Will AI take my job? What the sewing machine already taught us
The tailors who smashed the sewing machines
In 1830, a French tailor named Barthélemy Thimonnier patented a sewing machine and set up a Paris workshop with about 80 of them, making army uniforms. In 1831, a crowd of roughly 150 to 200 tailors stormed the workshop and destroyed the machines. They believed the machine would end their trade.
It didn't. Sewing machines spread anyway, and what followed was the opposite of what the tailors feared. Clothes that once took days could be made in hours, so they became cheaper, and far more people bought them. Ready-made clothing turned into a giant industry.
The numbers from one city
In Philadelphia, after the foot-powered sewing machine arrived in 1846, clothing work exploded:
| Year | Clothing workers | Share of all jobs |
|---|---|---|
| 1850 | 10,532 | 18.2% |
| 1880 | 34,548 | 19.8% |
More than three times as many people worked in clothing 30 years later. The machine didn't remove the work. It removed the slowest part of it, and demand did the rest.
The honest part: the work changed. Many skilled hand-tailors lost status, and factory jobs were often lower-paid. Technology rarely leaves jobs exactly as they were. That's the real lesson, and it applies to AI too.
The same pattern, again and again
Power looms: 98% of the labour automated, more weavers hired
Economist James Bessen found that 19th-century power looms automated about 98% of the labour needed to weave a yard of cloth. Yet factory weaving jobs increased, because cheaper cloth created much bigger demand.
ATMs: the machine that was meant to replace bank tellers
The US now has over 400,000 ATMs. Between 1988 and 2004, tellers per branch fell from about 20 to 13, but cheaper branches meant banks opened more of them (urban branches rose 43%). Teller jobs didn't disappear. They shifted from counting cash to selling and helping customers.
The pattern is consistent: automation shrinks the task, lowers the cost, grows the demand, and moves people to the higher-value part of the work.
Canva didn't kill design. It created designers.
When Canva arrived, many designers worried. If anyone can make a poster in five minutes, who pays a designer? Today Canva reports 265 million monthly active users and over 31 million paying subscribers. Hundreds of millions of people now design something every month, people who never would have opened professional design software.
Simple design became something everyone can do. That pushed professional designers toward what templates can't do: brand strategy, original ideas and work that stands out from the millions of templates everyone else is using.
We help small businesses automate recurring work like attendance, payroll, orders and WhatsApp replies, so your team spends time on customers, not copy-paste.
See business automationWhat's actually different about AI
AI is the next step in a long line of technology updates. What makes it feel different is how it learns. Earlier machines followed fixed instructions. AI learns patterns from huge amounts of human experience: text, code, designs, records.
That's also its limit. AI is built from what millions of people have already done. It is excellent at repeating, combining and speeding up known work. It does not decide what's worth creating, it doesn't know your customers, and it has no goal of its own. Direction still comes from people.
For a business, this means your own records are the real asset. Companies that track daily sales, attendance, orders and customer questions can see patterns and automate the repeated steps. Companies with no data have nothing for AI to work with.
The knowledge paradox: AI helps experts most
It sounds backwards, but the evidence suggests AI rewards people who already know their field.
Two studies, two very different results
- A clear, well-defined task: in a GitHub experiment with 95 developers, the group using Copilot finished a set coding task 55% faster.
- Real, complex work: in a 2025 randomised trial by METR, 16 experienced open-source developers took 19% longer on real issues when allowed to use AI tools. Afterwards, they still believed AI had made them 20% faster.
The lesson isn't "AI is useless." It's that AI speeds up work you already understand and can check. If you can't judge the output, it can feel productive while quietly making things worse.
Like the sewing machine, AI makes a skilled person faster and lets them try more designs. For someone with little knowledge, it's risky: whatever it shows looks like the whole answer. AI can write thousands of lines of code in minutes, but someone still has to know whether that code is correct, secure and maintainable. Testing and maintenance are where weak knowledge gets exposed.
The rule: AI is powerful when you hold the steering wheel, with a clear goal, real knowledge and your own ideas. It's dangerous when you let it drive.
Why change feels sudden but isn't
New technologies take time to change how work is done. Electricity was well known in the late 1800s, but US factories didn't fully use it until the 1920s. Economists estimate it took around 40 years for electricity's productivity gains to show up, because factories had to be redesigned around it.
AI will follow a similar path. The headlines move fast, but businesses change at the speed of their processes, data and people. The World Economic Forum's Future of Jobs Report 2025 puts the expected shift by 2030 in numbers:
| By 2030 (WEF estimate) | Figure |
|---|---|
| New jobs created | 170 million |
| Jobs displaced | 92 million |
| Net change | +78 million |
| Share of job skills expected to change | 40% |
| Workers who will need reskilling | 59 in 100 |
| Employers planning to upskill staff | 77% |
Displacement is real, so pretending AI changes nothing would be wrong. But the bigger number is the opportunity, and it goes to people who adapt.
What to do now: 7 steps
- Go deeper in your craft, not shallower. Knowledge is what lets you direct and check AI.
- Set the goal before you open the tool. Know what you want to produce and for whom.
- Hand AI the repetitive work. First drafts, data entry, summaries, routine replies, boilerplate code.
- Be original. AI recycles what millions already made. Your ideas, taste and experience are the part it can't copy.
- Check everything that matters. Test the code, verify the facts, read before you send.
- Start tracking your data. Daily records of sales, time, orders and questions are the fuel for useful automation.
- Learn from someone who's seen the industry change. A good guide saves you from the hype and the mistakes.
For business owners: start by listing the tasks your team repeats every day. Those are the first things to automate, and the people doing them can move to work that grows the business.
Coming next: a deep dive into AI
This post is the big picture. If you want to go further, our next post is a deep dive into AI, written from my own experience of using it in real client work.
- My personal experience: where AI saved me hours, where it got things wrong, and how I caught it.
- How to use AI professionally: giving it clear direction, checking its work, and keeping control of quality.
- Real case studies: AI and automation in actual businesses, from websites to payroll systems.
Tell us what you'd like it to cover, or ask Lenso, our assistant, in the corner of this page. Your questions will shape the next post.
Tell us what to coverFrequently asked questions
Will AI take my job?
AI is more likely to take over tasks than whole jobs. The World Economic Forum expects 170 million jobs created and 92 million displaced by 2030. People who learn to direct AI are best placed to benefit.
Which jobs are safest from AI?
Work that needs judgement, original ideas, trust and hands-on skill: building relationships, leading projects, solving new problems and checking AI's output. Purely repetitive tasks are the most exposed.
Is AI like the sewing machine or the Industrial Revolution?
In one key way, yes. Like the sewing machine and the power loom, AI makes work faster and cheaper, which tends to grow demand and shift people to higher-value work. The difference is that AI learns from data, so it can handle more kinds of tasks.
Does AI make beginners as good as experts?
Not reliably. Studies suggest AI helps most when the user understands the work and can check the result. Without that knowledge, AI output can look right while being wrong.
How should a small business start with AI?
List the tasks your team repeats every day, start recording the data behind them, then automate one process at a time, such as attendance, order updates or customer replies.
Related
- Zoho and WhatsApp business automation
- Five signs your business needs automation, not another hire
- An AI assistant trained on your own data
Sources
- Barthélemy Thimonnier (Wikipedia)
- Encyclopedia of Greater Philadelphia: Garment Work and Workers
- James Bessen, "Toil and Technology", IMF Finance & Development (2015)
- TechBriefly: Canva hits 265 million monthly active users (Feb 2026)
- GitHub: Quantifying Copilot's impact on developer productivity
- METR: Early-2025 AI and experienced open-source developer productivity
- Stanford: The productivity paradox and the dynamo (Paul David)
- World Economic Forum: Future of Jobs Report 2025
This article shares research and opinion. Forecasts like the WEF's are estimates, and outcomes vary by country, industry and role.