At the Threshold of the Compute Era, Looking Back at the Dawn of Electricity
What the adoption of electricity, telephony, television, computing, and the internet can teach us about the recurring tests of the AI era.
This series began with an intuition.
When electricity first appeared, European nobles were eager to try its novelty and convenience, but only a few with genuine power needs would transform their production with it. Today’s compute and AI are in exactly the same position: no matter how the models evolve, one day they will be available to everyone and affordable by everyone. Electricity was never scarce, and neither will compute be. What is truly scarce is the ability to envision what new lifestyles people can create on top of it after it becomes ubiquitous.
Following this intuition, I discovered that history not only provides us with a metaphor but a complete set of graded exams—electricity, the telephone, television, computers, the internet, five revolutions of general-purpose technologies, each of which gave the same exam to a generation of entrepreneurs, and each generation failed in the same places. This series aims to lay out this exam, along with the answers of past generations, the rare correct answers, and the mistakes, on the table of the compute era.
The series currently contains four pieces:
The People Swept Away by Electricity is a list of wrecks. It tallies the failures, deaths, and detours in the history of electricity’s widespread use: Edison’s defeat with direct current, the ongoing frequency wars that still divide Japan’s power grid, the bodies of linemen hanging above Manhattan and spitting flames, the collapse of Insull’s financial empire, the bet on electric vehicles a century too early, and the Telharmonium—“streaming” half a century ahead of its time. History is written by the victors, but the lessons are hidden in the losers’ accounts.
Recombination is the first exam question. The value of general-purpose technology does not lie in the technology itself, but in the recombination it enables—movies only became movies when they stopped imitating plays, and factories only had assembly lines when they removed their drive shafts. New technologies are always first dressed in the clothes of the old era, and the revolution begins when someone undresses them.
The Migration of Scarcity is the second exam question. Technological revolutions do not eliminate scarcity, they merely relocate it: as something becomes abundant, its value shifts to its complements. Losers cling to the ruins of old scarcity (copper wires, printing presses, whole machine manufacturing), while winners move ahead to new bottlenecks (content, operating systems, attention).
Translate the User is the third exam question. Ford’s “faster horses” are often misread as “ignore the users,” but its true meaning is the opposite: users are always right, but they express symptoms, and entrepreneurs must diagnose the causes and prescribe solutions that users cannot imagine. Western Union and AT&T both made the same mistake ninety years apart, and Kodak died by serving film rather than memory—companies that fail to understand their users die by taking their words literally.
The three exam questions together form one sentence: Technology is a variable, demand is a constant, structure is the answer. The mediocre chase variables, the failures cling to old structures, and those who transcend time do the same thing—fitting constant human needs into the new structures allowed by new technologies.
This series will not end here. It is the beginning of a book. What will follow is applying this framework to every specific corner of the compute era: which industries still operate under “steam engine layouts,” where scarcity is currently migrating, what today’s “faster horses” look like, and—just as in every previous cycle—the bubbles, scams, regulation, and accidents of this round will reappear in what form.
Over a hundred and forty years, this exam has been given five times, and the answers are already public, but the people who get it right are still few. Not because the questions are hard, but because getting it right requires betraying one’s existing assets, identity, and profits. Knowledge can be inherited, but betrayal must be done personally.
This book is for those who are willing to do it personally.