Tom Harrison
Aug 28, 2017 · 3 min read

I have 30 years in the software industry, and am now working in the area that involves AI. I have several points, but I hasten to say I agree with your premise, if not the conclusions.

Automation of the form you present, and that which Musk and other visionaries have raised alarms, is the tail of the dog. From wet nose to rump, the dog has been automated.

Automation, computerized or other has been on a job-killing tear since the industrial revolution. The steam engine, then trains and cars put horses out of business. Electrification made longer workdays possible. Massive machines made underground coal mines unnessary. Fertilizer and pesticides and more big machines made family farms outmoded. Mass production of goods put the cobbler and seamstresses out of business. And robots took the people out of mass production. Walmart killed the grocer and pharmacist and general store. Amazon is killing Walmart.

These changes start with more employment. Then as they kick in, they kill the jobs the formerly “enhanced”. Above all, they wipe out the people whose lives and towns and habits, beliefs and expectations revolve around those jobs. Coal country. Steel country. Car country. Farm country. Not what they were fifty or a hundred years ago.

As you point out, the current change is characterized by its rate: while the first chunk took 50 years, the second took 25, the third has happened in about a decade. And it’s accelerating.

If you believe we’re in a time of social unrest in the world as I do, I think you need not look much further than automation through increasingly rapid technical abilities. Sure, standards of living seem to have increased in aggregate, but a great and rapidly increasing number of people have and will be displaced.

So this is the part of your premise I agree with. I guess the nature of how to address the consequences seems less clear to me. This is far from something new, and there’s a lot of momentum.

The concept of intelligent machines operated by despotic corporations has alarmed us for many decades. As a kid in the 60s, technology seemed wondrous. As I studied and worked with computers in the 80s notions of AI and evil corporations seemed more comical — computers were stupid and corporations even more so.

Today, AI is quite real. Systems that my team at work make are able to make predictions and inferences far better than people. Cars drive themselves better than people do. Today.

There is no bright line dividing intelligence from rote repetition, there is only the speed that useful intelligence can be employed. Slow computers don’t seem so bright, and take forever to learn. Fast ones may not be creative or intuitive in the way people are yet, but they do amazing things now, and the rest is coming quickly. Exponentially.

So if you or I think our knowledge-worker jobs are safe, we must think again. A common tool used by all software developers points out my errors of thought, allowing me to focus on how to solve the hard problems. Software to translate between languages used to be a party joke; now it’s better than most humans in many cases. How about journalism. Baseball stats are child’s play. So are obituaries, market reports, crime blotters, puzzles, and horoscopes. Other writing tasks like spelling, grammar, and style? Not to mention publishing, distribution and marketing.

So don’t get too comfy.

My final point is with your call to action. I am afraid that it’s too late for a populist, anti-technology uprising. And this despite having read and watched every dystopian book and movie ever made. To be sure AI and other forms of automation are growing fast. But far more important, power in the form of wealth is now more concentrated that ever before at least in modern history.

As far as I can tell, there are really only two jobs that are safe: soldiers, and the people that automate them.

)
    Tom Harrison

    Written by

    30 Years of Developing Software, 20 Years of Being a Parent, 10 Years of Being Old. (Effective: 2019)