Saturday, February 12, 2011

Patents and progress in the method of discovery

I am often shocked by the rapid advances of scientific techniques, and how major accomplishments from a decade ago can be repeated with a tiny fraction of the effort. Recent advances in biology -- particularly with DNA sequencing and synthesis - may be exceptional, but the basic principle surely holds across all sciences: advances in the methodology of discovery will increase the ease with which future discoveries are made.

Through the recent past, the most general advances are probably related to computing -- the storage, transmission, and analysis of data. These advances are continuing, along with advances in artificial intelligence and robotics. Regardless of whether these advances will create a technological "singularity", they will surely contribute to additional discoveries in all fields of science.

Recently, a "robotic scientist" has even been built. Specifically, this "scientist" is a geneticist, with a very limited toolkit (both in design and implementation of experiments). However, as a proof of concept, it does raise the specter that robotic systems will soon be able to make discoveries that traditionally required a lot of time from highly trained scientists. Perhaps over the course of a decade, we will see 100-fold reductions in the price of discovery, across many fields of research.

If we put this rate of advancement in the context of patent law, we see a startling incongruity. Patents grant a 20 year monopoly to the inventor of a new tool or process, supposedly providing incentive for research and development. This monopoly allows the inventor to capture all of the benefits of the new invention for himself -- which makes like more costly for everyone else, but seems fair as long as we assume that it was necessary to produce the invention in the first place. While this assumption has always had some problems, it seems extremely shaky in light of the current rate of progress.

No comments: