Friday, December 2, 2011

Apple v Samsung

The Apple v. Samsung patent lawsuits worldwide have captured Silicon Valley mindshare. The latest issue is secrecy: how can  U.S. District Judge Lucy Koh balance the rights of the companies to maintain trade secrets with the need for court hearings to be open so that the public understands and consents to the courts' findings.

The really big issue, though, is the patent system. In a recent talk, John Naughton, Professor of the Public Understanding of Technology at the Open University, develops ideas on how technologies evolve, the increased proliferation of new technologies, and factors that dictate the success of a new technology.

Listening to Naughton, my first reaction was to ask whether today's patent system is proper.

Naughton argues that new technology mostly evolves from combining existing technology in new ways. With each new technology, the combinational possibilites increase. Thus, especially in fields like software and electronics, the rate of new technology proliferation skyrockets. This creates two important intellectual property quesitons:

  1. What incremental combination of technology deserves a patent? Every new combination of technology, or combinations that are "significantly" novel (whatever significantly means)?
  2. If the increment of combinational difference required to qualify for a patent is small, then the number of patent applications and, presumably, patents will proliferate exponentially.
In theory, patent applications require more than just a novel technology for approval. They also require a novel application. So, combining technology randomly isn't enough to thwart the current patent system. The new technology combination has to do something useful. Which leads to ...

My second reaction was to wonder whether generating valuable patents is a reproducible process that doesn't require invention, that ah-hah! moment of insight.

In his talk, Naughton also argues that consumer acceptance drives the success of a new technology more than the intrinsic value of the new technology. The classic VHS versus Betamax example is a good example. Betamax was measurably the superior technology, but consumer adoption favored VHS, and then VHS economies of scale killed Betamax technology innovation.

With this in mind, it becomes feasible to develop a patent strategy that has nothing to do with the intent of patents (to protect inventors against copies of their novel ideas) and everything to do with speculation. Assuming that consumers are technology kingmakers, a reasonably astute technologist could guess at 10 or 20 possible directions consumer demand could lead. For example,
  • Take basic human desires like eating, listening to music, or playing games.
  • Think of 10-20 different ways today's products could evolve for each of those desires.
  • From a catalog of existing technologies, combine technologies to create new products that do things like eating, listening to music, or playing games in novel ways.
  • Apply for patents on the all these "inventions." If money is a constraint, rank by factors such as market size and estimated product price. The cost of a patent applications could be lowered through repetition, and the time for patent approvals is increasing so that the costs of final patent arguments and issuance are further out when there is more certainty about the value of pursuing a particular patent.
It's not hard to imagine software that could replicate this algorithm someday. What if someone patented this algorithm? What if the algorithm worked and generated valuable patents that forced companies to license or buy the intellectual property? It's possible this system would miss some ah-hahs!, but it's also possible that it would extract a lot of money from the current patent system because it focuses on what consumers are likely to want in the future incrementally rather than non-linearly.

But then, what is an incremental versus a non-linear innovation? In market terms, incremental innovations may cause a non-linear response before a truly non-linear innovation can. For instance, in the case of silicon versus gallium arsenide and other integrated circuit substrate candidates, silicon keeps winning in the market with incremental innovations on top of a huge silicon infrastructure investments. The non-linear advantages of gallium arsenide should have won (like Betamax should have won), but can't beat silicon's infrastructure advantages.

Is the current patent system working? No, and it's creating ancillary problems like too many sealed filings in Apple v Samsung that undermine the courts' legitimacy. It's also likely that both Apple and Samsung have innovated in the smartphone and tablet markets in ways that both unwittingly benefit from. A new intellectual property system has to reconsider what inventions deserves protection (ownership), and what an individual or corporate inventor is being protected against (ownership rights).