I want to associate myself with Adam’s excellent comments about Jonathan Zittrain’s book. I haven’t read the book yet, so I won’t try to comment on the specifics of Zittrain’s argument, but it strikes me that if Adam is summarizing the book fairly, ZIttrain’s thesis is strikingly similar to the thesis of Larry Lessig’s Code: The open Internet is great, but if we don’t take action soon it will turn into a bad, proprietary, corporatized network. I’ve been mildly surprised at how little comment there’s been on how spectacularly wrong Lessig’s specific predictions in Code turned out to be. Lessig was absolutely convinced that a system of robust user authentication would put an end to the Internet’s free-wheeling, decentralized nature. Not only has that not happened, but I suspect that few would seriously defend Lessig’s specific prediction will come to pass.
But while Lessig’s specific prediction turned out to be wrong, the general thrust of his argument—that open systems are unstable and will implode unless managed just right—is alive and well. I think that basic claim is still wrong. And I think it’s not a coincidence that these kinds of critiques often come from the left-hand side of the political spectrum (I don’t actually know Zittrain’s politics, but Lessig is certainly a leftie). It seems to me that left-of-center techies are in a bit of an awkward position because on the one hand they’ve fallen in love with the open, decentralized architecture that is epitomized by the Internet, but are predisposed to criticize the open, decentralized economic system called the free market. As a result, they wind up taking the somewhat incongruous stance that to preserve the decentralized nature of our technological systems, we need to have more centralization of our economic and political system. Zittrain’s choice of the Manhattan Project as a metaphor for the way to preserve the Internet’s openness is particularly striking, because of course the Manhattan Project was the absolute antithesis of the philosophy behind TCP/IP. It was a hierarchical, secret, centrally planned effort that left no room for dissent, diversity or public scrutiny.
The typical life cycle of a technology goes something like this. Technologies are usually originated in the laboratory, where a small number of geeks invent it and explore its capacities. Once the time comes for commercialization, often proprietary versions are the first out of the chute because it’s easier to monetize closed platforms, and therefore to raise the capital necessary to deploy them quickly. So in the early years of any new technology’s growth, it often looks like the proprietary technology has a decisive advantage (think AOL in 1993). Then, as the technology begins to mature, the disadvantages of closed technologies become apparent. They can only grow and evolve as fast as their owners can manage them, and as their owners get larger and more bureaucratic (think AOL in 1998), these platforms begin to stagnate. Meanwhile, the open alternatives, which are not held back by centralized management, continues growing rapidly, equalling and then quickly surpassing the closed competitors. Finally, the open platform’s lead gets so large that the closed platform, facing a choice between interoperability or irrelevance, is forced to open itself up to the competition (think AOL in 2003).
And once an open platform has become firmly established, proprietary firms stop trying to dislodge it. Instead, they try to build new proprietary technologies atop the underlying open architecture. Mac OS X, for example, is a thin layer of proprietary software atop a big stack of open technologies. Similarly, Facebook is a thin layer of proprietary code atop a lot of open Internet technologies. But that means that even as a company is trying to establish the dominance of their new, proprietary platform, they’re reinforcing the primacy of the underlying open architecture. Which means that that open architecture remains available to be built on further by anyone who cares to do so. And that, in turn, ensures that the process I described in the previous paragraph can begin again at another layer of the software stack.
What happens, though, is that every time this process begins in a new sector of the economy—online access in the mid-1990s, e-commerce in the late 1990s, broadband access in the early 2000s, social networking today—folks on the left-hand side of the political spectrum begin wringing their hands about how this time is going to be different. Sure, open platforms have had a good run so far, but now the big, bad corporations are going to take over. And because the specific names and technologies are different each time, it’s always possible to come up with a theory about the specific developments that will bring about the triumph of the walled gardens. Their warnings invariably turn out to be overblown, but by the time it becomes clear that the last round of predictions were wrong—Lessig’s 1999 predictions about e-commerce destroying the Internet, say—there’s a new threat to obsess over.
This, incidentally, is why it annoys me so much when libertarians denigrate the value of open platforms. The triumph of open architectures over the last couple of decades has been a vindication of the Hayekian idea of spontaneous order. AOL tried to build a closed, centrally-planned network, and it got clobbered by the Internet for precisely the same reasons that the US economy outperformed the Soviet economy: central planners lack sufficient information to satisfy the needs of millions of users with diverse needs. What the answer to the Lessigs and Zittrains of the world isn’t that open systems are bad. It’s that precisely because open systems are good, they’re unlikely to be dislodged by closed systems in the marketplace. Even when the structure of the market is far from ideal, as it is, for example, in the broadband duopoly, the open architectures have turned out to be far more robust than anyone expected. Indeed, the only way they are likely to be undermined is if you give the federal government—the largest, most bureaucratic hierarchical institution on the planet— too much control over them. Which is why the “Manhattan Project” is a terrible metaphor for the kind of effort we need.
Just to be clear, this is all based on Adam’s summary of Zittrain’s book, so my characterization should be taken with a grain of salt. I hope to get a chance to read the book myself, and I’m sure I’ll have further (and perhaps different) thoughts then.