Jim Bessen — “Is technological innovation on the Web different?” (Oct. 5, 2010)
ø
Jim Bessen gave last week’s “Web Exceptionalism” presentation at the Berkman Fellows Hour. Jim asked, “Is technological innovation on the Web different?”
What follows here is an attempt at recapping a lively, high-level Fellows Hour discussion. To paraphrase [a no-doubt cringing] Wordsworth, liveblogging — as I’ve just come to learn — might best be described as a spontaneous overflow of note-taking recollected in tranquility, and the practice supplies ample opportunity for missing nuance and eliding worthy discussion points, both in the note-taking and in the tranquil recollection. That’s my long way around to an apology for the summation failing to live up to the event. (And sorry, too, for the time lag in getting this posted.) Here goes:
At the outset, Jim makes clear that when talks of “innovation,” he means “technological innovation” and not cultural or production innovation. Web-related innovation is of course a part of that, but his principal focus is on how the Web facilitates innovation more broadly — and whether that facilitation is, per the Fellows Hour theme, “exceptional.”
To be sure, Jim says, the Web enables a whole new level of collaboration (take, for example, open-source software). Digital media and the Internet reduce the cost of communication, with the result that you can cheaply and easily “pool knowledge” all across the globe. The conventional wisdom is that Internet collaboration has triggered a “revolution of innovation.”
Jim isn’t convinced, and he urges us not to buy into the hype, which is in large part predicated on mythologies of pre-digital innovation (i.e., an inventor has a moment of inspiration, followed by several moments of perspiration, obtains a patent, and starts a company). In point of fact, the past was not so different. Pre-digital innovation was fraught with collaboration. Blast furnaces, steel minimills, U.S steamboats, the early PC — all of these significant technological advances were the products of extensive collaboration. And indeed, biological innovation was until very recently an extremely collaborative effort. Jim pauses for a moment over to wonder aloud why this history of pre-digital collaboration is “a hidden history.” “Propaganda,” in support of the cult of the inventor, may have something to do with it.
So the phenomenon of collaboration does not, in Jim’s view, make the Internet exceptional on its own. Another possibility is scalability. Larry Lessig has written of the Internet that entry costs to the technology are very low, such that scalability might be the Net’s exceptionality trait. Jim is not sure this is any different from conditions in the 19th century.
But collaboration nowadays, with Internet technology, is for the first time truly global, isn’t it? Jim counters that researchers have recently unearthed records of substantial communications between French aviation innovators and the Wright brothers. Sure, overseas collaboration can happen more quickly online, but Jim is not sure the rate of exchange is enough to be truly consequential.
Does the Net’s ability to pull in vast numbers of collaborators make it an exceptional engine of innovation? Jim argues that the size of an innovation network is not critical. Technological innovations generally — before and now — emerge from a core network of a handful of collaborators.
Jim suggests a recurring pattern for technological innovation. We see early-stage collaboration, innovators exchanging knowledge until the point of a major breakthrough — the Kitty Hawk, Bessemer steel, the Apple II — triggers a shutdown of collaboration, while a single agent consolidates its gains into market dominance. There are, of course, exceptions — Jim mentions Cornish mining and steam engines.
Jim considers that the Internet may indeed be exceptional in that online collaborative innovations tend to resist market domination/ collaboration failure. Linux is by all accounts a mature OS. The breakthrough moment has come and gone, and large firms are generating the majority of code. Here, however, the firms participate in and benefit from the incremental innovation that large firms are good at, but there’s no domination. Consider as well smartphone OS: Jim observes that contrary to many projections, we are seeing a space created for innovation (in this case, apps) that large companies do not dominate.
Finally, innovation accomplished on the Web — at least in cases where collaboration survives the “major breakthrough” stage — may be distinctive in the level of customization it offers. Open-source software is accessible, and therefore customizable, and Jim posits that Web innovation is more open in the long-term to customization.
To sum up: Web-facilitated innovation is exceptional for its resistance to single-player domination and, relatedly, for its susceptibility to customization.
*** Discussion follows ***
Q: Let’s not forget that the Bessemer steel case differs from Linux on the ground that open-source licenses leveraged IP in the latter case to prevent the innovation from being captured.
Jim: It’s important to note that the Bessemer innovators licensed the process, mastered the technology as others had not, and then they built U.S. Steel. But generally, yes, folks seeking to dominate a market will deploy patents to that end, as they can. And it’s not just IP rights that would-be dominators will leverage into supremacy: often they leverage other assets, as Microsoft hoped to do with Windows and Internet Explorer.
Q: Can we point to instances in which we had significant post-“forking” innovation? Maybe the OS X-Windows case?
Jim: This is one important mechanism for preventing dominance. There’s the more formal mechanism of licensing, too. Apache is a license that could be taken private, but there is so much community innovation around Apache, it resists any single-player dominance.
Q: Isn’t it the case that large-scale participants tended to exert a disproportionate influence on early-stage innovative collaboration?
Jim: In my historical examples, there weren’t big companies involved.
Q (follow-up): But collaboration on the Web — even early-stage collaboration — tends to find itself “dominated” by consistent/insistent participants.
Jim: There was a significant social network phenomenon in the pre-digital era. The International Fraternity of Mechanicians would share technologies within the community but enforce a patent against out-group infringers. Let’s take care to distinguish between what I’m talking about — market dominance — and contribution dominance. How is it that in the open source context, hobbyist contributions have given way to large-firm professional contributions, but there is still no one exerting market dominance? That’s the point of interest for me.
Q: Might we be understating the historical significance of the rise of the corporation, or the rationalization of scientific practices? These are radical changes.
Jim: It’s been shown that involvement of an original scientist in a biotech corporation is directly related to its success.
Q (follow-up): . . . but you can’t innovate now without having a Ph.D. or corporate affiliation.
Jim: Mark Zuckerberg?
Q (follow-up): Point taken.
Jim: Several of the collaboration examples I described predate the modern corporation.
Q (follow-up): Nonetheless, it’s important to remember that the macroeconomy changes right alongside the technology — and it’s not easy to control for that.
Q: And let’s not forget [Clayton] Christensen’s “Innovator’s Dilemma”: the market dominator is unlikely to see the innovations that will disrupt its domination.
Jim: Yes. And large companies are less able to leverage success with one technology into another area.
Q: Sarnoff was able to leverage radio into TV, but Microsoft couldn’t leverage OS supremacy into browser supremacy.
Jim: And indeed, Microsoft’s ability to leverage anything today is much less than in the case of the OS/IE failure fifteen years ago. Windows is just now coming up with a phone OS. The horse has left the barn.
Q: How do the phenomena you describe square with the patent push phenomenon?
Jim: In the past you’d see the heavy waves of patenting late in the life cycle. Firms would only take out patents after they had staked out ground they wanted to dominate. Now the patents are coming first — and in many cases inventors have no intention of marketing the technology.
Q (follow-up): Might aggressive patenting be a rearguard corporate action to use IP to take back some of the ground they’re losing elsewhere?
Jim: A lot of this is orthogonal and has to do with problems in the legal system.
Q: We’re seeing less de facto standardization from big firms. What role will public standards organizations play in web-era innovation?
Jim: I’m not sure I agree with the premise: consider Apple’s iPhone apps standard. But I do agree that public standard bodies have become more influential. This may have to do with the greater modularity that comes from collaboration.
Q: Does the ability to interact quickly and electronically beat down the need to hash out standards?
Jim: The standards bodies still exist and flourish.
Q: HTML won in industry after industry, precisely because it’s a crappy, anything-goes standard.
Q: The Web may require more from us — whenever you want to work at a level where you have to have a common substrate (as you increasingly need to do on the Net) — standards are required.
Q: Maybe one way to see standards is that you dampen innovation in one space to enhance innovation above it.
Q: What happens if the courts retrench on patentability and reject software and business model patents?
Jim: We’ve concluded through research that these patents function as a 10% to 20% tax on innovation. The big payers are the large companies, but every startup that makes any progress has to worry about patents and has had a patent asserted against them. If thrown out, restricted reasonably, you’re alleviating the tax and improving incentives to innovate.
Q: Are we headed in that direction?
Jim: The Supreme Court made a step in the right direction with Bilski. The Federal Circuit has made moves to restrict the availability of the doctrine of equivalents in proving infringement. There’s an awareness now that things are screwed up. The train may be turning around. A number of patent law doctrines and Patent and Trademark Office practices affect the issuance of software patents. They’ve all changed over time. Five years ago, pretty much any software was patentable, but now it’s getting more difficult. Patent litigation has tripled from the early 1990s. It’s finally leveling off, but the litigation rate for software patents continues to increase. Common problems include overlooked prior art, vaguely-stated claims subject to broad interpretation in patentee-friendly fora like the Eastern District of Texas. Although courts may take baby steps in the right directions, the situation likely will continue to deteriorate until a crisis hits the large IT firms and they become politically active.
Q: What role does copyright play in locking down software?
Jim: This is not an issue with software innovations.
Q: It is possible to infringe a copyrighted header file if you use it, but typically you can rewrite the header file in a manner that does not, strictly speaking, infringe the copyright (although it will infringe a patent). And of course, some expression is not copyrightable. If there’s only one way to write code to accomplish a result, then you run the risk of infringing copyright, but in that case copyright’s “merger doctrine” would likely preclude infringement liability.
Q: Trafficking in technologies of subversion — hacking DSS, jailbreaking iPhones — is easily accomplished online. That’s a consequential aspect of Web exceptionalism.
Q: But of course, that same quick-and-easy distribution has empowered the big-firm objects of subversion in the first place.
Q: The sort of online subversion we see tends to be directed at attacking models of rent extraction.
Jim: to sum up, I see the following aspects of the Web as “exceptional” promoters of innovation: (1) faster diffusion of information, (2) lower cost of entry (possibly), (3) a culture of civil disobedience, (4) the relative inability to leverage dominance from one area into another, (5) greater customizability, and (6) the ability of communities to create their own defensive perimeter to allow collaboration without subjecting a technology to control (i.e., through GPL). That last bit might be “legal exceptionalism.”
