means doing genuinely new things; its one-word name is technology.

The two interact: if everyone copies the resource-heavy Western lifestyle, environmental and resource constraints will bite. When scaling from 1 to n hits a wall, only new 0-to-1 technology can break through it — so technology matters even if globalization is all you care about.

3. Why 0 to 1 Is So Hard

Society defaults to 1 to n because copying is simply easier. Whoever attempts 0 to 1 faces the problem of exceptionalism: a founder claiming to do what no one has done must ask whether that is insight or delusion. Hollywood shows the dark side — roughly 20,000 people move to Los Angeles each year convinced they will become stars, and almost none do. Startups suffer this less than Hollywood, but not zero.

Education can't fix this, because education is 1 to n at its core: watch, imitate, repeat. Learnable mechanics — incorporating properly, pitching VCs — get you maybe 30% of the way; the decisive leap can't be taught. That is also why business-school case studies mislead: successful companies each solved the 0-to-1 problem in their own unrepeatable way.

All failed companies are the same; they botched the 0 to 1 problem.

New things also break statistics: with a sample size of one, the standard deviation is infinite, so probabilistic thinking has nothing to grip. The alternative is calculus-style determinism — Apollo engineers computed exactly where the moon would be. But our society calls people who claim to know the future prophets, and treats all prophets as false; Steve Jobs walked that line about as closely as anyone can.

No one would want to ride in a statistically, probabilistically-informed spaceship.

4. Four Theories of the Future

Where does vertical progress go from here? Thiel sketches four candidate trajectories.

1. 1

   Convergence

   Growth took off with industrialization but will slow and flatten toward an asymptote.

2. 2

   Cyclical

   Progress advances and retrenches forever. True for most of history, but implausible now — accumulated knowledge won't simply be lost and relearned.

3. 3

   Collapse

   Some technological advance ends up destroying civilization.

4. 4

   Singularity

   Technology accelerates toward an event horizon — an AI, say — beyond which prediction breaks down.

Thiel's bet: people overestimate the convergence and cyclical theories, and correspondingly underestimate collapse and singularity.

5. Why Companies — and Why Startups

Why organize technological work in companies at all? The Coase Theorem answers: firms exist where internal and external coordination costs balance. A totalitarian state has near-zero external coordination costs — it just commands — but its internal costs are crippling, which is why central planning fails, as Hayek and the Austrian School showed. A lone contractor has zero internal costs but must negotiate every single relationship. Firms settle at the optimum in between.

Past about 100 employees the character of a firm changes: people no longer all know each other, politics arrive, and signaling that work is being done starts to beat doing it. Multi-floor or multi-city offices, hired consultants, and outsourced key engineering are red flags investors take seriously. Path capping friends at 150 — echoing ancient tribal sizes — points to the same natural limit. Startups matter precisely because they are small enough to escape these costs.

• The negative reason to found: you simply can't build new technology inside big companies, governments, or nonprofits — their bureaucracies can neither pay people right nor give them real recognition.
• Money is a weak motive: the class cited research that happiness stops tracking income around $70,000 a year.
• Fame is a dubious motive; wanting to change the world is a better one — the American founding in 1776 was itself a kind of startup.

The costs of failing are misread. The financial hit is smaller than people assume; the real damage is nonfinancial — a failed startup may teach you nothing except how to fail, and leave you more risk-averse. A failed 0-to-1 attempt at least teaches you a great deal; a failed clone — 'Groupon for Madagascar' — leaves you nowhere. And since people are not lottery tickets, 'just try again' is not a strategy.

6. Where to Start: Three Questions

• What is valuable?
• What can I do?
• What is nobody else doing?

The first question separates business from academia: academia's cardinal sin is plagiarism, not triviality, so much of its output is esoteric and useless — but no company survives on non-valuable work. The second demands real capability, not talk. The third, the most often overlooked, guards against slipping back into copying.

What important truth do very few people agree with you on?

The business version asks: what valuable company is nobody building? The test is disagreement itself. 'Our education system is broken' fails — not because it's false, but because everyone already agrees, which is why education startups crowd in, mostly running in globalization mode. A right answer has the shape 'most people believe X, but the truth is !X.' Finding one is rare and tricky, but the search itself is richly rewarding.

Then vs. now (2026)

Self-check quiz

Pick an answer to reveal the explanation.