---
title: AI Is Doing to Thinking What the Industrial Revolution Did to Muscle, Sequoia Argues
description: Sequoia partner Konstantine Buhler argues the cost of machine thinking is falling tenfold a year and the AI era's defining application is not yet built.
author: Darie Nani (Editor-in-Chief)
date: 2026-09-08T14:38:38.447Z
updated: 2026-09-08T19:02:26.243Z
canonical: https://www.sovereignmagazine.com/article/sequoia-cognitive-revolution-ai-industrial
image: https://cdn.nanimediahouse.com/sequoia-cognitive-revolution-5180.webp
categories: Artificial Intelligence
content_type: Analysis
region: Global
publication: Sovereign Magazine
schema_type: Article
---

Sequoia Capital partner Konstantine Buhler published an essay on September 2 arguing that artificial intelligence is doing to human thinking what the Industrial Revolution did to human muscle. Over the past two centuries machines took over almost all physical work, he writes, and now machines are taking over cognitive work. He calls this the Cognitive Revolution, and he argues that the price of machine thinking is falling so fast that people will buy far more of it than humans could ever have supplied.

## Buhler Opens With a Waymo Ride That Uses Neither His Muscle Nor His Mind

Buhler opens with his morning commute in a Waymo. The car does physical work, moving five thousand pounds of metal and glass across the San Francisco Bay Area, and it does cognitive work, reading the road and making predictions. Those predictions, he writes, result in 17 times fewer serious-injury crashes than human drivers. Buhler writes that the ride uses none of his muscle and none of his mind, because the car has taken over both.

He argues that physical work went from about 99 percent biological to about 99.9 percent machine over roughly two centuries, and that cognitive work is about to make the same change, from about 99 percent human a century ago to about 99.9 percent machine in the near future. That will happen, he writes, "not because humans will think less, but because machines will compute much much more."

## Half of a $120 Trillion World Economy Has Barely Started to Mechanize, Buhler Writes

Buhler puts the world economy at about $120 trillion and says it splits roughly evenly between physical and cognitive work. The physical half has been mechanizing for 200 years, he writes, and the cognitive half has barely started.

He argues the cognitive half will mechanize much faster because the price of machine thinking is falling faster than the price of mechanical work ever did. Sequoia says "Intelligence per Watt" has been falling more than 10 times a year. Buhler cites the economist William Jevons, who found in 1865 that more efficient steam engines multiplied Britain's coal consumption. Cheaper cognition will follow the same pattern, Buhler argues, and the latent demand for it is "almost unbounded" because "most problems on Earth currently go un-thought-about, because thinking is expensive."

## Buhler Calls Software Engineering the First Skilled Job Machines Now Mostly Do

Buhler compares AI-assisted software development to the Spinning Jenny, the frame James Hargreaves built in 1764 that let one person spin eight threads at once. Writing software is the first cognitive task, he writes, where the machine now completes most of the work and the engineer's job is to direct, review and correct it. "It's no accident," he adds, "that this is where the Cognitive Revolution found its first commercial traction."

Textiles started the Industrial Revolution and the automobile defined it, Buhler writes, and he argues that AI's equivalent of the automobile has probably not been built yet. His candidates are science conducted at machine speed, a personal agent that runs a person's life the way a chief of staff runs a CEO's, and something new in how people connect and coordinate. In 1800, he notes, the automobile would have been hard to predict.

## The Shift May Take Five Years Where the Last One Took Forty

In 1800, Buhler writes, more than 75 percent of American workers worked the land. Today farming is about 1 percent of US employment, more people are employed than at any point in history, and they are far richer than their ancestors. Machines, he writes, "eliminated categories of work and expanded the total amount of it."

Buhler says the same pattern is already showing up in AI: the data show that demand for software engineers, the occupation most exposed to AI, "has inflected," or turned back up, after a period of acceleration. Sequoia plans to publish similar hiring data for legal services and financial services in the coming week.

Buhler cites the economic historian Robert Allen, who named the period from 1780 to 1840 "Engels' pause": British output per worker rose 46 percent while real wages rose only 12 percent. Between 1840 and 1900, output per worker rose 90 percent and real wages rose 123 percent. The aggregate story will be triumphant, Buhler writes, but the individual story frequently will not be, and "the difference this time is speed." The Industrial Revolution gave a farmhand's son 40 years to become a factory hand, he writes, and this revolution may give five. A pause is not inevitable, Buhler writes, and he quotes Mark Zuckerberg: "There is no rule that AI must increase automation faster than it increases individuals' capabilities or demand for new skills."

## Cheap Thinking Would Make One-to-One Tutoring and Drug Research Affordable

On education, Buhler cites Benjamin Bloom's 1984 finding that the average student tutored one-to-one with mastery learning outperformed 98 percent of students in a conventional classroom. The ratio of teachers to students has limited education for all of history, he writes, and the marginal cost of an infinitely patient, endlessly knowledgeable tutor is now approaching zero.

On health, he points to [AlphaFold](https://www.sovereignmagazine.com/article/deepmind-alphagenome-atlas-dna-variant-prediction), which predicted structures for virtually every known protein. Thousands of diseases have no treatment, Buhler writes, because the cost of investigating each one was never justified by the number of patients, and he argues those diseases become worth investigating once that cost collapses. He also cites a shortage of physicians in the United States and worldwide, and argues that cheaper cognitive tools will let more patients receive higher-quality care.

## Wanting, Choosing and Being Trusted Stay Human Longest, Buhler Argues

> "The machine can draft the treaty. Someone still has to sign it."
> — Konstantine Buhler, Sequoia Capital

What stays economically human the longest, Buhler argues, is "wanting things, choosing between them, being accountable for the choice, and being trusted by other people." A machine can produce the work, he writes, but a person still has to decide what is wanted and answer for the choice.

Buhler cites chess as an example: Deep Blue beat Garry Kasparov in 1997, and more people play chess today than at any point in history. After AlphaGo beat Lee Sedol, researchers found that the novelty of moves played by human Go professionals increased significantly. Superhuman machines, he writes, "made human play more creative, not less."

## Sequoia Partner David Cahn Asked Where AI's Revenue Would Come From

Sequoia partner David Cahn made the skeptical case in [AI's $600B Question](https://www.sequoiacap.com/article/ais-600b-question), asking where the revenue would come from to justify the capital being spent on AI infrastructure. In this essay Buhler takes the demand side of that question, arguing that buyers will want far more machine thinking than exists today once it is cheap enough.

Buhler writes that the Cognitive Revolution will be "turbulent, unevenly distributed, and deeply uncomfortable," and that it will end the way the last one did, "with the world unrecognizably better." His closing line is "We have made this trip before." The full essay is [The Cognitive Revolution (When Machines Do the Thinking)](https://www.sequoiacap.com/article/the-cognitive-revolution).

## FAQ

**Q: What is Sequoia's "Cognitive Revolution" thesis?**
Buhler argues that AI is moving human thinking into machines the way the Industrial Revolution moved human muscle into machines. He writes that cognitive work will go from about 99 percent human to about 99.9 percent machine, and that because the cost of machine thinking is falling more than tenfold a year, people will buy far more of it than before.

**Q: Does Sequoia think AI will cause mass unemployment?**
Buhler does not forecast permanent mass unemployment. He writes that past mechanization, including the fall of US farm labor from more than 75 percent of workers in 1800 to about 1 percent today, eliminated categories of work while expanding the total amount of it. His concern is speed: a transition that took 40 years in the Industrial Revolution may now take five, he writes, and individuals can lose out even if the economy as a whole gains.

**Q: What is "Engels' pause" and why does Sequoia mention it?**
"Engels' pause" is the economic historian Robert Allen's name for the period from 1780 to 1840, when British output per worker rose 46 percent but real wages rose only 12 percent. Buhler cites it to show that wages can lag productivity for decades and that people can lose ground for a generation while the overall economy grows.

**Q: Which AI applications does Sequoia think matter most?**
Buhler calls AI coding the first cognitive task where the machine now does most of the work, and argues the defining application of the era has probably not been built yet. His candidates are science conducted at machine speed, a personal agent that runs a person's life the way a chief of staff would, and new tools for how people connect and coordinate. He also names education and health as markets where cheaper thinking makes previously unaffordable services possible.

**Q: Who wrote the Sequoia essay?**
Konstantine Buhler, a partner on Sequoia Capital's seed and early-stage team, wrote "The Cognitive Revolution (When Machines Do the Thinking)," published September 2, 2026.