🔍 Read the full analysis: Finding A Productive Balance With AI 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
ThorstenMeyerAI.com’s Nov. 12 article argues that AI’s measurable productivity effect across the economy remains small, even as some studies report substantial gains within particular workplaces. Drawing on Robert Solow’s work and the history of computers, it explains why broad productivity improvements can take years to appear in national statistics.
ThorstenMeyerAI.com published an installment of its “Work After Work” series on Nov. 12, 2026, examining whether artificial intelligence is raising productivity across the economy or mainly improving results in particular workplaces. The article says executives estimate AI has added about 0.2 percentage points a year to productivity growth so far, while one study it cites found a 15% gain for customer-support agents—a contrast that underscores how local results may not yet show up in economy-wide measures.
The article defines labor productivity as the output produced per hour worked. It argues that sustained productivity growth matters because it compounds: at 2% annually, output per hour doubles in about 35 years, compared with roughly 70 years at 1%. Those are illustrative growth calculations, not forecasts of what AI will deliver.
Its historical account draws on economist Robert Solow’s 1957 analysis of U.S. output per worker-hour from 1909 to 1949. Solow attributed 87.5% of the increase to technical change and 12.5% to increased capital use, such as machinery. The article uses the finding to explain why productivity gains can come from changes in technology, methods, organization and skills, not just from buying more equipment.
The article also compares AI with the spread of computers. U.S. nonfarm business productivity growth slowed after 1973 even as computers spread, then rose to about 3.1% annually from 1995 to 2004, based on the article’s calculation from a Bureau of Labor Statistics index. It cites a 2000 estimate by Federal Reserve economists Stephen Oliner and Daniel Sichel that information technology and computer production accounted for about two-thirds of the one-percentage-point productivity-growth acceleration they examined.
Infographic 01 · The sector clock
Productivity moves the sector clock
US farms produced nearly three times their 1948 output in 2023, with only about 9% more inputs in total. Over the same decades, farming's share of US jobs fell from about one in eight to 1.5%.
- Field12.1%
- Factory29.9%
- Services58.0%
Lebergott · own calculation
- Field1.5%
- Factory18.9%
- Services79.6%
ILO estimate
Farm output +210% (1948-2023, 1.52% a year); all inputs +0.12% a year (about +9% in total); total factor productivity +1.40% a year (USDA ERS). 1950: services = remainder, incl. unemployed.
Sources: USDA ERS, Agricultural Productivity in the U.S. (updated 25 Aug 2026); US employment shares: S. Lebergott, in US Census Bureau, Historical Statistics of the United States (1975; 1950, own calculation, see episode 2); ILO modelled estimates via World Bank (2025).
Powered by Thorsten Meyer AI
Why Economy-Wide Gains Take Time
The distinction between workplace-level improvements and aggregate productivity is central to judging AI’s economic impact. A tool can help a defined group complete tasks faster without immediately changing how much the whole economy produces per hour. Gains may be uneven, confined to certain tasks or firms, or offset by the time and resources needed to adopt new systems.
For workers and businesses, the stakes include more than speed. Productivity growth can support higher output and living standards over time, but the benefits depend on how widely improvements spread and whether they translate into measurable economic output. The article’s figures offer a reason for caution: early case-level gains are not, by themselves, evidence that AI has already lifted national productivity growth by a similar amount.
The historical comparison is useful but limited. Computers eventually coincided with faster U.S. productivity growth, but that later acceleration does not prove AI will follow the same course or on the same schedule. The reported 0.2-percentage-point annual estimate is presented as an executive assessment, not as a definitive measurement of AI’s contribution.
Products Worth Considering
Samsung 32-Inch Odyssey G5 QHD 180Hz Gaming Monitor
Amazon Kindle Scribe Colorsoft 64GB with 11-Inch Color Display and Premium Pen, Graphite
Logitech MX Master 4 Ergonomic Wireless Mouse with Haptic Feedback, Graphite
As an affiliate, we earn on qualifying purchases.
From Solow’s Paradox to AI
In a 1987 review, Solow observed that a technological revolution could be widely felt while productivity statistics showed a slowdown. The comment became known as the Solow paradox. The article presents the later productivity acceleration of 1995–2004 as an example of how the effects of a technology wave may become visible in aggregate data only after a lag.
Its broader illustration comes from U.S. agriculture. The article reports that farm output in 2023 was nearly three times its 1948 level, while total inputs had increased by about 9% over that period. It attributes the difference to productivity improvements and links those changes to a long-term shift in where people work. This example shows how higher output can result from changes in production, rather than simply adding more labor and resources.
The source identifies the piece as Episode 10 of “Work After Work,” Arc 2, following an installment about GDP. Its focus here is narrower: output per hour and the challenge of determining whether a new technology is raising that measure.
““Productivity isn’t everything, but in the long run it is almost everything.””
— ThorstenMeyerAI.com, summarizing economist Paul Krugman
What AI’s Productivity Figures Show
The source does not identify the study behind the 15% customer-support gain or give its methods, sample size, time period or comparison group. Readers therefore cannot assess from this article alone how broadly that result applies. The source also does not specify which executives supplied the estimate that AI has contributed about 0.2 percentage points a year, or how they calculated it.
The article does not establish that AI caused any future change in national productivity statistics. It also does not provide enough detail to determine which tasks or industries account for the reported firm-level benefits, whether those improvements persist, or how they compare with adoption costs. The figures are presented as early indicators, not a settled economy-wide accounting.
The Test Is Broader Measurement
The next test is whether reported gains spread beyond individual teams and appear in sustained measures of output per hour. That requires evidence across firms and industries, with clear time periods and comparison baselines, as well as accounting for the resources used to implement AI.
The article does not announce a forthcoming study, release date or policy action. For now, it places AI in a familiar historical debate: visible technological change can precede a measurable change in productivity. Whether AI will produce a broad acceleration, and when, remains an open question.
Key Questions
What does productivity mean in this article?
It means output per hour worked: how much value workers produce in an hour. The article focuses on labor productivity, not total output alone.
How much has AI increased productivity so far?
The article says executives put AI’s effect on productivity growth at about 0.2 percentage points a year. It does not provide the underlying calculation or identify the executives, so this should be treated as an attributed estimate rather than a definitive measurement.
Does a 15% gain in customer support prove AI raises productivity across the economy?
No. The article cites a 15% gain in one study involving support agents, but a result in one setting does not establish an economy-wide effect. The source does not give details of the study’s methods or scope.
Why does the article compare AI with computers?
Computer adoption spread during a period when measured productivity growth slowed, before U.S. productivity growth accelerated in the late 1990s and early 2000s. The comparison illustrates that the economic impact of a technology may take time to appear in broad statistics; it does not prove AI will follow the same path.
Source: ThorstenMeyerAI.com
