Law in the Internet Society

The attention switch

-- By SamuelRoth - 16 Oct 2014

In “The Attention Economy and the Net,” Michael Goldhaber suggests that the Internet has transformed attention into an emergent form of value, and predicts that success and failure in the internet society will be determined in large measure by the amount of attention one can garner for oneself or one’s project.

I think that Goldhaber is exactly half-right. As part of our analysis in class of the new economic forms of the Internet age, we have distinguished between functional bitstreams, such as software, maps, and mathematical equations, and non-functional bitstreams, such as music, video, and literary text. Looking at the changes the Internet has wrought, we can observe that attention has grown increasingly important to the success of functional bitstreams. In the context of non-functional bitstreams, however, attention is playing a smaller and smaller role.

Non-functional bitstreams: from 15 minutes to 15 people

In the pre-Internet economy, non-functional bitstreams needed a great deal of concentrated attention to survive. The logistics of nationwide distribution through narrow channels—a handful of radio stations in each market, one or two TV stations that played music, a similarly limited number of record stores with finite shelf space—meant that a music act had to sell tens of thousands of albums at a time or face exclusion from the industry altogether. For instance, “a classical album [was] considered a substantial success if it [sold] 70,000 copies as a new release.”

By effectively eliminating marginal costs, reducing fixed costs, and widening the channels of distribution, however, the Internet made it possible for producers of non-functional bitstreams to distribute their works little by little, finding niches and expanding them over time. Sustaining a high level of public attention, which had been the sine qua non of success in the pre-Internet aesthetic markets, ceased to be a prerequisite for access to the channels of distribution. It was no longer necessary to be world-famous for 15 minutes; artists could instead focus on being world-famous to 15 people.

More to the point, because the internet allowed niche, low-attention products to proliferate, consumers of non-functional bitstreams no longer had to choose from a small set of options, each in its way made inferior by the need to cultivate wide-spread attention all at once. Thus, the availability of low-attention non-functional bitstreams on the Internet began to threaten the viability of the high-attention non-functional bitstreams of the pre-Internet economy. See, e.g., Matt Richel, Record Labels Assert Control in Cyberspace, in which early Internet-based music distribution companies, which of necessity specialized in the work of independent artists, express their apprehension that the major record companies were ultimately drawn to the Internet to “brake the momentum of the Net as an alternative source of music” and “‘maintain control of distribution.’”

In other words, the music giants didn’t come to the Internet to fend off pirates; they came to fend off an army of unpolished, low-attention artists who, collectively, were much better at giving audiences what they wanted to hear than a handful of high-powered, high-attention acts. The giants’ ability to build a brief crescendo of attention for their non-functional bitstreams was no longer enough.

Functional bitstreams: We’re going to need a lot of eyeballs

Functional bitstreams, meanwhile, have had an opposite experience of attention inversion. In the pre-Internet economy, the relative simplicity of non-functional bitstreams, together with economic and technical limitations on distribution, meant that UNIX could percolate in Bell Labs and a handful of computing centers for over a decade before it faced much external pressure to innovate. At IBM, a single employee might manage a whole programming language, interacting with individual users in a hub-and-spoke paradigm to implement changes over a period of years. Even as Linus Torvalds harnessed the contributions of a worldwide network of programmers to build Linux by means of the “bazaar” model, he was able to run that network more or less as a singular, benevolent dictator.

The zero-marginal-cost nature of the Internet, however, means that functional bitstreams can be distributed to a much broader audience than before. Moreover, many of the functional bitstreams themselves operate on the Internet, so flaws found in one copy can be exploited instantaneously in thousands or millions of other copies operating on other network switches. Moreover, despite AT&T’s dream of an operating system that would run everywhere, it’s only with the Internet that functional bitstreams have actually had to confront a world’s worth of diversity in hardware.

All of this has meant that while the pre-Internet functional bitstreams could sip attention slowly over time, like the new Internet artists do, functional bitstreams now require fairly intense attention all the time. Take the Heartbleed bug, for instance: In April of this year, Internet users and developers learned that OpenSSL? , one of the key functional bitstreams underlying security on the Internet, had developed a serious technical flaw, putting many people’s private data at risk. How could this have happened? As it turned out, the software foundation committed to maintaining the software had just one full-time employee and took in $2,000 in annual donations. In the networked world, functional bitstreams that fail to maintain a high degree of attention quickly become outmoded or actually dangerous.

Eric S. Raymond, summarizing Linus Torvalds’ development strategy, once wrote, “Given enough eyeballs, all bugs are shallow.” We have begun to see the significance of an unstated corollary: “If it is essential that all bugs remain shallow, then you need a lot of eyeballs.”

Which raises something to ponder: When it was logistically necessary for the artistic industries that the public devote a high degree of attention to a relatively small number of products, those industries had the natural advantage that, from a technical standpoint, the market could only support a small number of products at one time. Today, we need to distribute attention judiciously among functional bitstreams, lest the network fall apart. How will we accomplish that in the Internet society, where the market for functional bitstreams supports so very many projects into which talented individuals can sink their limited attention?

This is a thoughtful and valuable response to Goldhaber. If it is going to use the analytic distinction between functional and non-functional bitstreams, it would benefit from applying also the distinction between producing and distributing digital goods, which is also part of the idea I presented. Because the anarchic, or non-propertarian modes of distribution through sharing are orders of magnitude more efficient than previous forms of physical goods distribution, as you indicate there are forms of communication now possible for small cultural producers to compete effectively against industrialized producers of music, video, and journalism. That doesn't mean these distribution systems are unrelated to the attention economy, however. Some depend on advertising revenue gained by monetizing attention or viewership, or are forms of personal promotion based on attention conscription. (Twitter is a very interesting phenomenon to consider in this respect.)

On the production side, however, it seems to me that the argument would benefit more from some effort to quantify its conclusions. There are naturally bugs in software, and some number of people required to catch them. But the numbers involved are unstated, unestimated, not even speculated about. It turns out, for example out that maintaining OpenSSL (a client of mine) has involved the attention of seven people, two of whom were producing more than 90% of the work. There are now, after Heartbleed, something approaching eight people doing twice as much work overall. They are providing better security for tens of thousands of commercial products (one company, HP, found more than 6,000 of its products and services alone dependent on OpenSSL, for example), and hundreds of millions of websites, still using approximately one billionth of the human population. That's because the provision of software, for the very reasons you suggest, is much more efficient than the distribution of culture.

I think, therefore, that the essay's conclusions depend on assumptions about scale that should be made explicit, and perhaps considered against the available facts.


Webs Webs

r3 - 04 Jan 2015 - 14:46:09 - EbenMoglen
This site is powered by the TWiki collaboration platform.
All material on this collaboration platform is the property of the contributing authors.
All material marked as authored by Eben Moglen is available under the license terms CC-BY-SA version 4.
Syndicate this site RSSATOM