Cherry-Pickers Beware!

Tuesday, September 14, 2010

Googling around for something else, I ran into a piece by Alexis Madrigal of the Atlantic on "What's Wrong with 'X is Dead'." In that piece, Madrigal considers the latest attention-grabbing (but wrong) technological obituary -- Wired's "The Web is Dead. Long Live the Internet," which Paul Hsieh's GeekPress first brought to my attention. I duly smirked at Hsieh's quip about not getting the memo about "Macintosh, Linux, E-mail, [and] Facebook" all being dead, and duly followed his link to an entertaining list of several other technological "deaths."

In his Atlantic piece, Madrigal considers a couple of interesting aspects of the Wired argument. First, he looks at it as an example of a type of argument about how technological progress occurs. Second, he looks at the notion that such a change is "inevitable."

Regarding how Chris Anderson of Wired portrays technological progress, Madrigal notes that the argument was discredited long ago and simply doesn't stand up to any degree of scrutiny. He quotes Australian historian Carroll Pursell's summary of the argument:

An obsession with "innovation" leads to a tidy timeline of progress, focusing on iconic machines, but an investigation of "technology in use" reveals that some "things" appear, disappear, and reappear...
Afterwards, Madrigal notes that, quite often, "newer and older technologies happily coexist." My experience in the submarine force gives me an excellent example of exactly that: We used the energy from a nuclear reactor to heat water for use in a steam engine. I guess the internal combustion engine wasn't the final nail in the coffin of steam power.

In addition to Anderson's failure to consider how technological progress has occurred in the past, Madrigal notes that, while the graph of bandwidth share in his article (upper right) seems to support the idea of the web dying, with its share (red) in decline, that simply isn't the case. According to BoingBoing's Rob Beschizza, who considers the same data:
So with actual total traffic as the vertical axis, the graph would look more like this. [lower right -- ed]

Clearly on its last legs!

Assuming that this crudely renormalized graph is at all accurate, it doesn't even seem to be the case that the web's ongoing growth has slowed. It's rather been joined by even more explosive growth in file-sharing and video, which is often embedded in the web in any case. [bold added]
Madrigal doesn't put it this way, but a major problem with Anderson's argument is that there is an enormous amount of missing context -- in terms of both the history of technological progress in general and the relationship of web traffic data to all other Internet traffic. (Beschizza notes later that it may even be somewhat misleading to compare traffic volume for those different uses directly: "Does 50MB of YouTube kitteh represent more meaningful growth than a 5MB Wired feature?")

Another point that leaps out from a reading of Madrigal's article is that the "tidy timeline of progress" makes the conceptual error of ignoring how technological progress occurs, which much more often than not, entails building on the previous work of others. That is, one could think of such progress as representing division of labor (or even collaboration) over time.

Shifing gears to Madrigal's second point -- that he suspects that Anderson is allowing personal interests and ideological considerations to cloud his judgment... Near the end of his article, Madrigal notes:
[T]he great irony is that with this article, Anderson has done a masterful job of showing exactly how and why human beings try to shape the technological narrative of their worlds. We make arguments for personal and intellectual reasons based on our experience, desires, and ideological leanings.

Anderson doesn't work on, nor believe in, the economics of content on the web, and so while he's making his case against the web generally, he's also making the specific point that print and tablet editions of Wired make sense, but its website (which he doesn't edit) does not.
I don't know much about Madrigal or Anderson as thinkers, so I cannot speak to either of whether Anderson is a libertarian or Madrigal understands that many strands of libertarianism have a somewhat Marxist view of history, but I thought the following point on the "inevitability" of the change Anderson claims is occurring sounded spot-on:
Later, Anderson writes, "This is the natural path of industrialization: invention, propagation, adoption, control."

I wonder how many historians of technology would agree with him. It sure seems suspiciously like a "tidy timeline of progress," tinged with a little libertarian cynicism.
This is not to say that patterns of development in new industries don't naturally occur in economies. However, Anderson is wrong about at least part of what he thinks these patterns are, and these mistakes color how he looked at the data. In any event, even such patterns are not inevitable, given the pervasive interference of the government in our economy -- another piece of context Madrigal indicates that Anderson ignores.

If we assume, for the sake of argument, that Anderson is a libertarian, and that he is attempting to argue that "his" editions of Wired make sense, one conclusion is inescapable: he did himself no favors by trying to "shape the ... narrative," as Madrigal puts it, in an unobjective way.

-- CAV


Galileo Blogs said...

Nice call-out on the manipulation of statistical data, and why it was done.

Clever, but dishonest, people can manipulte statistics and graphical representations of statistics in many ways. As an economics professor of mine once said, "If you torture the data enough, nature will repent!"

Actually, nature never repents, despite all the efforts to try to make reality other than what it is.

Gus Van Horn said...


Sadly, even people who mean well sometimes fall for this or do it, out of carelessnes or, perhaps, desperation.

For example, I encountered a similar thing once -- manipulation of the horizontal AND vertical axes of a series of global temperature graphs -- in an anti-AGW book I was reading.

While I think it's a mistake to mix the political and scientific arguments together in that debate, the fact was that, thanks to such mixing, global warmists were able to call Ian Plimer out on exactly that, opening up the entire anti-global warming hysteria camp to smears about statistical manipulation.

Galileo Blogs said...

Good point, Gus. That is why one must *always* fairly represent the facts, even when some of them are inconvenient to one's own position.

A is A. Gee, who said that??

Gus Van Horn said...

I'll use this moment at the keyboard to note that Plimer, ironically, used exactly the quote you mention about making data repent.