Thursday, December 7, 2017

Dilation and contraction.

Our standard narrative is deeply invested in the idea that progress is ever accelerating and that we are always on the cusp of an unimaginable leap forward. One of the favorite devices used to sustain this belief is dilation/contraction. The rate of technological change and its impact on society in the past is underestimated (particularly when describing the periods of the late 19th/early 20th centuries and the postwar era) while the current rate of change is overestimated, often wildly so when it comes to near-future predictions.

The following by Jim Hoagland is a good example.
Driverless cars and trucks rule the road, while robots “man” the factories. Super-smartphones hail Uber helicopters or even planes to fly their owners across mushrooming urban areas. Machines use algorithms to teach themselves cognitive tasks that once required human intelligence, wiping out millions of managerial, as well as industrial, jobs.

These are visions of a world remade — for the most part, in the next five to 10 years — by technological advances that form a fourth industrial revolution. You catch glimpses of the same visions today not only in Silicon Valley but also in Paris think tanks, Chinese electric-car factories or even here at the edge of the Sahara.

Technological disruption in the 21st century is different. Societies had years to adapt to change driven by the steam engine, electricity and the computer. Today, change is instant and ubiquitous. It arrives digitally across the globe all at once.

It is essential to note that, despite decades of serious and aggressive research and development, none of these technologies currently exist (at least not at the levels implied here) and, with the exception of autonomous vehicles, they probably won't exist a decade from now. Even with AVs, ruling the road will probably take decades unless we start heavily regulating non-autonomous cars.

(As we've pointed out before, there is an important distinction between driverless cars and driverless trucks. While both are coming, the current economic case for autonomous trucking makes far more sense and the technological challenge of driving a relatively small number of routes is considerably less. Long haul truck driving has the potential to go away quite suddenly.)

It's true that outside factors often slowed implementation in fields like electricity in the past. Factories had to be reconfigured. Power lines had to be laid. This meant that the adoption of certain technologies was delayed, particularly in certain localities, but even with this, the rate of technological change around the turn-of-the-century and, to a somewhat lesser extent, in the postwar era, was stunning.

More importantly in this case, (barring the special case where a new technology can be uploaded without any kind of change or upgrade of hardware) the same basic issues still apply. Even with AV's, there's still development time, infrastructure and adoption to consider.

The day is coming when you'll be able to say an address into your phone and have your car on its way without ever leaving the comfort of your couch. The prognosis for Hoagland's other just-around-the-corner innovations is considerably less promising. There are daunting problems involved with using AI to take over complex, badly understood tasks that largely lack reliable and agreed-upon metrics of success. While as for flying cars, the obstacles are huge and we have a history of failed promises going back literally a century. Just because some compulsive liar whose primary accomplishment has been getting gullible venture capitalists to write him ginormous checks makes this particular promise doesn't mean you should report it in the pages of the Washington Post as a done deal.

But even allowing for the vanishingly small chance that all of these things do come to pass roughly in the time frame given, we would still be nowhere near the level of technological change that people in the late 19th and early 20th centuries had to adapt to in virtually every aspect of their world, from day-to-day living to industry to transportation to mass media to telecommunications to medicine to agriculture and to all the others that I have invariably left off the list.

2 comments:

  1. I thought you would find this interesting, the November issue of Car and Driver included a very long feature on the future of driverless cars. It was edited by Malcolm Gladwell and features some great quotes from tech luminaries. It's basically a 2017 version of the 1900's Scientific America articles you have been posting. https://www.caranddriver.com/features/autonomous-addressing-the-totality-of-the-driverless-car-feature

    ReplyDelete
    Replies
    1. In addition to the stuff on Tech, I've also been working on a piece about Gladwell so I definitely need to read this.

      Delete