🔍 Read the full analysis: Could AI Transform Work As Much As Farming Did? on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A November 3 article uses agriculture’s long shift from human labor to machinery, improved seeds and fertilizer to examine whether AI could reshape work. The historical account shows that farming’s share of employment fell sharply while food production grew, but the author’s suggestion that AI could automate up to 90% of a sector is a thesis, not an established forecast.
Thorsten Meyer’s November 3 article argues that the history of farming offers a useful test for claims that AI could transform work: agriculture’s share of employment has fallen dramatically over time, even as food output rose. The comparison is historical, not a forecast; Meyer’s suggestion that AI could automate up to 90% of a sector within a decade or a few decades is his thesis, not a confirmed projection.
Meyer cites an estimate by economic historian Stanley Lebergott, published by the U.S. Census Bureau, that about 73.7% of the U.S. labor force worked in agriculture in 1800. The figure is based on estimates rather than a census count, and includes enslaved people. The article notes that the comparable share is now about 1.5% in the United States, while agriculture still employs a much larger share of workers in countries such as India.
The shift did not come from a single invention. Meyer describes successive changes in farm tools, chemical fertilizers and crop varieties. A steel plough developed by John Deere in 1837 helped address sticky Midwestern soil; later came mechanical equipment. The article also points to the Haber-Bosch process for making ammonia fertilizer and the spread of short-straw wheat during the Green Revolution.
These changes coincided with much greater agricultural production. India’s wheat harvest rose from 12.3 million tonnes in 1964-65 to 26.4 million tonnes in 1971-72, according to figures cited in the article. Globally, Meyer reports, cereal production increased 4.2-fold between 1961 and 2023, while population grew 2.7-fold. These figures describe changes in output and population, not a direct measure of how much any one technology caused.
The Employment Lesson from Farming
The historical example challenges the idea that fewer workers in a sector necessarily mean less of its output. Farming now uses a much smaller share of the U.S. workforce than it did in 1800, yet food production is far higher. That matters to current debates about AI because automation may change the number and kind of jobs needed to produce goods and services, without necessarily shrinking production.
But the comparison does not establish what will happen to office, service or professional work. Agriculture changed over roughly two centuries in the United States, through many technologies and social changes. AI adoption could follow a different pace and affect tasks across many sectors at once. The article’s history is evidence that large employment shifts can occur; it is not proof that AI will produce the same outcome.
The distribution of gains also matters. Rising output does not by itself show whether former farm workers found comparable work, whether wages improved, or how the benefits were shared. Those questions are central to assessing AI’s effects on workers and communities.
Products Worth Considering
Logitech MX Keypad Programmable Macro Pad, Graphite
Apple MacBook Air 15-inch (2026) with M5 chip, 16GB memory, 512GB SSD, Silver
LG UltraGear 45GX900A-B 45-Inch Curved OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
From Farm Labor to Higher Output
Meyer sets the farming example within a series about how economies are organized into agriculture, industry and services. Around 1800, most U.S. workers were tied to food production because farms relied heavily on human and animal labor. The article gives estimates from other countries to show that high agricultural employment was not unique to the United States: Japan’s primary-sector share was estimated at about 85% in 1872, while India’s 1951 census counted cultivators and agricultural laborers as about 70% of workers.
The historical figures are not all directly comparable. Some cover agriculture alone, while others include forestry and fishing; some are reconstructions rather than official counts. Meyer also cautions that early U.S. estimates are uncertain. The article says the U.S. farm-worker headcount continued to rise until about 1910, even as agriculture’s share of the workforce eventually declined. In Germany, it cites a fall in agriculture’s employment share from 43% in 1882 to 25% in 1939, while the number of farm workers remained between 8 million and 10 million.
The comparison with present-day work also has a key limitation: agriculture is a specific sector, shaped by biological processes, land and food demand. AI systems can be applied to a wide range of tasks, but their effects on employment depend on how organizations use them, what tasks they can perform reliably and whether new work is created.
“Farming shows a whole sector can be emptied of human hands. It also shows that it took about 200 years in America.”
— Thorsten Meyer, article author
AI’s Pace and Job Effects
The article does not establish how many jobs AI will replace, alter or create, nor which workers or sectors will see the largest effects. Its “up to 90%” formulation is not backed in the supplied material by a defined measure of automation, a specified sector or a cited forecast. It should not be read as a verified estimate that 90% of jobs will disappear.
It is also unclear how quickly AI adoption will spread or whether productivity gains will lead to new roles, reduced working hours, lower costs, higher profits or some combination. The farm comparison cannot settle those questions. Its employment and production figures describe different periods and countries, and the causal contribution of particular technologies is not isolated by the statistics presented.
The agriculture figures also carry limits: several early values are estimates, country definitions differ, and the source material does not provide the full underlying datasets for every comparison. Those caveats do not erase the broad historical shift, but they limit how precisely the numbers can be compared.
Further Episodes and Better Measures
Meyer identifies this as Episode 2 in the “Work After Work” series, following an earlier installment on economic sectors. The next episode is expected to address farm machinery, according to the article’s note that “Episode 3 tells that part.” No date for publication is given in the supplied material.
For readers tracking AI’s effect on work, the next useful evidence will come from measured changes in employment, hours, wages, output and task composition, rather than predictions alone. Results will need clear time periods and comparison baselines, and should distinguish jobs eliminated from tasks changed or new positions created. The farming history offers a comparison point, while the scale and timing of AI’s effects remain open.
Key Questions
What is the article’s main claim about AI and farming?
Thorsten Meyer argues that farming’s long shift toward automation can help readers think about AI’s potential effects on work. He does not establish that AI will follow the same path or on the same timeline.
Did 73.7% of Americans work in agriculture in 1800?
That is Lebergott’s estimate for the U.S. labor force in 1800, cited by Meyer and published by the Census Bureau. The article says the early figures are estimates, not census counts, and that the labor-force total includes enslaved people.
Does the article say AI will automate 90% of jobs?
No. Meyer describes automation reaching up to 90% of a sector within a decade or a few decades as his thesis, not a fact. The source does not define a sector or provide a verified forecast supporting that number.
Does fewer farm labor mean less food production?
Not in the historical comparison presented. Meyer reports that global cereal production grew faster than population between 1961 and 2023, while agriculture’s employment share fell in many countries. Those figures show the trends together but do not isolate the effect of any single technology.
What remains unknown about AI’s effect on work?
The scale, timing and distribution of AI’s effects remain uncertain. The article does not quantify likely job losses or gains, identify which occupations will be most affected, or show how productivity gains will be shared.
Source: ThorstenMeyerAI.com
