I’m retired now but I worked in software development for over 30 years. I think it’s quite ironic that the Frankenstein monster of computer-based automation has finally turned on its creator. During the 1980’s and 90’s I used to say that the main result of my work was to facilitate unemployment because a lot of what I was doing was developing ways that a computer could do what a human did, but cheaper, faster, and more reliably. Though formally the work I did was to increase efficiency and effectiveness in tasks that needed to be performed by a business to keep the business going. That is, to make more money than it spent, which companies have been trying to do way before there were computers.
I worked on developing solutions that used software to replace humans. I lost count of the number of people who lost their jobs as a direct or indirect result of the work I did. To be honest it was for the better in many cases, but nonetheless, I helped create the entity that replaced them at the job they were currently doing to earn a living. In a fair number of cases the actual person or persons affected found new and better employment elsewhere (i.e., they were re-skilled, even if often reluctantly). But in terms of what they had been doing, their job was now gone, lost to the hands of humans. The work they once performed was now performed by a computer.
OK, now to the irony. This phenomenon is now about to happen to the very people who perpetrated the changes that affected so many others[i]. Developers are now, or will shortly be, losing their jobs because of automation. This time automation will be applied to their jobs. The work (that is the actions that produce outcomes) will still be performed, of course, but by a computer that has mastered their skills but expects no compensation in return. Namely machine learning, artificial intelligence, and robotics. Those who are swift of mind and action will see, and in many cases have already seen, this coming for a while. They have moved forward, along the arrow of time, to keep ahead of, or at least abreast of, changes. But humans can learn new stuff only so fast, and the ability to do so slowly diminishes over time.
Machine learning can cause a computer to learn at not only a higher rate than humans, but at an ever-increasing rate. Thus, over time, widening the knowledge gap between computers and humans. This is done using feedback loops, called backpropagation, coded into the instructions of the computer. Changes to the inputs to the computer’s instructions are made every time the instructions run which, of course, changes the outputs that are the result of each run. Over time the outputs based on these changing inputs get reliably closer and closer to what they are expected to be. When the computer gets as close as it can get, usually determined by a human, it is assumed that the instructions can produce reliable results even with inputs it has never “seen” before. This is quite a marvelous thing when you think about it.
Sooner or later humans reach their capacity for learning, either because the pace of change in their field has slowed down or the type of work has become irrelevant. In any case it can be done faster, cheaper, and more predictably by some means of automation.
Come to think of it, this is the way “progress” has always worked. However, that does not change the irony. No human is safe from, or immune from this phenomenon. You might ask, how will people earn a living if they can no longer stay at least a little bit ahead of the curve? After all, the paradigm has always been that human effort produces work, which is compensated for by money, so there is this unbroken connection for a person from effort to money. Finally, as money is accumulated, wealth is created. Wealth is then exchanged for sustenance. Sustenance thus provides the motivation to keep a cycle (or loop) going.
Maybe this “unbroken connection” is what will change in the future. Maybe more people will be on what many people now call “welfare”, however it won’t be called that in the future. There won’t be any social stigma about it at all. It will simply be the way that people get sustenance. If it’s true that artificial intelligence and its means of coming about: machine learning and robotics, will eventually create more wealth, then at least some of that wealth will need to be distributed, without dependency on their effort, to more and more humans. If not, sustenance will diminish, and human population will crash increasing the odds that the technological singularity[ii] will occur much faster than anticipated. This will happen simply because there will be fewer humans to oppose it. Those who remain will become more and more dependent on automated means for their sustenance. AGI (artificial general intelligence) will have arrived. Not because artificial intelligence has become smarter, but because humans have become less so.
[i] Not literally the same people. For example, I escaped unscathed and have been able to retire reasonably well-off. It’s the people who follow who will have to fight for their meals. I guess the sins of one generation are always visited upon the next.
[ii] The book “The Singularity is Near: When Humans Transcend Biology” by Ray Kurzweil was published in 2005. In it he uses the term “singularity” usually without the term “technological”. The addition may be to differentiate it from the same term used in Physics to denote the interface between a blackhole and everything else.
Birkdale Computing 