This excerpt is particularly interesting:
"...the algorithm generated a bizarre flower-like pattern of holes that no human would have thought of trying."
Science Fiction? Machine that makes a better machine than what humans can create! I am reminded of what William Gibson said during an NPR interview:
"... the future is already here. It's just not very evenly distributed."
Not too far in the near future I believe computations that mimic biological progressions will allow us to build individualized and personalized solutions for our everyday needs. The person with issues of severe dry scalp in Winter will have a custom shampoo synthesized for her in a matter of minutes.
All this is exciting yet I also foresee the complexity that organizations like USFDA are going to face when computers will start to create customized Rx. For example, custom pharmacological cocktails for every person in a matter of hours or minutes - How does one regulate it? See a Wired article on pharma cocktails here.
This inevitability is a lot closer than we realize. Let us indulge for a minute - one of the main issues with customized pharma solutions is the speed of the computing power for analysis, for example, of biological side effects and reactions. Specifically, today's microprocessors run close to 200 Fahrenheit and fail when pushed farther. Diamond wafers are around the corner with diamond microchips running at speeds that would liquefy silicon. Manufactured diamonds are becoming prevalent, read detailed story from Wired here. 4Ghz CPU? 400Ghz may not be far!
Therefore, this is not an indulgence but an opportunity in desperate need of hype, frenzy, marketing and Davos panels for the preeminent entrepreneurial workshops of the World like Silicon Valley and Israeli invention machinery to go after.
So far I have touched on formulated products. Perhaps insights can be gleaned from the discrete manufacturing industry like the automotive industry. Base car platforms are being personalized, such as the Scion, that have started to democratize the "made to order" car previously the domain of brands like Ferrari.
Yet, as I touched on Patents previously in (algorithmic) Evolutionary Innovation - Pt. I, the bigger concern will be with Intellectual Property. The traditional and current patent office could be considered a detriment to the future. Why? Because in its current state, it will have to grant every permutation of the algorithmically evolved invention a patent, every algorithm in itself will be a patent and then the methods of dispensation of the invention will be a patent also.
Here is an article with an example to validate my concerns above. This article relates tough questions faced by the USPTO today due to the great excitement brought about when and since the first "business methods" patent was awarded in 1998.
In case of algorithmically evolved inventions, a solution could be to take the bold step of declaring the duplication of Nature as Open Source, while the effort of deciphering Nature licensable.
Unfortunately, governments through history have not been able to seize the possibility and design the future. Mostly it is that the future seizes them and the opportunity is lost to the chagrin of the knowledge creators, who then have to find new horizons to practice their art and science.
By the way, I tried in vain for fifteen minutes to search for a simple statement describing the purpose of the US Patent & Trademark Office i.e. Who do they serve? What do they do? And how does that come about? This continues to be an area of improvement for USPTO.
Caring and nurturing of the seeker, the discoverer and the inventor through the USPTO has brought great success to the USA in the past two centuries. It must take the lead here to enable us to continue to remain at the head of the World in knowledge creation and the ability to leverage it for growth.