🔍 Read the full analysis: The 15-Hour Week Was The Forecast. AI Had Other Plans 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 18 article from ThorstenMeyerAI.com revisits John Maynard Keynes’s 1930 prediction that rising prosperity could eventually make a 15-hour workweek possible. It says income per person has risen broadly within his forecast range, but the U.S. manufacturing workweek has remained near 40 hours for decades; early AI evidence shows time savings, not a broad reduction in recorded work hours.
A November 18 analysis from ThorstenMeyerAI.com revisits economist John Maynard Keynes’s 1930 vision of a possible 15-hour workweek, finding that rising incomes came closer to his forecast than shorter working hours did. The article says U.S. manufacturing workers averaged 41.8 hours a week in October 2026, while early estimates suggest generative AI has saved some work time without yet producing a clear reduction in recorded hours.
In “The 15-Hour Week Was the Forecast. AI Had Other Plans,” author Thorsten Meyer distinguishes between Keynes’s numerical forecast for living standards and his suggestion that remaining work might be shared through three-hour shifts or a 15-hour week. The essay, published in 1930 as “Economic Possibilities for our Grandchildren,” predicted that living standards in “progressive countries” could become four to eight times higher over a century. Meyer’s calculations using the Maddison Project Database put real income per person in 2022 at about 4.4 times its 1930 level in Britain and 5.5 times in the United States.
The workweek data tell a different story. The article cites U.S. manufacturing production-worker averages of 51 hours in 1909, 44.2 in 1929 and 41.8 in October 2026. It says the Fair Labor Standards Act established a 40-hour threshold for overtime pay beginning in October 1940, after a staged reduction from 44 hours. The law did not prohibit longer weeks, but required overtime pay above the threshold.
On artificial intelligence, the article reports an estimate that U.S. workers saved about 2.1% of all work hours with generative AI in August 2026. It also cites a Danish study that found no measurable effect of AI chatbots on recorded hours. These findings concern time saved or hours recorded; they do not establish that AI has reduced the standard workweek across the economy.
AI Time Savings Have Not Cut the Week
The comparison matters because productivity gains do not automatically translate into more leisure. Keynes’s vision connected greater prosperity with the possibility of sharing necessary work more widely. The figures presented in Meyer’s analysis suggest that income growth and shorter working hours have followed different paths: living standards rose substantially, while the U.S. manufacturing week settled around a 40-hour benchmark.
For workers and employers, the early AI evidence leaves open how time saved through automation will be used. It could support higher output, different tasks, reduced hours or other changes; the cited estimate alone does not show which outcome is occurring. The Danish study’s reported lack of a measurable effect on recorded hours also cautions against treating AI use as proof that working time has already fallen.
The broader point is not that AI cannot shorten work, but that technical capacity does not determine how productivity gains are distributed. The article’s historical comparison gives readers a reason to distinguish tools that save time from workplace or policy changes that convert those savings into shorter schedules.
Popular right now

Bose QuietComfort Wireless Noise-Cancelling Headphones, Moonlight Grey

Roborock Q7 M5+ Robot Vacuum and Mop with Self-Emptying Dock

Bose QuietComfort Ultra Wireless Noise Cancelling Earbuds, 2nd Gen, with 20W 2-Port Wall Charger
As an affiliate, we earn on qualifying purchases.
From Factory Hours to Keynes’s Forecast
Keynes wrote “Economic Possibilities for our Grandchildren” during the economic turmoil preceding the Great Depression. He argued that technology and accumulated wealth could ease what he called the struggle for subsistence. His remarks about a 15-hour week came as a possible way to share the remaining work, not as a dated prediction that the week would reach 15 hours by a specified year. The distinction matters: the article treats the income projection as a forecast and the shorter week as a proposed possibility.
The cited U.S. factory-hours series shows a substantial fall in the early 20th century, from 51 hours a week in 1909 to 44.2 in 1929. The Fair Labor Standards Act, signed in 1938, phased in the 40-hour overtime threshold by October 1940. Meyer says average manufacturing hours have since stayed near that level, citing 38.1 hours in 1940, 40.5 in 1950 and 39.8 in 1970, alongside the preliminary October 2026 figure.
The article also notes a different European path: following a seven-week strike in 1984, Germany’s metalworking settlement reduced the week to 38.5 hours, with the West German metal industry reaching 35 hours by 1995. These examples provide context for how working time can change through labor agreements and law, not technology alone.
““Five working days of eight hours, with no overtimes, is to constitute a week’s work.””
— Ford company paper, quoted in Meyer’s article on the 1926 schedule
What the AI Figures Cannot Show
The available figures do not establish whether generative AI will lead to shorter working weeks over time. The 2.1% estimate for August 2026 is described as a share of all U.S. work hours saved, but the source material does not specify its method or a comparison baseline. The Danish study is summarized as finding no measurable effect on recorded hours; the supplied material does not provide the study’s sample, timeframe or detailed results.
The October 2026 U.S. manufacturing-hours figure is identified as preliminary and applies to production and nonsupervisory employees in manufacturing, not all workers. The income comparisons are Meyer’s calculations from the Maddison Project Database and end in 2022, rather than covering the full hundred years of Keynes’s forecast. The supplied source excerpt also ends before completing its discussion of the growth comparison, so its full supporting detail is not available here.
The Next Test Is Working Time
The immediate next step is to see whether later research and labor-market data show AI-related time savings translating into fewer recorded hours, shorter standard schedules or a different distribution of work. That requires evidence over time and across occupations, rather than a single estimate of hours saved.
Readers can also track revisions to the preliminary October 2026 manufacturing figure and compare it with hours data for workers beyond manufacturing. For now, Meyer’s analysis documents a historical contrast: substantial gains in income per person, a U.S. factory workweek still near 40 hours, and early AI findings that do not yet establish a broad reduction in working time.
Key Questions
Did Keynes predict that the workweek would be 15 hours by 2030?
No. In the account presented by Meyer, Keynes gave a dated numerical forecast for living standards but described a 15-hour week as one possible way to share remaining work. It was not a specific deadline for reaching that schedule.
How close did living standards come to Keynes’s forecast?
Meyer’s calculations using Maddison Project Database figures say real income per person in 2022 was about 4.4 times the 1930 level in Britain and 5.5 times in the United States. The source describes Keynes’s forecast as a four-to-eightfold increase in “progressive countries.”
What is the latest U.S. workweek figure cited?
The article cites a preliminary average of 41.8 hours a week in October 2026 for production and nonsupervisory employees in U.S. manufacturing. It is not a figure for every U.S. worker.
Has generative AI shortened the working week?
The cited evidence does not establish that. Meyer reports an estimate that generative AI saved about 2.1% of all U.S. work hours in August 2026, while a Danish study found no measurable effect of AI chatbots on recorded hours. Time saved is not the same as a demonstrated reduction in scheduled or recorded working hours.
Source: ThorstenMeyerAI.com
