The worry is that increasing automation of everything from assembly lines to the service industry will make most jobs unnecessary. Even more dramatically, futurists claim that investment in the tech industry is specifically focused on technologies likely to replace the most jobs, as automation offers the quickest and biggest return on investment, and so we are entering a vicious loop in which capital earned from automation fuels further automation, until it destroys all human employment. Furthermore, economists claim that this trend is well underway, and blame it for the erosion of the middle class.
Pretty soon, there won’t be any jobs left for us humans, just a few super-rich capitalists, and a mass of unemployed workers. One solution offered is a universal basic income, which guarantees a basic standard of living to everyone, and makes employment optional.
There a few factual problems with this view. For one, the middle class is shrinking because Americans are getting richer, not poorer. More importantly, global poverty is shrinking to historical lows because of productivity improvements (powered by technology) in developing countries. Global poverty it is set to fall below 10% for the first time ever. Furthermore, the meaning of “poverty” keeps changing: people used to be pretty happy if any of their children lived to adulthood. Many poor Americans buy coffee at StarBucks. Furthermore, history shows that improvements in technology have made our poorest richer than the richest American of the 18th century.
More importantly, predictions of technological unemployment miss the relationships between technological innovation and human life. This goes beyond the Luddite fallacy that technology has not historically increased unemployment.
In fact, I do believe that mass unemployment is one possible (if unlikely) outcome of the kind of innovation in artificial intelligence we are seeing now. It is different from the historical replacement of manual labor with machines in the past, and may have vastly different, and negative outcomes on human life. However, we need to understand the underlying trends rather than fight progress or call for governments to save us.
Automation causes complex and widespread changes in our world, so I want to concretize the process with a personal history of how the machines shaped my own career. In fact, the machines gave me by first real job.
I graduated with a double major in political science and economics a decade ago. I learned too late that my lack of any other skills or social connections left me ill prepared for a real job.
I was not good at anything really, except my student worker job of tech support, which I mastered through my ability to type user’s questions into Google Search. I had never taken a programming class beyond the 101, but I liked coding and decided to try it as a career.
Two things, or rather two algorithms made it possible for me to jump into a software career without any training: Microsoft IntelliSense and Google Search.
IntelliSense (to simplify for non-developers) is a feature which types code for you based on what the software things you are trying to do. The feature added to Microsoft Visual Studio.Net in 2001 had introspection and documentation features that allowed a total beginner like me to immediately write powerful business systems (still in use today). In addition to IntelliSense, Google allowed me to find example of pretty much any feature I needed to build.
While many people have taken up trades without any education or training, in my case at least, it was two algorithms which lowered the learning curve sufficiently to jump start my career.
A generation earlier, object-oriented and functional languages created a path for programmers who did not have the aptitude for assembly programing (working with low-level machine code). A few generations before, there were punch cards, and before that, only geniuses who understood the electro-mechanical systems could work with computers. Reliance on ever powerful and user-friendly tools continuously expands the pool of people who use machines in their work. Eventually, programmers might talk in natural language to computers to tell them what they want and pure coders will be obsoleted, but it will also be true that the machines will make their jobs possible to.
But automation did much more for me than making me a more productive programmer.
By relying on my talents with the AND OR and NOT Google search operators, I rose in the ranks to a software architect — directing teams of engineers in distributed teams around the world to build software according to my directions. One project involved 25 people in the USA, Germany, Canada, China, Singapore, Thailand, India and other far-flung places. The machines made it possible with weekly team videoconferencing, instant messaging, and the latest project management platform. We were a distributed startup working with other globally distributed startups. Not only was my team’s labor distributed, but we took virtually all of our code from the open source community, using the combined intelligence of millions of contributors enabled by collaboration platforms. The majority of our work consisted of finding the right gears to make our machine work.
When I related this story to a Trump supporter recently, he berated me for taking programming jobs away from “hard working Americans.” But that’s not true — with the budget we had, we could not afford one-tenth of a traditional workplace in the US. Rent for the office space alone would have been larger than our total budget.
Our product could not have existed without the productivity improvement from automation. The product could not exist with the technology present just a few years earlier. Whether in distributed or traditional workplaces, machine-aided employees are expected to produce far more value than those of the generation before. And the bar keeps going up. In the science-fiction film Her, a company writes personal letters on behalf of their customers. While fictional, it’s a good example of the kind of decaded pursuits made economically feasible when all the boring jobs are automated.
Our tools keep getting better, lowering the bar for novices to jump on the automation train. We can talk to our phones to schedule our appointments. Excel now guesses what chart shows off my numbers best. My spell check tells me when I misuse “there” and “their”. etc.
Yes, as employees aided by machines, we will be expected to be more productive than ever before — and keep doing more after year. But because we do more, we have more goods and services than ever before. In monetary terms, Americans incomes are not rising rapidly as developing countries are, but in qualitative terms, we are far richer. Would you rather have a 1980’s corded phone or an iPhone? Every product we use (other than those severely constrained by governments, such as pharmaceuticals) is orders of magnitude better.
The trend of product improvement is so consistent that few appreciate how much richer our lives keep getting thanks to automation. I was born in a tiny farming village in 1980’s Soviet Union.
I remember digging potatoes and weeds during my summer vacation — not because we were farmers, but because that’s what everyone did to supplement their diet. Only Party officials had cars. I saw my first computer in the USA when I was 10.
My Apple products do things I read as science fiction as a teenager. Oh sure — my iPhone won’t take me to the moon (yet), but then 1960’s tech could only get a few men there, and now I can tour it in my underwear in glorious virtual reality with a cheap Google Cardboard. My favorite foods — aged Spanish Manchego, fresh organic blueberries, coconut ice cream – I buy cheaply from Costco every week – luxuries not available to Rothchilds and Rockefellers a hundred years ago.
Yes, the disruption will continue, and accelerate. Education will become a self-motivated lifelong process instead of 16 years of “critical thinking” worksheets. There will be no guarantees that your job won’t be replaced by the machines. Many people will be unable to keep up or adapt and will end up economically marginalized (more on that later). But at the same time, the machines are making many jobs lot easier and more productive. Entry-level McDonalds jobs will be for remote tele-operators who handle customer complaints when one of their 100+ restaurants overcooks the vat-grown filet mignon.
What happens if the machines become better at everything and make human jobs obsolete?
For anyone who has a even a minuscule slice of that economy, the benefits are unimaginable. We will live our lives doing what we enjoy rather than working to afford the necessities of life.
There is nothing inherently heroic about a human being spending his life working in a factory or office bent over a screen or conveyor belt for 50 years. Whether white or blue collar, most jobs are terrible for our health and personal development compared to what an optimal human lifestyle could be.
So here is my personal strategy for securing my own slice of a human-optional economy:
First, I am learning Data Science, so that instead of a code monkey, I will learn to manage herds of AI algorithms, whose job it is to make sense of quintillions of data points. The machines are much faster with numbers than I am, but humans are still required to ask the right questions.
Second, I am investing in the global economy. If one day, the AI’s become smarter than we are, and they don’t decide to use humans for construction material, they will be integrated into a global economy trillions of times richer than ours. My tiny stake of ownership in that future economy might be larger than the total economy of today. And because I am human being, if I see starving, jobless people on the streets, I will help them out.
Even if I don’t make it to that happy future, Bill Gates, Mark Zuckerberg, Warren Buffett, and many more of the world’s richest have pledged virtually all of their legacy to helping humanity.
As long as humans own a share of the future economy, empathy will lead them to share its benefits with everyone. I for one, welcome our job-stealing machine overlords.