🔍 Read the full analysis: AI’s Place In The Next Big Shift In Work on ThorstenMeyerAI.com
Get smart everyday buys 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 report uses agriculture’s long decline in employment share to examine what AI could mean for work. It documents major gains in farm productivity, but its suggestion that automation could reach 90% of a sector is the author’s thesis, not a measured forecast.
A November 3 report by Thorsten Meyer uses agriculture’s long shift from a dominant source of employment to a small share of the U.S. workforce to examine what AI could mean for jobs. The history shows that a sector can produce much more with far fewer workers, but the report’s suggestion that automation could displace up to 90% of work in a sector is Meyer’s thesis, not a confirmed forecast about AI or a specific industry.
Meyer cites estimates that 73.7% of the U.S. labor force worked in agriculture in 1800, compared with about 1.5% today. The early figure comes from economic historian Stanley Lebergott’s estimates, published by the U.S. Census Bureau; it is not a direct census count. The estimate covers a labor force of 1.9 million people aged 10 and older and includes enslaved people. Meyer says the uncertainty in early data makes “about three in four” a more appropriate summary than treating the percentage as exact.
The change took place over generations, through improvements in equipment, fertilizer and crop varieties—not one sudden invention. The report points to a steel plough developed by John Deere in 1837, Fritz Haber’s 1908 ammonia patent and the spread of short-straw wheat in the 1960s. India’s wheat harvest rose from 12.3 million tonnes in 1964–65 to 26.4 million tonnes in 1971–72, according to the figures cited by Meyer. The report attributes the development and spread of dwarf wheat to plant breeders around Norman Borlaug, drawing on the Nobel Foundation.
Employment shares did not fall in lockstep with worker numbers. In the United States, the number of farm workers continued rising until about 1910, when it reached an estimated 11.8 million. Meyer also cites estimates that the world’s cereal harvest grew 4.2-fold from 1961 while the population grew 2.7-fold. These figures illustrate changes in production and employment; they do not establish how AI will affect other occupations.
Farm Productivity and the AI Debate
Agriculture provides a real example of output rising as a sector’s share of employment fell. That matters as businesses and workers consider whether AI could automate tasks at scale: the farming record cautions against assuming that fewer workers in a field necessarily mean less production. It also shows that productivity gains can accumulate through several technologies and take a long time to reshape employment.
But the analogy has limits. The report describes machinery, plant science and fertilizer changing farm production over roughly two centuries in the United States. It does not establish how quickly AI will spread, which tasks it will replace or how many new jobs may arise. For workers, employers and policymakers, those distinctions matter: historical productivity gains are evidence of what happened in farming, not a guarantee of similar outcomes elsewhere.
Popular right now
Vitamix 5200 Professional-Grade Blender, 64 oz. Container, Black
Dyson Car+Boat Cordless Handheld Vacuum
Ninja Flip Toaster Oven & Air Fryer, Large Capacity, Stainless, SP151
As an affiliate, we earn on qualifying purchases.
How Farming’s Workforce Changed
The report compares the United States with several other countries to show that agriculture once employed a large share of workers in different economies. It cites historical estimates of roughly 85% in Japan in 1872 and 83.5% in China in 1952, both for the broader primary sector, which includes activities such as forestry and fishing. In India, cultivators and agricultural laborers accounted for 69.7% of workers in the 1951 census, Meyer says. The figures come from different dates, definitions and source methods, so they are not a directly comparable survey.
In the United States, the agricultural employment share had fallen substantially by the modern era, while farm output expanded. Meyer cites U.S. Department of Agriculture data showing 2023 farm output at nearly three times its 1948 level, with only a small rise in total inputs. The report also notes that agriculture remains a major employer globally: it cites an International Labour Organization estimate of 928 million people working in agriculture in 2025, about one in four workers worldwide. The shift is uneven, not complete.
“The share fell long before the headcount.”
— Thorsten Meyer, in the November 3 report
AI’s Pace and Reach Remain Unknown
The report does not identify an AI-driven employment decline already underway or provide a forecast tied to a named industry, country or set of occupations. Its 90% figure is Meyer’s hypothesis; the report does not supply a model, timeframe definition or evidence that would make it a confirmed prediction.
It is also unclear how closely AI will follow agriculture’s path. Farming’s history involved changes to physical equipment, biological inputs and land-based production. AI may affect a different mix of tasks and occupations. The report does not quantify the likely balance between jobs lost, tasks altered and new jobs created, or explain how benefits and costs might be distributed among workers and businesses.
Further Episodes Will Extend the Comparison
Meyer describes this report as Episode 2 of “Work After Work,” a series examining how the economy is built. He says Episode 3 will cover mechanical reapers, threshers and tractors, extending the agricultural timeline. That material may add detail on how machinery changed farm work, but it will not by itself resolve how AI adoption will affect employment.
For readers following the broader question, the next evidence to watch is industry-specific: which tasks organizations automate, how quickly they adopt AI, and what happens to employment and output afterward. Until those outcomes are documented, farming is best treated as a historical comparison rather than a direct prediction of AI’s effects.
Key Questions
What is the report’s main point about AI and work?
Thorsten Meyer uses agriculture to show that a sector can produce more while employing a much smaller share of workers. He presents that history as a comparison for AI, not proof of what will happen to other jobs.
Does the report predict that AI will eliminate 90% of jobs?
No. Meyer says automation could reach up to 90% of a sector within a decade or a few decades, but identifies this as his thesis. The report does not establish that 90% of jobs will disappear or provide a substantiated forecast for a particular sector.
How long did agriculture’s employment shift take in the United States?
The report describes a change that unfolded over about two centuries. The share of U.S. workers in agriculture declined over time; the number of farm workers, however, continued to rise until about 1910.
How many people still work in agriculture worldwide?
Meyer cites an International Labour Organization estimate of 928 million people in 2025, or about one in four workers worldwide. The figure refers to global agricultural employment, not the share of workers in any single country.
What evidence would clarify AI’s effect on jobs?
Useful evidence would include documented changes in specific tasks, employment and output after organizations adopt AI, along with the timeframe and industry involved. The report does not provide those measurements for AI.
Source: ThorstenMeyerAI.com
