Wednesday, January 28, 2015

Are we dumbing ourselves down?

kw: book reviews, nonfiction, technology, automation, surveys, critiques

Six years ago Historian George Dyson wrote on, "What if the cost of machines that think is people who don't?", summarizing an article by Nicholas Carr. Writing about 60 years earlier, in "The Feeling of Power" Isaac Asimov presented a future in which small calculating devices had so usurped arithmetical abilities, that a man who rediscovers paper-and-pencil methods of addition and subtraction is a phenomenon (Strangely, I haven't been able to find a date for this story).

Nicholas Carr has continued to think and write about automation and its effects on us. His recent book The Glass Cage: Automation and Us explores mainly the uglier side of current trends in technology. The best quote in the book is, "How far from the world do we want to retreat?" (p. 137)

Every person will have a unique answer to this question. For people like me, the answer would be, "Very far indeed, for long stretches of time, but with the option to return to full engagement at times and for durations of my choosing." For most of my life I have been more comfortable with machines than with people. Yet I need human contact…just not on the constant basis required by extroverts.

Technology is ancient and continuing: Stone tools as old as 3-4 million years; the successive technologies after Stone of Bronze, Iron, Steam and now Electronics; pocket computers we call "phones" for which making calls is now a minor function. The first time I saw a cell phone in use, some 15 years ago, there were two girls about 7 years old, running together through a park, each talking on a phone to someone else. (Note to self. Try making most calls while walking or jogging. Might be a good way to shed that next 5 pounds or so.) I recall predicting that during my lifetime, our "phone" would be installed in the mastoid bone at puberty and be entirely voice operated. Little wire to a microphone embedded in our jaw somewhere, and software filtering to subtract out the effects of flesh-to-bone conduction of our voice.

I am no slouch when it comes to computer use. I've been what was once called a Power User since the 1960s, when computers were too big to fit in most bedrooms. The motto of the Elephant Club: Don't Trust a Computer You Can See Over. Except today, a new club—the Power Tower Club?—might need a new motto: Don't Trust a Computer Smaller Than a Toaster Oven. Sure, my wife and I have a laptop, but my favorite workstation is a tower 18" tall (46 cm) with a pair of screens that gives me about a meter-wide view into cyberspace. For some of the work I do, that much screen real estate is essential. But do you know what one of my favorite activities is? A few times monthly I am a Historical Interpreter at Hagley Museum, in the Machine Shop, demonstrating machine tools (lathes, drill presses, mills, etc.) from the 1860s and 1870s, powered by a water mill in Brandywine Creek.

I wonder, though, if some machine workers of the early 1800s thought it was somehow "cheating" to use a power tool, when they were perfectly capable of making parts using hand tools. Probably not! Particularly for machine work, one needs a peculiar combination of intelligence and patience. I often point out to museum visitors that cutting the teeth on a medium-sized gear (5" thick and a foot in diameter; about 120 mm and 300+ mm) took a week in the 1870s. You set up a machine with eye and hand. You monitor the machine by ear; by the end of apprenticeship a machinist knows the changes in cutting sound that herald trouble on the way. So you need the brains to set the right index for a 17- or 19-tooth gear on a 40:1 indexing attachment, and the patience to listen for trouble for the next 60 working hours of your life…with resetting of the cut and rotation of the piece about 4 times per hour. Fast-forward to the modern era: Such a gear, if needed today, could be produced in a few minutes using electromachining, or in about an hour on a more conventional NC mill. Those old-time machinists would drool!

In many areas we are going through a transition, and Mr. Carr points out several of importance. The airline industry was among the first to automate wayfinding and autopilot aircraft control. If needed, any modern jetliner, and many smaller planes, are capable of taking off, flying themselves, and landing, without the pilot doing a thing. The trouble is, machines break, thunderstorms and solar flares disrupt communications and sometimes damage equipment, and because no program is totally bug-free, a rare combination of factors puts the autopilot's program into a confused state. In all these cases, the "fail safe" provisions immediately turn control over to the pilot. A few times, this has caused crashes, typically with the loss of everyone on board.

This brings to mind another principle that seems to be lost on modern engineers and programmers, "fail soft". Is it really appropriate for all the software to totally cut out so instantly? If the plane is at all still capable of level flight, the autopilot needs to alert the pilot(s) while keeping the plane on some standard course, giving the humans time to get their brain in gear. There may still be cases such as the "standard course" being straight into a mountainside (and I am reminded of the crash of a small plane in Malaysia in 1991, that effectively decapitated the Conoco corporation), but further development of "standard course" back-up routines ought to take care of that.

Such issues multiply when we come to the driverless car. It sounds seductive. Plot your course on a GPS navigator, press "GO" and take a nap, or play cards, or read or whatever. But the "lanes" in which an airplane "drives" are a few miles wide. Highways lanes are 12 feet (3.66 m) or less. For most of a plane flight, course corrections are few and may be hours apart. On the road, course corrections can occur minute-by-minute and even second-by-second. I have read a time or two about auto-driving "road trains", made up of dozens of autos on a superhighway at superspeed, inches apart to take advantage of drafting. Now suppose a solar magnetic storm disables half a dozen GPS satellites, the road-side "driving aid" equipment being relied upon by the cars, and perhaps some of the electronics in the cars. What is the "fail soft" scenario? Is one even possible??

We are in transition, all right. Casualties of all kinds are one price of progress. YouTube abounds with videos of people so engrossed in texting as they walk that they walk into fountains, manholes, lampposts and each other. We can expect the phones to become even smarter, so they would be on the lookout for such events. Maybe blare, "Look up, dope!", and make a red, flashing screen as such incidents approach. Smart phone technology is not yet complete, nor even appropriate for human use. It's why I use a flip phone that can call and make texts. Period.

In a late chapter Mr. Carr writes of the young Robert Frost and his poems about scythe work. The scythe is an extraordinary instrument. Using one creates muscle tone around your rib cage that no other exercise can match. Learn to use a scythe properly, and use it frequently, and you'll never have back problems. It exemplifies the kind of work that keeps a fellow close to the earth. Even as we try to re-educate America for a supposed post-manufacturing economy, there are huge numbers of jobs that remain very, very hands-on. A company may outsource its call center, computer programming, and database management to drones in India or China. You can't outsource construction, electrical work, plumbing and paving, nor landscaping or even repairing (and washing!) your car. Yeah, I know most New Yorkers just look puzzled when asked by a tourist where to buy gas…like they'd know! But deep in the bowels of the city are track workers and subway car maintenance folks that they'd suddenly feel a great need for if there were a month-long strike.

I have major mixed feelings about automated medicine, though. In certain cases, the Caduceus program has been able to make quick diagnoses where medical experts were baffled. But in others it has been embarrassingly off the mark. In medicine as in many other areas, the term "robotic" is being misused, most notably with the Da Vinci Robot for precision surgery. Let us reserve the term "robot" for autonomous devices such as the Roomba vacuum cleaner. The Da Vinci device is actually a tele-operated "Waldo" with vision magnification and down-scaled, feedback-enhanced motions so a surgeon can operate on something half an inch across while feeling like the object is the size of a basketball. I was once trained on a soldering Waldo used for attaching leads to integrated circuits. It worked at 25x, so a millimeter looked and felt like an inch. It greatly simplified the job. By the way, "Waldo" comes from an old story (1942!) by Robert Heinlein, where the concept was first made public.

Do I want a doctor to cede control in an operation to a robot? Probably not. Diagnosis? Not without human review. Only humans have a sense of what is sensible! How about prescribing? Ditto. I prefer the physician to not only make the decision, albeit aided by the computer system, but also to discuss it with me, because in teaching me how and why he chose a certain medication or treatment, he's rethinking it in a way that is useful to him and may cause him to realize something extra he might at first have missed (Feminists out there, I'd have used "she" and "her" if I had a female physician). "Thinking out loud" is often the most useful kind.

Artificial intelligence, once it gets on a par with us as a conversationalist, will still be quite different from us, so it could provide a very useful function: serving as a "straight man" to our musings, asking questions no human would think to ask, and adding a powerful level of synergy. A very neglected area of ergonomics has been to determine what tasks humans will always do better than machinery, and which tasks should be at the top of the list for turning over to machines. The various Zooniverse projects, citizen science at its best, primarily take advantage of our superior visual abilities. We can recognize the difference between spiral and elliptical galaxies at a glance; or different kinds of beetles; or see that a certain black-and-white blob is a rock and its shadow rather than a penguin.

Some might see The Glass Cage as a Luddite polemic. Not at all. Mr. Carr points out that we are a technological species. We can't live without it. Even the prototypical "cave man" was no naked savage killing prey with teeth and fingernails. The tool kit of Paleozoic people included dozens of tools that require skill to produce but reduce either the effort or the danger of doing the work. I, for one, am glad of today's technologies. I am equally glad that I can pick and choose which to use and which to ignore. Looking around the room I am writing in, I see several thousand objects, nearly all artifacts. Only the insect collection and a few shelves of mineral and fossil specimens are not technologically produced (though I used technology to mount and display them!).

Physically we are more "gracile" (that is, thinner and weaker) than the Cro-Magnons of just 20,000 years ago. They are called "modern" in an anthropological sense, but the technologies they inherited from their ancestors, and added to in following millennia, resulted in a modern civilization in which we don't need the great strength they required for day-to-day living. Our teeth are a little smaller, and as our jaws shorten, most of us need our wisdom teeth removed. All this results from technology. Will this age of intelligent machines cause our brains to atrophy? It's not likely. Our descendants will probably think differently than we do, just as we think differently than our grandparents who mostly grew up without radio, television, airplanes or automobiles.

I'm thinking of my own grandparents, here, all of whom were born in the late 1800s. I know most millennials are of an age to be my own grandchildren. We got our first television when I was 8 years old. Black and white, in a console the size of a divan. Our phone was on a party line. The most local of calls were made by clicking the button 1, 2, 3, or 4 times. "In-exchange" calls needed only 4 or 5 digits. All long distance calls went through an operator. So the change in the "thought world" of today's young adults is as different from mine as mine is from my own grandparents. It is another side of progress. So the book is more of a call to enter the future thoughtfully. We are creating it, after all.

No comments: