rotation is great but someday machines will do something else & it’ll be amazing


Yglesias has just published a great post about robots’ persistent failure to put us all out of work. You should go read it immediately, forsaking any activities that contribute to GDP.

This is part of Matt’s thesis, of course: that information technology is contributing to welfare in ways that don’t show up in productivity numbers. This makes some sense. Right now the digital music player on my kitchen counter is humming some vaguely new-wavey Millennial band, and by doing so offering me both greater control and higher fidelity than its radio forebears. Those differences are not necessarily captured in its price tag.

On the other hand, I spent part of Sunday reading a great Wired story about sneaking paper maps out of post-Soviet Russia, smuggling briefcases full of cash into Parisian cafes while trailed by KGB agents. It sounds like it was considerably more trouble than I had this morning, when I researched, located and priced out Slovenian geospatial data from their government’s (non-English) website. It took me about one and a half cups of coffee to get through it, which were much more pleasant than flying across an ocean and maybe getting injected with polonium (the Parisian cafe bit sounds okay).

Still, I think Yglesias is probably right: there’s value showing up off the books, there’s goofing off, there’s overstatement of the importance of the IT-relevant section of the economy, and there’s genuine value being created. And it looks like it’s been a mistake to just assume that the last effect will surely swamp the others.

This makes IT innovations a different beast from affirmed productivity boosters like HVAC and dishwashers. Those things manipulate actual atoms, and seem to produce wealth and surplus time in a way that software might not.

Here’s the thing, though: from a certain perspective, those are pretty much all the same machine. You can create an astounding number of things with a motor and some switches and some surrounding plastic and metal bricabrac, including not only the aforementioned appliances but your dryer, refrigerator, automobile, running water, vacuum cleaner, garbage disposal and washer/dryer. Lately these appliances been tarted up with LEDs and microcontrollers, but they all come down to a fairly simple switch mechanism detecting input and, when appropriate, setting a motor turning. Powered rotary motion has been a great friend to humanity (not least thanks to the satisfaction to be had from an afternoon spent browsing websites about mechanical linkages).

But even the noble rotary motor has its limits, tasks that humans can do but which we cannot plausibly harness the awe-inspiring might of rotation and switches to accomplish. Preparing good meals, cleaning bathrooms, navigating road traffic–all of these require a bit more finesse. Or at least a much more complicated collection of switches and motors. Perhaps we won’t be able to deliver that finesse, but it sure looks like we’re close. The Roomba knockoff in my living room isn’t perfect, but it really has dramatically reduced the percentage of time I spend on certain classes of housework.

I won’t dwell on autonomous vehicles. But c’mon. Even the most curmudgeonly unmanned vehicle skeptic can buy a startlingly cheap drone. Even he must admit that we could deploy slow, unmenacing delivery droids to toddle along our sidewalks today, if we weren’t so convinced that actual cars, with all the inalienable rights American society affords them, were just around the corner. A truly enormous economic class of atom-moving will be converted from labor to capital in short order. It will involve information technology, yes. But this will be no Facebook.

And from there, who knows? I’m not sure that Asimo will be scrubbing my toilet anytime soon; perhaps one of his descendants will lift my withered body back into its goo-filled pod after every fresh organ installation. Moving atoms around is very hard if you’re facing a problem you can’t cheat your way out of with photolithography, and I doubt that a Google Car will be the harbinger of robotic burger chefs.

Still. The car thing. That will be a big one. We’re going to notice it.

One Response to “rotation is great but someday machines will do something else & it’ll be amazing”

  1. universal basic income | Tom Lee

    […] Software stuff, mostly — probably a bunch of machine learning projects that promise to finally invent a machine that does more than rotate, plus maybe one or two discount-rate materials science startups (I hear you can make quantum dots […]

Comments are closed.