6-07-03 Diamonds and Water

The Harvard Business Review published an article by their editor-at-large Nicholas G. Carr titled "IT Doesn't Matter," wherein he asserts that IT is now a commodity and simply a step in the ongoing industrial revolution that began with steam, trains, electricity, telegraph, telephones, etc. I would provide a link to this article, but I find it telling that it is not available publicly (as I will revisit at the end of this weblog).

He makes a critical economic assumption: "what makes a resource truly strategic ... is not ubiquity but scarcity." In the midstream of economics ("all things being equal") this is typically true. But I would assert that computers still represent things on the edge, and economists have the "diamonds and water" observation that says "economics is not consistent across the spectrum." That is, water is incredibly precious and important, and yet it is common enough to be very cheap (or free; no one charges you for dipping a pail in a river), and diamonds are (outside of some industrial purposes) not particularly important, and yet they are expensive. Personally I think IT is more like water or air to businesses. It should not be scarce, but if there is a threat to its availability it suddenly becomes crucially important. Thus I do not think it can be considered using mainstream economic theories, as Carr seems to do.

I would also note -- in reference to his dissing of commodities -- that Warren Buffet made billions by investing in commodities.

Dawn observes: "How can something that requires this much human input be equated to steam and electricity? Do computers write code all by themselves yet? If not, you can't call IT a commodity."

He says "from a strategic standpoint, they (Industrial revolution technologies) became invisible; they no longer mattered." And I think if IT technologies could become invisible then they wouldn't "matter" unless they became unavailable, but because they require human interaction, and a lot of it, they are still very important. However, there are two types of human interaction that are required:

1. The human interaction to install and configure the machines and to maintain and update them. Here we are quite primitive; it's rare that you can plug into the network, for example, without having to do a lot of messing around. As things improve, there may come a day when this aspect of the machine becomes "invisible" and at that point I can imagine this aspect of the industry becoming "invisible." As I would hope that it would. Chuck was struggling to configure his Linux box and we observed that there are people who love to tinker with this kind of stuff. I used to do lots of work on my own cars, but I don't find it productive anymore, and I used to know a lot more about the innards of the machine than I do now, because I no longer find that productive. The essence of abstraction is to hide the parts that "don't matter." Note that Carr says "IT management, should, frankly, become boring," with which I absolutely agree. Dealing with configuration, installation, maintenance and update issues do nothing to move us forward or build anything. The easier and more automated those can be, the better.

2. The human interaction to create new content. There are aspects of this that can be automated -- for example, Google's "content" is the network of references that it creates to other content -- but at some point the words, whether they are text or code that produces new ways of creating other content, must be created by humans. Someday, AI may make inroads on this but with our current understanding of consciousness I can't imaging machines doing these things from end to end.

So while point one must be minimized in order to make computers cheap enough, point two is really the whole point of computers at all -- Steve Job's famous "Bicycles for the Mind" quote says it best. Without the human factor, the bicycle just sits there. And there are lots of improvements we can make to the bicycle, the paths, and the whole experience, and that's where IT does matter.

Here's an article in Fortune that rebuffs the HBR article. If you Google, you can find many other commentaries on the HBR article, but I was unable to find the article itself, except as a pay-per-download via Amazon. I think this little twist may be the biggest flaw in Carr's argument -- IT obviously matters enough to Carr himself that he wants to sell his article this way, rather than just post it.