The Age of Automation

How Tech Is Coming For Our Jobs & Lessons from the Industrial Revolution

This is a weekly newsletter about how people and technology intersect. To receive this newsletter in your inbox each week, subscribe here:


The Age of Automation: How Tech Is Coming For Our Jobs

In the 1820s, a French mathematician used the word “travail” to describe the effect produced when one billiard ball hits another. People soon began to use the English translation—“work”—to describe a steam engine’s effort converting steam to motion. Over the ensuing decades, “work” came to broadly convey the effort that people spend on labor in exchange for money. Two centuries later, VCs like me co-opted the word to flood your inbox with vague references to “the future of work” 🤷‍♂️

I’ve written extensively about how work is changing, but I haven’t looked at how technology is replacing human jobs and how we can adapt to it. That’s my goal today.

I’m going to look at:

  1. Where automation stands today and where it’s heading,

  2. How technology will replace rote and thankless tasks,

  3. Why we need to revamp education alongside automation, and

  4. What lessons we can learn from the Industrial Revolution.


The Outlook for Automation in 2021

Work has been an integral part of people’s self-worth for hundreds of years. In England, last names were once determined by profession: Smith for blacksmiths, Miller for grain mill operators, Weaver for clothes makers. Children inherited both surname and profession, apprenticing to their parents and taking up the family trade. While we no longer take our names from our work, jobs remain central to identity: the question “What do you do?” is typically answered with what you do for work.

This is what makes the prospect of automation scary: in addition to losing a living wage, millions of people stand to lose their sense of purpose. And the numbers are scary; according to the World Economic Forum’s 2020 Future of Jobs Report:

  • 85 million people will be replaced by machines or algorithms by 2025

  • By 2025, humans and machines will do the same amount of work

  • 83% of businesses plan to accelerate their pace of automation

Companies are moving a greater share of work from humans to machines:

This means that there will be a massive shift from “jobs of the past” (data entry, administrative work, factory workers) to “jobs of the future” (AI and ML specialists, information security, robotics engineers):

Some predictions are even bolder: one expert portends that 50% of all jobs will be automated by AI in the next 15 years. Notably, these will be both low-wage jobs like truckers and factory workers, and high-wage jobs like radiologists and accountants.

The COVID pandemic has been an accelerant. Hotels are turning to robots to deliver fresh towels—unlike humans, robots don’t transmit the coronavirus. Workers once crowded into recycling facilities; now, robots use AI to sort through plastic, paper, and glass. AMP Robotics (Series B; $75M raised), which makes these recycling robots, has seen a 5x increase in customers since March.

Many employers don’t intend to hire back workers laid off in 2020. Instead, automation will replace workers and be more efficient, convenient, and cost effective.


Automating Rote, Uninspiring Work

In the 1980s, Dunkin’ Donuts released a famous commercial that came to be known as “Time to Make the Donuts”. In the ad, a middle-aged man drags himself out of bed in the wee hours of the morning, puts on his Dunkin’ Donuts uniform, and groans, “Time to make the donuts.” He shuffles out the door, and the commercial cuts to him returning home, bone-tired and muttering, “Made the donuts.” The cycle repeats again, and again, and again. The commercial was so popular that Dunkin’ made over a hundred different versions.

This is the type of work that technology should be replacing: the monotonous, thankless tasks. In the early years of America’s manufacturing boom, one worker on Henry Ford’s assembly line complained:

“It don’t stop. It just goes and goes and goes. I bet there’s men who have lived and died out there, never seen the end of that line. And they never will—because it’s endless. It’s like a serpent. It’s just all body, no tail.”

If technology does its job, the exhausted donut maker and the exasperated Ford worker will avoid the worst parts of their jobs and move on to more interesting work.

Startups are using artificial intelligence and machine learning to transform work. Within our portfolio at Index, five companies working on this are Aurora, Gong, Covariant, Brightback, and Instabase:

The economist Paul Krugman once said, “Productivity isn’t everything, but in the long run, it is almost everything.” At its core, automation unlocks productivity. My partner, Sarah Cannon, wrote a piece titled Freeing Humans to Do Their Best Work. That phrase captures the promise of AI: if technology does its job, workers can focus on more high-value and rewarding tasks.

Some people call these new in-demand skills “androrithms”—inherently human skills like emotional intelligence, creativity, and intuition. But shifting to a new skillset means that we need to revamp education to ensure that innovation doesn’t leave workers behind.


Reskilling Workers

In 2019, I had the chance to tour one of JD.com’s Beijing warehouses. JD—a $150 billion online retailer—operates 800 such warehouses across China, totaling 20 million square meters or about 3,000 football fields.

As we walked through the warehouse, our tour guide waved his hand toward the workers scanning packages and operating conveyor belts. “By next year,” he declared, “every single one of these workers will be automated.”

There was pride in his voice, but I remember feeling a chill—what would happen to the workers? This is the other critical piece of automation: reskilling workers in jobs of the past, for jobs of the future.

In America, only 0.1% of GDP is spent on helping workers retrain—less than half of what was spent 30 years ago. This is a policy failing. (House and Senate Democrats introduced a $25 billion workforce-retraining bill in May, but it hasn’t gotten passed Republicans. Republicans prefer to encourage retraining with tax credits.)

With the public sector leaving workers vulnerable to automation, the private sector is rising to the opportunity. There’s been a 5x increase in employer provisions of online learning opportunities to their workers, and 66% of businesses expect to see a positive ROI within a year of funding reskilling. Even Amazon—known as a soulless corporate overlord—is spending $700 million to retrain 100,000 employees. The initiative—called Upskilling 2025—teaches workers both skills they can use to get other jobs inside of Amazon (e.g., software engineering), and skills they can use to get jobs outside of Amazon (e.g., nursing).

The key question to answer is: Who’s paying for reskilling? Workers are ill-equipped to pay, and government hasn’t risen to the challenge. The most interesting companies figure out a clever mechanism.

Some companies—like Lambda (Series C; $122M raised), Flockjay (Series A; $14M raised), and Microverse (Seed; $3M raised)—use Income Share Agreements (ISAs), which align incentives with learners. Flockjay, for instance, only gets paid if students land a new job paying at least $40,000 per year. The average graduate gets a job paying $75,000.

These programs can be life-changing:

When Raven Winchester lost her job as a janitor last year, she enrolled in Flockjay’s 10-week course. She was hired as a sales development representative at LaunchDarkly—one of the businesses where she used to scrub toilets. As she tells it:

“The first day I stepped into the office an official employee, I just looked around at all of the things that I used to have to f*cking clean. It was an emotional moment. I cried like a baby.”

Other companies manage to get employers to pay for reskilling. Guild Education (Series D; $229M raised), for example, provides “education-as-a-benefit” to customers like Walmart, Chipotle, and Disney. A low-skilled Chipotle worker, for instance, can get her education paid for by Chipotle. Chipotle is willing to pay because offering the benefit reduces employee turnover. It’s less expensive to reimburse tuition than it is constantly need to hire new workers.


A Lesson from the Industrial Revolution

To learn how we might adjust to work’s current upheaval, we can look to the Industrial Revolution.

In 1814, the Boston Manufacturing Company built America’s first textile mill. An enterprising businessman had seen power looms on a visit to England, and he managed to memorize them in enough detail to recreate them across the Atlantic.

Immediately, workers were more productive. Output went up, and prices went down. Wages, though, remained low and stagnant for decades. The skills demanded by textile mills were different, and schools couldn’t adapt quickly enough. But by the end of the 1800s, workers were earning more than twice what they earned in 1830.

It took time, but a labor market for skilled mill workers developed. By the end of the 1800s, weaving a yard of cloth took only 2% of the human labor it took at the start of the century. Lower costs meant that demand skyrocketed. A single shirt had cost the equivalent of $3,500 in the 1300s. Now, everyone could afford clothes.

We’ve seen this same phenomenon in recent decades. Many people thought ATMs spelled the end of bank tellers. But ATMs drove up demand and lowered costs for banks, allowing banks to open more branches and hire more workers. As a result, there are more bank tellers today than there were in the 1990s.

This isn’t to gloss over the challenges in adapting to innovation. In the short term, the Industrial Revolution hurt workers: in 1800, a weaver could expect to earn 6 shillings a day, but by 1826, a weaver could expect to earn that same amount in a week. (In England, textile workers calling themselves “Luddites” stormed factories and burned equipment. The British Army was sent to fight the Luddites and at one point, there were as many soldiers fighting the Luddites as there were fighting Napoleon. Today, calling someone a “Luddite” refers to his or her fear of technology.)

The key is to efficiently and economically retrain displaced workers, mitigating negative short-term effects of innovation. In the long run, automation lowers costs and increases demand for workers, leading to more jobs. And importantly, these jobs are more highly-skilled and better paid.


Final Thoughts

Technology can do some jobs better. Just as power looms transformed productivity in the 1800s, a bricklaying robot today can lay more than 3,000 bricks in an eight-hour shift, 10x what a human can do. JP Morgan now uses AI to review commercial loan agreements; what used to take 360,000 hours of lawyers’ time can be done in seconds.

In 1930, John Maynard Keynes predicted that because of innovation, we’d all be working 15 hours per week in 2030. I’m sorry to say that that’s not going to happen. What Keynes missed is that innovation lowers costs and stirs up demand, representing more of a shift in how humans work than in the overall amount of time worked. But, if technology does its job, everyone will ultimately earn higher wages and do more rewarding work.

In the 1800s and 1900s, America shifted from an agricultural economy to a manufacturing economy. Today, we’re shifting again, this time toward a creative economy or knowledge economy or ideas economy.

In How Technology and COVID-19 Are Reinventing Education, I wrote that we’ll acquire more horizontal “human” skills early in life, then reinvent our job-specific skillsets throughout our careers:

Today’s graduates are expected to hold between 15 and 20 jobs (!) over the course of a career. Many haven’t been invented yet: 85% of today’s college students will have jobs in just 11 years that don’t currently exist.

We can point to historical examples to see that technological innovation is a tide that lifts all boats. This should compel us to build and invest in the companies leading the way. And just as critically, we should acknowledge that reskilling workers is equally important. We should also be designing business models (and policies) to ensure workers aren’t left behind by the pace of change.

For hundreds of years, jobs have been the lifeblood of our economy and the bedrock of our lives. We’re now facing the greatest disruption to work in at least 200 years. How we automate and how we respond to automation may be the defining problems of the 21st Century.


Sources & Additional Reading


Thanks for reading! Subscribe here to receive this newsletter in your inbox each week: