🔍 Read the full analysis: Why 80% Of AI Careers Are Services on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A November 9, 2026, essay from ThorstenMeyerAI.com argues that services became a dominant share of U.S. employment well before factory jobs peaked, and uses Baumol’s cost disease to explain why many service costs rise. It cites a study in which an AI assistant increased customer-support agents’ resolved issues per hour by 15% on average, while saying the effect on service-sector employment remains an open question.
The piece, part of the series “Work After Work,” argues that the shift toward services did not begin only after manufacturing employment declined. Its calculation from U.S. Bureau of Labor Statistics payroll data puts services at 62% of nonfarm payroll jobs in 1939, 72% in 1979 and 86% in 2025. The BLS series excludes farm workers and self-employed people, while the separate ILO estimate covers all workers; the two percentages use different populations and should not be treated as directly interchangeable.
To explain why some services become more expensive, the article revisits the work of Baumol and William Bowen. Their 1966 study of performing arts compared a live string quartet, which still requires four musicians to play for about 45 minutes, with manufacturing, where machinery and process changes can reduce labor per unit. Baumol’s 1967 paper developed the argument into a broader account of uneven productivity growth: wages can rise across the economy even where output per worker is difficult to increase, putting upward pressure on the cost of those activities.
The essay compares U.S. prices since 2000, reporting a 98% decline in the official quality-adjusted price of televisions alongside increases of 274% for hospital services, 188% for college tuition and 143% for car repair. These figures are presented as price changes over the stated period, not as proof that one factor alone caused them. The article also cites a study of 5,172 customer-support agents in which an AI assistant raised the number of issues resolved per hour by 15% on average.
Infographic 02 · Baumol's cost disease
Why the workers moved to services
Baumol's mechanism in four steps, and what it looks like in US jobs.
- 1966Baumol and Bowen: the quartetA 45-minute Schubert quartet still takes about three hours of human work
- 1967Two kinds of workWhere labour is a tool, machines cut it again and again; where labour is the product, they can’t
- WagesPay rises everywhereTeachers, nurses and musicians must earn more too, or they leave
- ResultCosts rise, people moveIf we keep buying the same mix, more and more workers end up in the slow sector
Service industries, share of US nonfarm payroll jobs
US jobs, annual average 2025 (millions)
Payroll shares exclude farm workers and the self-employed (BLS establishment survey). Health spending: national health expenditure as a share of gross domestic product.
Sources: W. J. Baumol, “Macroeconomics of Unbalanced Growth”, American Economic Review 57(3), 1967, BLS Current Employment Statistics (own calculation of shares), CMS, National Health Expenditure, Table 1.
Powered by Thorsten Meyer AI
AI Meets Labor-Intensive Services
The question is not just whether AI can perform particular tasks. It is whether it can raise productivity in services where human attention is part of what customers receive, such as teaching, medical care and live performance. If tools let workers handle more work without reducing service quality, they could ease some cost pressures. The cited customer-support result offers one measured example, but it does not show that similar gains will occur across the wider service economy.
Employment effects are a separate issue from productivity. Higher output per worker could reduce the number of workers needed for a given volume of work, support greater demand, or change the tasks people perform. The source frames the outcome as unresolved and provides no economy-wide estimate of jobs lost or created by AI. Readers should distinguish the study’s reported productivity result from broader predictions about the future of service work.
The price comparisons also matter because major service expenses affect households and public budgets. But the article’s own contrast between U.S. and German health spending—18.0% versus 12.4% of the economy—points to factors beyond the basic productivity argument. The figures do not by themselves explain the difference between the countries or establish what policy changes would narrow it.
Products Worth Considering
Bose QuietComfort Wireless Noise-Cancelling Headphones, Twilight Blue
Bose QuietComfort Headphones (2nd Gen), Wireless Over-Ear Noise-Cancelling Headphones, Eucalyptus Green
Apple MacBook Air 13-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, 12MP Center Stage Camera, and Wi-Fi 7, Sky Blue
As an affiliate, we earn on qualifying purchases.
From Quartet to Service Economy
Baumol and Bowen studied the finances of theatre, opera, music and dance for the Twentieth Century Fund. In their 1966 book, Performing Arts: The Economic Dilemma, they described how a performance’s labor requirements resist the same kinds of reductions seen in factory production. Baumol later generalized the idea in his 1967 American Economic Review paper, “Macroeconomics of Unbalanced Growth: The Anatomy of Urban Crisis.”
The article links that theory to the long change in U.S. employment. It notes that manufacturing jobs reached a peak of almost 20 million in June 1979, but says the services share had already grown from its 1939 level before that peak and continued rising afterward. This challenges a simple account in which services expanded only because factories contracted: the figures indicate a longer-running shift, though they do not isolate every cause.
Baumol’s distinction is not absolute. The essay notes that technology has made some service work more productive, including through self-service retail and faster travel for lecturers. Its central contrast is between activities where labor can be reduced while producing the same output and those where the amount of human involvement is closely tied to the service being delivered.
““Human ingenuity has devised ways to reduce the labor necessary to produce an automobile, but no one has yet succeeded in decreasing the human effort expended at a live performance of a 45-minute Schubert quartet much below a total of three man-hours.””
— William Baumol, quoted from his 1967 paper
What AI’s Productivity Gains Mean
The source does not identify the customer-support study’s publication, date, research design or comparison group, so those details cannot be independently evaluated from the supplied material. Its 15% average gain applies to the studied agents and work setting; it is not evidence that AI produces the same result in hospitals, schools, repair shops or other services.
It is also unclear whether the reported productivity increase improved service quality, reduced costs for customers or changed staffing. The essay does not provide evidence that AI has already reversed the long-term growth of service employment. Its employment question—whether workers are displaced or move into other roles—remains open.
The price figures require care as well. The article gives a period and comparison for each listed price change, but the supplied material does not detail the underlying index methods or explain how much of each change reflects productivity, wages, policy, demand or other factors. Its claim that service prices are affected by Baumol’s mechanism is an economic interpretation, not a complete causal breakdown of each price series.
Evidence Needed Beyond Support Work
The immediate next step is to test whether the customer-support result holds across other organizations and types of service work. Useful evidence would identify the study’s methods, measure service quality as well as issues handled per hour, and track what happened to workers’ responsibilities and staffing after the tool was introduced.
For the broader labor-market question, readers will need employment and productivity data over time, alongside clear definitions of which workers and industries are counted. The source provides historical employment shares through 2025 and a single AI productivity finding; it does not report a future milestone or forecast for service-sector jobs.
Until broader evidence emerges, the article’s main point is a framework rather than a settled prediction: services grew for decades before the 1979 manufacturing peak, and some service activities are harder to make more productive than factory production. Whether AI changes that pattern at scale remains to be seen.
Key Questions
What is the news development in this article?
ThorstenMeyerAI.com published an analysis on November 9, 2026, connecting the growth of service employment with Baumol’s cost disease and considering whether AI can raise productivity in service work.
Does the source say 80% of U.S. jobs are services?
It cites an ILO estimate of 79.6% of all U.S. workers in services in 2025. Separately, its calculation from BLS nonfarm payroll data puts the share at 86%; that series excludes farm workers and self-employed people, so the figures use different definitions.
What is Baumol’s cost disease?
It describes how costs can rise in work where productivity is hard to increase, even as wages rise alongside more productive parts of the economy. The source uses live performance as an example because a quartet still needs musicians to play the piece.
Does the cited study prove AI will eliminate service jobs?
No. The article reports that an AI assistant increased issues resolved per hour by 15% on average in a study of 5,172 support agents. It does not establish job losses or predict effects across the service sector.
Source: ThorstenMeyerAI.com
