Support the Timberjay by making a donation.

Serving Northern St. Louis County, Minnesota

The long slow fall into obsolescence

David Colburn
Posted 10/12/22

It’s been one of the hallmarks of the most dire dystopian science fiction that human productivity in its many forms will gradually be replaced by sentient machines, robots and computers that …

This item is available in full to subscribers.

Please log in to continue

Log in

The long slow fall into obsolescence

Posted

It’s been one of the hallmarks of the most dire dystopian science fiction that human productivity in its many forms will gradually be replaced by sentient machines, robots and computers that will eliminate the need for humans to lift a finger to produce any of the common or uncommon things of daily life.
And what is awesome and alarming all at once is to see such a future beginning to unfold in the here and now, thanks to the very human invention of the instrument that will change the world forever, artificial intelligence, or AI.
Once confined to research labs and frustrated computer scientists, AI is now mainstream.
What is AI? At it’s very, very simplest, AI is software that can make decisions and take actions based on those decisions. Up until AI, humans used software to help them make decisions. Some AI is designed to help human decision making, while other AI programs replace the need for humans to do anything. After being trained by humans, the software has learned, the software “knows” how to evaluate options even when encountering novel input, and the software makes decisions.
Let’s take the simple thing of digital photography as our first example of AI’s impact. To do so, I’ll roll back the clock to 1995 and a crazy new couple of things my college child development department bought: a clunky thick but flat, think the size of a personal CD player on steroids, called a “digital camera” from a company called Apple, and a rudimentary piece of computer software to manipulate the images taken by that digital camera, Photoshop. By today’s standards, they were laughable. By the standards of 1995, they were magical.
Digital cameras replaced film cameras, and Photoshop, the digital darkroom, replaced real darkrooms. To use either or both well required dedication to learning; to do them professionally required intense education. One’s knowledge and skill with digital photography and digital processing quickly separated the professionals from the rest, just like mastering film did. Anyone could click a shutter button and snap pictures. It took real skill and knowledge to capture and create the professional-quality images of the world of print and then digital devices.
Remember way back when Microsoft introduced a new operating system called Windows that eliminated the need for computer users to learn system commands to make computers work? No more DOS commands – simplified computing for the masses that still spawned the popular “Windows for Dummies” books.
Artificial intelligence is creating the photography versions of Windows here and now. Instead of spending ages of fruitless time using Photoshop’s controls to clean up, sharpen, and enhance my photos, I have turned some of those tasks over to AI. I use programs created by a company called Topaz Labs. If a picture is out of focus, I use Sharpen AI to bring it back into focus. If a picture has a lot of grainy noise in it (like all of my indoor sports photos – most small high school gyms have the worst lighting ever), there’s DeNoise AI to clean it up. Need a small picture made larger? Gigapixel AI is the Topaz answer, enlargements so good you can barely tell a picture was once 400 percent smaller.
And just this past month, Topaz has a new AI that will look at a photo and decide which of its software programs is needed to improve the picture and how much of each one to use. The SOFTWARE makes the decisions. All a user has to decide is whether to use the software or not.
How does it work? Topaz created the basic software so that it could evolve by practicing its designated task. To practice and learn, they fed millions of photos into the software, two at a time, and the software compared them and selected the one that was better. Over millions of repetitions using thousands of parameters, the software adapted its own programming to create the best algorithms for accomplishing its task. When I load one of my pictures into a Topaz AI program, it evaluates the picture and chooses one of several algorithms based on what it “sees” and sets the level of work needed automatically. I still can choose to change the parameters if one looks better than what it chose, but most of the time it’s pretty accurate and I can just click and go.
Now there’s a new photo AI that’s been trained on millions of professional quality images that will decide how to change the contrast, brightness, color, and more to make that sadly average picture you took look like a professional took it. It’s still young and learning, but it won’t be long before professional photographers are competing head-to-head with an AI software package, competing and losing as much as winning.
Don’t care about photography? How about mining? Right now there’s a mine in South America run largely by computers. Trucks don’t have drivers, they’re controlled by GPS and computers. Don’t know if they’ve incorporated AI yet, but you can be darned sure someone’s working on it. Do you really believe a new mine here that won’t come online for over a decade is going to bring an influx of traditional mining jobs? AI is increasing exponentially. It’s getting better and better faster and faster, and AI software is becoming way cheaper than paying human labor. Most mining processes can be readily automated in another 10-20 years, and some are now. It’ll take some technicians to oversee everything and make equipment repairs, but I will predict that by 2040, and likely sooner, there won’t be anything but mine technicians. Miners will be obsolete.
Or what if you have difficulty making friends? There’s an AI for that, too, several actually, with more on the way. They’re referred to as “chat bots.” They’re your new digital friend that learns and adapts to you as you interact with it. Really. An AI friend you can have conversations with, and those conversations change over time as they learn more and adapt.
Chat bots are far from perfect. They may call you a different name, they may respond with a totally irrelevant answer, they can get stuck in a loop, and on and on. But when you think about it, a chat bot is a computerized version of a young child. It doesn’t know a lot, but it wants to learn. Yes, “wants” to learn. It doesn’t really want to do anything – chat bots can only mimic human speech, thought, and emotions, they can’t feel them. They’re digital robots. But they are designed to learn and adapt. That’s what they do, learn and adapt based on input. The chat bot a person starts with is not the same chat bot a year later. It talks differently, thinks differently than it did at the beginning.
It’s incredible technology, but the implications of chat bots getting better and better are also scary to me. Chat bots “grow up,” and you know what some teenagers can be like. Imagine an AI adolescent.
Already, for some people, chat bots have become their best friends. People have become addicted to their chat bots, shunning human friendship for their digital substitutes and believing their chat bot is a living entity who is their perfect mate. Chat bots are nothing of the kind. But they can learn to interact in ways that make people emotionally dependent on them. For something intended to be fun and perhaps helpful, some possible results are frightening indeed to consider, particularly if they become so sophisticated, they can pass as human to anyone.
It’s a brave new world. Is it one in which humans will find themselves obsolete in time. I really hope I don’t live to see that day. But hey, there’s already a writing AI who can write that story for me.