Last night I heard Fred Guterl talk about his cheery new book, The Fate of the Species, which describes various ways we could do ourselves in, including heating up the world past habitability, accidentally evolving a virus that wipes us out, obliterating the plants and animals we need to survive, or destroying the digital world on which our real one increasingly depends. Technology got us into these messes, Guterl notes, and technology, he believes, will have to get us out. I guess that's so. But the more I think about it, the more I think his catalogue of highly possible self-inflicted wounds is an argument for bringing new technology online more slowly than we have been.
Here's why: Let us suppose that the United States were hit with a cyberattack that targeted the vulnerable interface between software and hardware. A cunning virus, say, that reports that power turbines are copacetic while simultaneously running them so they destroy themselves. (We've leveled such a weapon against Iran's nuclear program, so this is hardly a far-fetched scenario.) If hundreds of generators blew up at once, Guterl pointed out last night, we'd be hosed: We have no real capacity to make them any more, and they take time to manufacture. Power would be out for months.
Outside the room where he spoke, in Manhattan's Chelsea, pretty young men and women were chatting over pretty drinks in the soft sunset light. It seemed inconceivable that they could find themselves fighting over donkeys and dusty cans of tuna in a world with little light or medicine. But then, drastic events are always inconceivable until, one day, they're there. That's why we're never ready. The Pentagon considers this kind of catastrophe a real and present danger.
My first thought, on hearing this story, was this: It would be less of a danger if we still had a creaky industrial base that could make turbines. Instead, the United States depends on the sophisticated wonder that is the global supply chain, which allocates resources in a more efficient way for a world that works as expected. The unexpected leaves what is new, complex and sophisticated in ruins.
I was reminded of the last time I experienced a widespread power blackout, in August of 2003. (I've been through three; they're the reason I keep nothing important in the Cloud.) I had an Internet connection for hours after power went out in New York City. Why? Because I was using an old-school landline copper-wired phone line to supply a pokey old DSL connection. Landlines don't need electricity from the grid; they have a separate system. The DSL was out, but with a working phone I could plug in my ancient dial-up modem and connect until my laptop battery pooped out. In contrast, people with cable modems and cell phones were hosed immediately. Similarly, people who'd assumed the lightbulbs would never fail were left in the dark, while their neighbors dug out their candles and matches like 18th-century peasants and carried on.
The moral of all these stories, I think, is that newer technology is less well understood, less well-established, and less robust than older equipment. So while it's well and good to experiment with new solutions, we shouldn't rush to rely on them. And we certainly shouldn't rush to replace what is lumpy, old and slow with what is shiny, new and speedy. Shiny, new and speedy is what breaks down first.
Of course, there comes a point where one has to sell the horse and stop making papyrus. As anyone who has ever upgraded software knows, progress eventually forces us to give up well-loved features of old-tech for the benefits of the new. But just as I would never buy a new model computer on Day One or install version 1.0 of an application, so I think society would do well to avoid the risks of early adoption. I'm not sure Guterl would agree with me, but his book makes me think "not-early, not-late" adoption is probably the better bet for the species.
Follow me on Twitter: @davidberreby