🔍 Read the full analysis: How AI Could Reshape The Work We Do 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 labor-intensive work to mechanized production to examine how AI could change employment. It reports that about 1.5% of U.S. workers are in agriculture today, compared with an estimated 73.7% in 1800, while emphasizing that the transition unfolded over generations. The author’s forecast that AI could automate as much as 90% of a sector is a thesis, not an established outcome.
Thorsten Meyer published an article November 3 arguing that agriculture’s two-century shift from human labor to mechanized production offers a historical comparison for assessing how AI could reshape work. The piece reports that farming’s share of U.S. employment fell from an estimated 73.7% in 1800 to about 1.5% today, while stressing that this precedent does not establish how quickly or how widely AI will displace workers. The AI productivity debate also examines how AI may affect work.
Meyer’s article, Episode 2 of the series “Work After Work,” follows changes in farming through tools, industrial chemistry and crop breeding. It cites an estimate published by the U.S. Census Bureau that 1.4 million of 1.9 million U.S. workers were in agriculture in 1800. The figures for years before 1840 are estimates, not census counts; the article also notes that the underlying labor-force count includes enslaved people.
The transition did not mean an immediate fall in the number of farm workers. The article says the U.S. farm workforce continued to grow, reaching about 11.8 million around 1910, even as agriculture’s share of the workforce declined. It also cites Germany, where farming’s employment share fell from 43% to 25% between 1882 and 1939 while the number of farm workers remained between 8 million and 10 million.
According to the article, agricultural change came through successive developments, including improved ploughs, mechanized harvesting, fertilizer and higher-yield crops. It cites official Indian statistics showing wheat production increased from 12.3 million tonnes in 1964–65 to 26.4 million tonnes in 1971–72 after new varieties and farming methods spread. Those examples illustrate rising output alongside a changing labor mix; they do not, by themselves, establish what AI will do to employment in other sectors.
What Farm Automation Can Show
Agriculture is a useful historical case because it shows that a sector can produce far more with a much smaller share of workers. The article cites International Labour Organization estimates that 928 million people worked in agriculture worldwide in 2025, about one in four workers. The proportions vary sharply: Meyer gives figures of about 43% in India, 1.5% in the United States and 1% in Germany.
For readers weighing claims about AI and jobs, the comparison points to two separate questions: whether technology can perform tasks people currently do, and what happens to workers and production as firms adopt it. In farming, the employment transition took a long time and coincided with increased food output. The article’s history does not show that displaced workers automatically found equally paid or secure work, nor that every industry will follow the same path.
Meyer’s suggestion that AI could automate up to 90% of a sector within a decade or a few decades is explicitly his thesis, not a measured forecast. The piece offers no sector-specific estimate, model or evidence that establishes that figure. Its practical value is as a historical analogy, not a prediction readers should treat as settled.
Products Worth Considering
Apple MacBook Pro 14.2-inch with M5 Max, 18-core CPU, 32-core GPU, 36GB Memory, 2TB SSD, Wi-Fi 7, Silver
LG UltraGear 34GX90SA-W 34-inch WQHD Curved OLED Gaming Monitor
Ledger Flex + Folio Crypto Hardware Wallet
As an affiliate, we earn on qualifying purchases.
A Long Shift in Farm Work
The article places early U.S. employment estimates alongside historical figures for other countries. It says economic historians estimate that about 85% of Japan’s gainfully occupied population worked in the primary sector in 1872, a category that included farming, forestry and fishing. For China, the national employment series began in 1952 with 83.5% of employed people in the primary sector; India’s 1951 census put cultivators and agricultural laborers at 69.7% of workers.
Those figures are not all directly comparable: they come from different dates, definitions and sources, and some are historical reconstructions rather than census results. The article notes, for example, that Germany’s 1882 occupational census may have undercounted women working on family farms. Its broad point is that labor-intensive agriculture was common before productivity gains reduced the number of people needed to grow food.
The technologies arrived in stages. Meyer cites John Deere’s account of a steel plough made in 1837, the development of industrial ammonia production after Fritz Haber’s 1908 patent, and the spread of short-straw wheat varieties during the 1960s. Together with machinery, irrigation and other changes, these developments helped alter farm output and labor needs over time. The series’ first episode set out a three-part framework of field, factory and office work; this installment focuses on the field.
“The farm is the best test case this season has for its big question, because here a whole sector of human work has already been almost completely automated.”
— Thorsten Meyer, in the article
Limits of the AI Comparison
The article provides historical evidence about agricultural employment and production, but it does not quantify how many jobs AI may affect, identify which occupations are most exposed or estimate the pace of adoption. No evidence in the supplied material establishes that AI-driven change will match agriculture’s timeline or produce similar employment outcomes.
It is also unclear what Meyer means by a sector becoming “up to 90%” automated: the article does not specify whether that refers to tasks, jobs, hours worked or another measure. The figure’s time range spans a decade to several decades and is presented as the author’s thesis. The article’s broader claim that farm output rose as employment shares fell should not be read as evidence that all workers displaced by automation secured new work or shared equally in productivity gains.
The supplied material ends during its discussion of industrial ammonia production. It does not provide the remainder of the article or a detailed account of later sections, including the full evidence behind its global production comparisons. Those details cannot be independently assessed from the material provided here.
Further Episodes and AI Evidence
The article identifies this installment as Episode 2 and says Episode 3 will cover mechanical reapers, threshers and tractors. That follow-up is expected to extend the agricultural timeline, though no publication date is given in the supplied material.
For the AI question, the next evidence readers would need is more specific: which tasks are being automated, how quickly employers adopt the tools, whether jobs are reduced or redesigned, and what happens to wages, hiring and output. The article does not set out a research schedule or cite a forthcoming study that will answer those questions. Its stated next step is the continuation of the series, while the pace and employment effects of AI remain unsettled.
Key Questions
What is the article’s main point?
Meyer uses agriculture’s history to show how technology can reduce a sector’s share of employment while production changes. He presents it as a comparison for thinking about AI, not proof of a particular AI jobs forecast.
How much of the U.S. workforce farmed in 1800?
The article cites a Census Bureau-published estimate of 73.7%, or 1.4 million of a 1.9 million labor force. The pre-1840 figures are estimates rather than census counts, and the estimate includes enslaved workers.
Does the article establish that AI will automate 90% of a sector?
No. Meyer calls that figure his thesis, not a fact. The supplied material does not define the measure or provide a model establishing the forecast.
How many people still work in agriculture worldwide?
The article cites an International Labour Organization estimate of 928 million in 2025, about one in four workers worldwide. National shares differ substantially, and the cited figures use different country contexts.
What does the article say happens next?
Meyer says Episode 3 will cover mechanical reapers, threshers and tractors. The supplied material gives no release date and does not identify a separate next milestone for measuring AI’s effect on employment.
Source: ThorstenMeyerAI.com
