this space intentionally left blank

July 23, 2009

Filed under: culture»pop»trendspotting

Cheap and Out of Control

Continued from.

Part 3: You keep using that word. I do not think it means what you think it means.

Before we go any further, I want to make something clear: I'm not opposed to "free" digital content, in either its monetary or political sense. Although I pay a small amount each month for hosting this site, it is served via Apache (free) on Linux (free), and is generated on the server by a set of (free) Perl CGI scripts. I log into the server using the PuTTY SSH client (free), and I view and test it using Firefox (free). Like almost everyone else, I use "free" ad-supported search engines and watch "free" ad-supported broadcast television. My current smartphone runs on a free, open-source operating system, and my previous phone OS is currently being open-sourced by its owners. I also eat the "free" samples at Price Club on Saturdays, if they're not too disgusting, which is a credit partly to my thriftiness but mainly to the strength of my stomach.

I don't have a problem with free, as long as we understand that "free" actually means "a wide range of well-known business models that shift costs to another location," also known as Chris Anderson's weak hypothesis in Free. If he'd written a book about that, I'd have little disagreement, but it would be a pointless book mostly composed of truisms. Much of Anderson's writing starts out in this mode. But inevitably, he keeps getting carried away and broadening it into a strong hypothesis that's untenable--either that the shifted costs, Heisenberg-like, cease to exist if he ignores them, or that "very cheap" is the same thing as free, or both.

That's largely the gist of chapters five through eight of Free. Anderson's wishfully-ambiguous conception of his subject from chapter four continues to shift to wherever his argument needs to be. It's incredibly frustrating--my notes are filled with repeated entries reading "so it's not really free, then." It's not that Anderson is unaware of these criticisms--he mentions them in passing at least a couple of times--but that he apparently dismisses them out of hand, or forgets about them in his rush to tell yet another overcooked, second-hand anecdote.

Chapter seven, for example, is devoted to Microsoft and the degree to which it has been threatened by free alternatives. Inarguably, Microsoft has been challenged in several markets by competing products that carry no up-front sticker price, and they've done their best to respond. The result has, in my opinion, been good for both Microsoft and for consumers--you can pry Firefox and Firebug out of my cold, dead hands, for example. But as a case study for how "free" will conquer all, you could not pick a worse company than Microsoft. In every example Anderson describes, by his own admission, they're thriving despite a paid-product revenue model. China? Heavily pirated and discounted, but still profitable. The desktop? Still controlling most of the market, and raking in money even on critical failures like Vista. The server? Incredibly, server software is one of Microsoft's biggest recent successes: IIS runs an astonishing majority of the web. Free software has challenged the software giant, but it shows no signs of killing them off anytime soon, and no cute Kubler-Ross reference on Anderson's part is going to change that.

But let's not get ahead of ourselves. Anderson opens chapter five with the story of Lewis Strauss, the man who coined his favorite phrase, "too cheap to meter." Strauss was discussing electricity, and of course you may have noticed the continued existence of power meters on buildings throughout the U.S. But, says Anderson (location 1219):

...what if Strauss had been right? What if electricity had, in fact, become virtually free?
Sure, and what if I had a pony? We can imagine all kinds of ways the world would be different if scarcity no longer applies--and Anderson does, laying out a vision of plentiful water, food, and clean fuel. But looking back, Strauss sounds like a crank. Anderson needs to show how his post-scarcity vision won't appear the same way in forty years, and using weasel-words like "virtually free" doesn't help.

Get used to it, though, because there's a lot of "virtually" free in Anderson's utopia, even though that's not the same thing as free at all. He seems to have an equivalence problem: make something small enough, and he'll swear it doesn't exist. For example, Anderson spends a lot of time in chapters five and six on Moore's Law and the price of transistors. He writes (location 1236):

In 1961 a single transistor cost $10. Two years later, it was $5. [...] Today, Intel's latest processor chips have about 2 billion transistors and cost around $300. So that means each transistor costs approximately 0.000015 cents. Which is to say, too cheap to meter.
Can you spot the fallacy? Yes, transistors are really cheap--which would be awesome if I bought computer hardware by the transistor. But of course single transistors are completely useless to me, or to anyone else. I need a bunch of them in a certain configuration, like the Core2 Duo in my laptop or the ARM in my phone, neither of which even remotely qualifies as "free." Anderson obviously knows this--he wrote the sticker price for an Intel chip in the previous sentence, for heaven's sake--but appears to be purposefully ignoring it.

He commits this same mistake when discussing Google (chapter eight is entirely devoted to Google, and is one of the most tedious things ever written). Google keeps building enormous, multi-million datacenters, but (he chortles) their cost-per-byte just keeps dropping! Why, they're practically free! Really? Is the company doing per-byte accounting, then? A huge datacenter may be a better value than the last one, but it still cost someone enough money to keep 70's-era The Who in guitar amps for at least a couple of years. I have the utmost respect for Google and their continued efforts to make their infrastructure both ecologically-friendly and energy-efficient, but their facilities are not "free" by any stretch of the imagination.

Anderson calls the combination of increasing bandwidth, processing power, and storage space a "triple play" that's "not too cheap to meter, as Strauss foretold, but too cheap to matter." (italics in the original) The elephant in the room is, too cheap for whom? All Anderson's examples revolve around fiber broadband and state-of-the-art PC hardware, probably because that's his experience. But even in this country, there are plenty of areas without a fast pipe, and plenty of people too poor to buy a machine that could fully exploit it. Not to mention the developing world.

Indeed, we might well ask "too cheap to matter" for what? In the last few years, commodity hardware has hit the point where it's sufficiently powerful for almost any local task (excepting, of course, heavy lifting like games and media production). I could run Word (or any similar native office suite) just fine on my old 366MHz Celeron. But according to Anderson the future is in the cloud, where an equivalent word processor will be implemented in a high-level scripting language that older hardware may struggle to interpret with the same responsiveness and power. A computer in a rural area (or a developing nation) may have difficulty pulling down pages fast enough to use those AJAX applications effectively. Anderson's hypothetical world is only free--or close enough that the cost can be waved away--for people who are urban, relatively wealthy, and have already sunk money into recent hardware. If you fall outside that cohort, the future of Free, isn't.

In interviews and responses to critics who have raised similar arguments about scale and definition, Anderson and his fellow travelers have not responded gracefully. He's not claiming that everything is free, Anderson says, just the important bits. But this has always been the problem with techno-utopian schemes, ranging from seasteading initiatives to the OLPC. The parts that he and his friends consider important (or unimportant, in the case of Google's extensive data-mining, for example) aren't necessarily the parts that translate across cultures, incomes, and geography. And while Anderson's demographic may not feel the cost of his revolution directly, it doesn't mean that it doesn't exist.

To his credit, Anderson points out one group for whom life is going to suck if his prediction comes true: the people driven out of business by the pursuit of Free's ideology. Wikipedia, he notes, has killed off what was left of the encyclopedia industry after Encarta demolished most of it. Craigslist has done a number on the newspaper industry. Anderson sees this as a "Robin Hood" transaction, decentralizing the flow of money, but admits that he could be wrong. We'll get to see in more depth how he thinks journalism (and the economy as a whole) can reinvent itself in chapters 9 and 10. As someone with no small amount of interest in the sector, and based on hints from Malcolm Gladwell's review, I can't help but dread it.

Future - Present - Past