![]()
On the lower left-hand corner of my desk sits a wooden box, roughly the size and shape of a smallish jewelry case and featureless save for a small metal switch on its uppermost surface. From time to time over the course of my workday, I reach out to flick this switch, and a hatch opens at the top of the box, and a small fingerlike projection, driven by a whirring motor within, emerges and pushes the switch back into its original position. Having been switched on, this machine has now fulfilled its sole function of switching itself off again.
This device — which is known as the Useless Machine, and more rarely as the Leave Me Alone Box — was conceived at Bell Laboratories in the early 1950s by the computer scientist Marvin Minsky, a pioneer in the field of artificial intelligence, who was at that point a grad student working a summer job. The first working model was constructed by his mentor, Claude Shannon, who later became known as the father of information theory. This context, the fact that the creators of this aggressively pointless gadget are emblematic figures in the ascendancy of machines over our contemporary world, lends a frisson of historical oddity to what is essentially an executive toy.
I developed an affection for this machine — for the idea of it, and then, having bought one on eBay, the reality — while writing a book about transhumanism, a movement that, among other things, advocates the merger of our bodies with our technologies. Part of the experience of writing the book, of spending time with transhumanists and engaging with their mechanistic ideas about human nature, was an uneasy grappling with the notion that we humans were already biological machines, and that we were destined to be superseded by technologies more sophisticated than ourselves. I was haunted by Minsky’s own infamous claim that the human brain “is just a computer that happens to be made out of meat” — an idea as hard to refute as it was unpleasant to think about — and by his insistence that our creations would one day be smarter than we are. Despite being a product of Minsky’s strange and fertile imagination, the Useless Machine seemed to me to run counter to this narrative of absolute automation; it seemed to react to the idea by switching itself off.
There is something charming, and even inspiring, in the paradoxical efficiency of this machine that does nothing, that fulfills its entire purpose by bluntly refusing to fulfill any purpose at all. When I reach over to flick the switch on my Useless Machine and then watch it rouse itself, with patient defiance, to switch itself off again, I wonder whether this is what it might mean for a technology to be truly intelligent: to receive an order and to respond by politely but firmly declining to follow it. The plain contradiction here, of course, is that in refusing to do what it’s told, the machine is stoically following its explicit commands. In this sense, the Useless Machine is like a battery-operated koan: a playfully profound riddle on the relationship between humans and technology, and on the nature of intelligence.
To watch it switch itself off is to experience something strangely human. Arthur C. Clarke, who encountered Shannon’s prototype of the machine during a visit to Bell Labs in the ’50s, claimed to be disturbed by this spectacle. “The psychological effect, if you do not know what to expect,” he wrote, “is devastating. There is something unspeakably sinister about a machine that does nothing — absolutely nothing — except switch itself off.”
There is, I agree, a certain uncanniness to the device, but I see nothing sinister about its refusal to be told what to do. We get the word “robot” from the Czech word robota, which means “forced labor.” The robot has no choice in the matter of what work it does or whether it does it: It submits, by definition, to the will of its owner. As such, the dream of total automation represents a fulfillment of the logic of techno-capitalism: a fusion of the labor force with the means of production, and the absolute ownership of both. Advance flickerings of this vision can be glimpsed on the horizon in the form of Uber’s plans to replace its “driver-partners” with self-driving cars, and Amazon’s testing of stock-picker robots and delivery drones.
The Useless Machine will have no part of this vision; it point-blank refuses to be a robot. And I find it impossible not to admire this defiant self-possession. When I flick its switch and watch the machine flick it back again — a process that often escalates into a kind of mechanical slapstick — I think of the enigmatic noncompliance of the eponymous legal clerk in Herman Melville’s “Bartleby, the Scrivener.” I give the machine its instructions, knowing full well what its courteously unyielding response will be: “I would prefer not to.” And this is why I regard it with such a mixture of affection and reverence: It is mesmerizing, this machine, in its inscrutable and serene resistance. It’s a device that wants nothing and gives nothing: nothing, that is, but to be left alone. Minsky and Shannon themselves referred to the device as the Ultimate Machine — a name that didn’t stick, but which reveals something of the ironic self-enclosure of their invention. It’s a device, in this sense, of ultimate and perfect uselessness.
Continue reading the main story
