Nobody Is Retraining a Truck Driver
Will AI take all our jobs?
This is the question I’m asked most often.
It’s often asked with a chuckle or laugh. It’s asked with the kind of nervousness that says, “I’m asking this question but really there’s only one answer I want to hear.” It’s a catchy headline-style question, but it’s not even the real question they’re asking.
They’re actually asking, “Will AI take millions of jobs (including mine), and will there be no new jobs?”
My response is the same each time.
Imagine we go back to 1950 and tell someone that “In 70 years, you’ll be able to communicate with almost anyone on the planet with three button clicks, and when you communicate, you’ll even be able to see them. In the process, however, all 1.5 million telephone switchboard operators will lose their jobs.”
This line almost always gets a chuckle. It’s easy to see that way more than 1.5 million jobs have been created because of the internet and video calling.
My take is that AI will displace lots of jobs and lots of workers, but I also think it will create more jobs and opportunities than we have now.
We have plenty of examples of this happening. Technological destruction has been occurring for thousands of years.
Spreadsheets and computers replaced human computers. Now there are millions of coders and data scientists around the world, and companies can’t hire enough of them.
The printing press replaced humans physically writing each copy of a book. Now we have more books than we can read.
Cranes have replaced humans dragging heavy things up ramps, yet we can’t find enough people to work in construction.
Warships that run on nuclear energy replaced ships running on coal and steam which replaced human rowers.
While I do think AI will displace millions of people, I think new opportunities will be created to make the world — both physical and digital — a much better and more productive place.
The AI-will-take-all-our-jobs crowd also seems to talk a lot about the need for retraining programs.
I’ve spoken with many people who think they understand what AI will and will not do, and there’s a sense of elitism to the retraining conversation.
When the conversation comes up, it often starts with someone saying something like this:
“I just don’t know how we are going to retrain some people. I mean, you can’t retrain a truck driver to become a coder.” Their implication is that everyone will learn to code, and everyone will build software to replace more and more jobs.
Nobody is retraining truck drivers.
We don’t have nearly enough of them. They do specialized work with lots of nuance that’s extremely difficult to automate. There will be automation in the trucking industry, but not in a way where all truck drivers lose their careers in one day.
Even if there is automation like autonomous caravans, where one human drives the lead-truck and autonomous trucks follow closely behind, that will still only make it so we don’t feel transportation bottlenecks and supply shortages.
The retraining will happen for the people who have repeatable 9-5 jobs that work mostly on computers. Those people will get displaced.
Well-created AI does noncreative, digital work better than humans do. It’s more repeatable. It’s much faster.
It’s not to say that all office workers will lose their jobs overnight. I don’t think that will happen either. What I do think will happen, and it’s already starting, is what I’m calling the 80% rule.
The 80% rule says that the complete technological destruction of any job description will come from 80% of that role’s day-to-day being automated. This rule applies to job descriptions that match all 3 of the following criteria:
They aren’t creating anything from nothing (design, code, multimedia, copy, anything qualitative)
They aren’t in leadership roles
They spend significant time using hard skills (that are largely noncreative) on a computer.
What will remain are corner-cases, one-offs, and tasks that aren’t currently ROI-positive to automate.
When the automations are released into production, they’ll be met, largely, with cheers. People are overworked and feel under appreciated. AI and automation will help reduce the workload, allowing people to focus more on the high-impact, high-value aspects of their days. And, they’ll have less to do.
How will the destruction begin?
Especially in large organizations where multiple people do the same role, when one person leaves, they won’t be replaced. Instead, an employee who stays will pick up the departed’s 20%. The good news is that I don’t think they will feel overworked.
Over time, another person will leave, and their role will not be filled. The responsibility will be divided among those remaining.
I don’t know what percentage of people will leave voluntarily and what percentage of people will be asked to leave. My gut says more people will leave the role, either to advance inside the organization or to find a new organization, than will be asked to leave.
As fewer and fewer people remain to do the 20%, the unit-economics of automating some of the corner cases and one-offs will change, making enhancements to the automation ROI-positive. More of the original job description will be automated, up until one person can do the work previously done by 5-10. 50 years ago it was common for an executive to have one or more secretaries just for sending letters. Now, with email, executives type their own emails and the people who used to send letters do far more creative work. Secretaries also used to screen phone calls. Now, everyone uses caller ID.
One customer service representative with great AI can do the work of 5 or 10 while the customer experience itself gets better.
Data validation can be done by an AI while the error rate decreases.
Five researchers (not hard-science research, but internet-based research) can be replaced by 2 with an AI and the research can be summarized, organized, and prioritized.
Which takes me back to my original point. It’s not the truck drivers that are being automated, it’s the desk jobs.
Plumbers, electricians, welders, and carpenters aren’t being automated by AI. There are no Zapier zaps or SaaS solutions for rewiring your house.
These jobs, while sometimes repeatable, have two unique characteristics. They’re fundamentally creative and they exist in the physical world, not the digital world.
There are Zapier zaps and SaaS solutions for updating project management platforms, doing research, finding data discrepancies, and doing customer service.
I know because I use them.
If both below are true, your job might be easily automated.
You had no problem going fully remote when lockdowns began (your work is largely digital)
Your work is largely task-oriented, routine, and detail-oriented
While “How to ensure you never get automated out of a job” is an essay for another time, it’s critical to remember that right now, there are more jobs open in the world than there are people to fill the roles. The demographics of the world are getting worse, not better.
More people are retiring than are entering the workforce. By a lot.
More people are becoming entrepreneurs and freelancers, leaving the corporate hierarchy completely, than ever before.
More people are searching for deep, purposeful work than ever before. Don’t believe me? Just ask 17–30-year-old what their biggest crisis is. Often, it’ll be something to do with “I just don’t know what my purpose is” or “I don’t know what I’m passionate about” or “I can’t figure out what I want to do.”
The truth is, I believe that the job displacement from AI will be a welcomed relief for the economy.
I think it’ll allow us to grow faster and solve new problems that aren’t being solved currently.
Problems like taking carbon out of the atmosphere.
Converting our energy from dirty fossil fuels to sustainable and clean.
Exploring how we can drastically increase the amount of healthy food we produce while decreasing the production cost.
Creating ways to lower the cost of daycare and increase the quality of education.
I’m not saying I don’t think AI taking jobs will be turbulent. I think there will be serious difficulties for some people and some companies.
I think the world will be a better, more creative place on the other side.
In the meantime, the 80% rule will apply. People who are displaced will need retrained, but they won’t be truck drivers.
Thanks to Courtney Thomson, Shannon Waller, and TJ Ryan for reading earlier drafts of this essay.