Same Job, Fewer Workers: Why Technology May Bring Mass Unemployment
In 1920, one roof took a crew a week. Today, tools and AI let one worker do more, from roofing to coding — a shift that may bring mass unemployment across every industry.
Abstract
In 1920, one roof took a crew a week. Today, tools and AI let one worker do more, from roofing to coding — a shift that may bring mass unemployment across every industry.
Post
Picture two roofing crews, living a hundred years apart. In 1920, a roofing crew climbs ladder after ladder, carrying shingles on their shoulders. Each shingle is nailed down by hand. The job takes the better part of a week for a single house. A modern 2020 crew uses a lift truck to move materials to the roof in minutes, and pneumatic nail guns drive a nail every second. That same house might be finished in part of a single day. This gap isn't merely a fun fact about roofing. It's a warning sign about where the whole economy is headed: mass unemployment.
The roofing example shows how technology multiplies what one worker can do. A century ago, every part of the job was manual. Shingles were cut by hand, hauled up narrow ladders, and nailed down one at a time. It took a large crew and many days of hard labor to finish even one average roof. Today, power tools handle most of that work. A modern crew, often smaller than a 1920s crew, can finish several roofs in the time it once took to finish just one. Fewer workers are now needed to produce the same amount of work, or even more.
This same story is playing out in every industry, not just roofing. Farms that once needed several dozen workers now run with a few people and heavy machinery. Factories that once used hundreds of hands on an assembly line now need only robotic arms and a handful of technicians. Office jobs are shrinking too, as software and artificial intelligence take over tasks such as scheduling, data entry, and basic customer support. Each of these changes follows the same math as roofing: technology lets fewer people do more work, so the economy needs fewer total jobs.
Some people push back on this, noting that older technologies, like tractors or assembly lines, eventually created new jobs to replace the ones they destroyed. That's true, but those machines mostly replaced physical labor, which left room for workers to shift into thinking jobs instead. Today's technology is starting to replace thinking work too. When software can plan projects, write reports, or answer customer questions, there are fewer new jobs left for displaced workers to move into. That's why many experts worry this wave of automation could shrink the job market faster than the economy can adjust.
Software programming shows this shift into thinking work clearly. Seven years ago, building an app or website took a team of programmers who wrote nearly every line of code by hand and spent hours tracking down mistakes. Today, AI tools can write code, catch errors, and suggest entire sections of a program in seconds. A single developer working with these tools can now do work that used to take a small team. Because of this, many tech companies need fewer entry-level coders than they did just a few years ago, even as the total amount of software being built continues to grow. This matters because programming was long seen as one of the safest, fastest-growing careers around. If a field like that already needs fewer workers, it suggests that even highly skilled, well-paid jobs are not safe from this trend.
The story of roofing is a small example of a much bigger shift. What once took a large crew a full week now takes a small crew a single day, and that same pattern is showing up in farming, manufacturing, office work, and even computer programming. As machines and software learn to do more with fewer people, the total number of jobs may continue to shrink for years to come. This doesn't mean the future has to be bleak, but it does mean society needs to take the trend seriously and start preparing now, because waiting will not slow it down.
Even the solutions that helped workers through past waves of automation may not be enough this time. Retraining programs assume a displaced worker can move into a new, growing field, but AI is now competing for many of those same jobs too. Unemployment programs were designed to support people between jobs, not those whose old jobs may never return. And the old belief that the economy always finds a way to absorb new technology may not hold up against a technology that can do both physical and cognitive work at the same time. If this truly is a different kind of change, then the policies and thinking that worked before may not be equipped to handle what comes next in society.
References:
- The Strange and Awful Path of Productivity in the U.S. Construction Sector
- https://www.nber.org/papers/w30845
- Robots and Jobs: Evidence from US Labor Markets.
- https://www.nber.org/papers/w23285
- AI Will Reshape More Jobs Than It Replaces
- https://www.bcg.com/publications/2026/ai-will-reshape-more-jobs-than-it-replaces
- AI Ate the Bottom of the Software Career Ladder, and the Stanford Data Shows Who Gets Left Behind
- https://techfastforward.com/articles/stanford-2026-junior-developers-down-20-percent-ai-erased-career-ladder
- Computer Programmers
- https://www.bls.gov/ooh/computer-and-information-technology/computer-programmers.htm
Buy me a coffee at:
https://buymeacoffee.com/clubtj
Visit my blog at:
https://www.quarkstochlorophyll.blog
© 2025 Tim Jackson. All Rights Reserved.
