Part I, “Automation Armageddon: a Legitimate Worry?” reviewed the history of automation, focused on projections of gloom-and-doom.
Part II, “Automation: Robots in Real Life” reviewed how robots are ubiquitous and create jobs.
My first real job was creating a print estimating and production control system for five print plants scattered around the US. Each had at least one web press. These are enormous presses that use large rolls of paper and run at high speeds. They’re typically used for high-speed, high-volume printing.
Each plant also had at least one and often several sheetfed presses. These are presses that print on cut sheets of paper. Sheetfed presses are typically higher quality, used for things like high-end flyers, posters, or the covers of booklets.
Both the web and sheetfed presses had at least four units, one for cyan, magenta, yellow, and black, CMYK. If that sounds familiar, it’s because your inkjet printer uses the same four print heads because, when the first three of those colors are mixed together they can create any other color. They can also create “process black” but it’s cheaper to just black ink, plus the black ink looks better.
Most presses had extra units, inline print presses, to add “spot color’ – specific colors that popped out because they were not mixed together.
The covers, created by the sheetfed presses, were combined with the innards of the “book” (which wasn’t a book but that’s what it was called), in the “bindery,” the part of a print plant that does just what the name implies.
Of course, the presses must be fit with printing plates created by a prepress group, fit correctly on the press units, and the whole thing tied together to work within a fraction of a millimeter.
I showed up to get a feel for the place and, three steps in, was grabbed by somebody who said “get out: no loose clothing is allowed in here, especially ties.” Stepping back he explained the web press pulls 4-6 rolls of paper that weigh about 550 pounds (250kg) each at high speeds. If a hand gets stuck inside the person will lose their hand or arm. If a jacket gets stuck inside your shoulder or side is going to be crushed. But if a tie gets caught up you’re “186’d,” the color number for Pantone blood-red ink.
All this came to mind because of a conversation I had discussing the origins of automation technology. My first thought (and that of the person I was speaking to) was mill technologies of the 1700s are the first true automation machines. These are the mills that automated creating thread, yarn, and cloth. He must’ve thought about it then posited the printing press is arguably the first real piece of automation tech. I agreed – and still agree – but something felt wrong.
Sleeping on it, I realized the problem. The automated mill technologies required essentially no skill. People feed the machines raw materials and collect the output but the appeal of those machines as a business is anybody could operate them. Six-year-old children worked in those mills.
The printing press, on the other hand, always required skilled labor.
With Gutenberg’s first press, the magic was not the press itself, which was the simplest component. The technology came from the interchangeable letters, the ability to assemble them into a frame, specialized inks and papers, and an understanding about how to combine it all. It was so complex that Gutenberg spent his entire sizable inheritance inventing the whole thing then went into debt and bankrupt trying to commercialize it.
The printing press required vocational skills that could not be learned overnight. Even the earliest printers often belonged to guilds which initially certified quality and skill. These later morphed into trade unions but that was not their original purpose.
Conversely, the mill workers required virtually no skill. Life was miserable for mill workers until they were unionized.
High-skill workers traditionally have more bargaining power than lower-skill workers. Of course, even the most skilled laborers are still vulnerable to union-busting, keeping in mind the PATCO air controller strike and Reagan’s union-busting response. However, absent an ability to order military air controllers in – something no private business would be able to do – those air controllers should have been union-busting proof due to the skill needed to do the job.
The worry with the upcoming disruption by artificial intelligence is the value of these skills, and the bargaining power they bring will decrease.
And maybe that’s true. Let’s think of a few examples.
Will we need radiologists when an AI can do the same diagnosis faster, more accurately, and at far less cost? What’s the point of pharmacists even today when a computer can sort through every imaginable drug interaction? I’ve been tailgated by reckless semi-trailer truck drivers, probably trying to make a delivery, wishing a more patient computer was behind the wheel.
While we surely care about the potential job loss, it’s hard to argue that lower healthcare and transportation costs at higher quality with increased safety sounds bad.
Getting back to the printing press, in September 1959, Xerox released the Xerox 914 plain-paper copier. It required no skill whatsoever. Place a printed on the glass, press the button, and an exact replica pops out on plain paper. Sixty years later, it doesn’t seem like much but at the time it was magic. I can’t find any statistics but the 914 didn’t seem to displace any jobs. Surely there were people operating mimeographs and Watt Copying Presses but there’s no record of mass firings. Most likely these people were reassigned to more interesting work (a low bar since few jobs are less interesting).
Gutenberg’s original press and the myriad that followed didn’t displace jobs for scribes. They seem to have morphed into working as printers or teachers since illiteracy before the press was widespread. Many were monks and I’d imagine they found some other monk job.
As books became cheaper, more people learned to read. As a market for books grew, people not only wrote more books but the quality increased since authors — who wanted to be paid — added their names. Gutenberg’s press did not eliminate any jobs but did eliminate ignorance and eventually led to the Renaissance.
This doesn’t mean that automation technology does not facilitate the destruction of jobs. The high cost of the machines, especially when first introduced, ensures only the wealthiest can afford them. Patents granted by governments or natural monopolies oftentimes cause control of the automation tech to be concentrated into a small group of people who, throughout history, abuse that power. But it’s not the technology per se, it’s those who own and control it.
Richard Arkwright is the original English mill owner and inventor of the modern no-skill factory and company town. Historians seem to agree he was a terrible person. But his automation and business practices are two separate issues. There was more than enough profit for Arkwright to treat his workers better but no regulatory nor financial incentive to do so, plus he was just a bad person.
I’ll admit that this line of reasoning, that it’s the people and not the equipment, runs perilously parallel to a “guns don’t kill people – people kill people” argument. But I think that’s a false equivalence. In the case of handguns and the AR-15/AK-47 style weapons, killing people is their primary purpose. They have no other utility value beyond killing people. In contrast printing presses, cloth mills, and the countless other types of automation tech we’ve listed have lots of utility value. Turning people into robotic-like wage slaves is an abusive side effect of predatory businesspeople, not the primary purpose of the technology.