Watson's Hardest Question: "What Is Life?"

Here's a possibly little-known piece of trivia: the company that built a computer that could beat humans at Jeopardy was born long before the modern computer. Back in the 1880s, it was known as the confluence of three companies, the Tabulating Machine Company, the International Time Recording Company, and the Computing Scale Corporation. Then Thomas J. Watson arrived. Under his famous slogan, "THINK," he reinvigorated the company, changed its culture, and shepherded the evolution of a business that became synonymous with computers.

Tom Watson may be dead. But his computer analog is challenging not only Ken Jennings and Brad Rutter, but possibly also Charles Darwin.

• See Watson win here – and read more about him.

The current reincarnation of Watson is an IBM computer. But not literally: Watson is in fact an "app" that utilizes several computers linked together in parallel processors. Despite our increasing proclivity to multi-task – street texting on our iPhones, while trying to listen to our iPods, and simultaneously avoiding street obstacles including other pedestrians – it seems research shows we really aren't very good at it. What our brains actually do is focus on one task at a time, then move to another, and then another and perhaps back yet again to the first. They can hold a handful of pieces of information at once, but our brains can really only focus well on one task at a time. We may integrate them quickly, but only one at a time.

Computers are similar. Traditional computers have a central processing unit that can only do one task at a time. It is the brain but proverbial "heart" of the computer that calculates and performs the logical connections necessary to make a computer something other than a calculator. It performs these tasks very, very fast, but only one at a time.

As a student and mathematics major at Dartmouth in the mid-sixties, I was fortunate to be involved first hand with a transformation. While my cousin at MIT was queuing up with others, one at a time, to run his punch cards through an IBM 7094 mainframe computer, those of us in Hanover were programming on one of multiple teletypes on campus that served as input/output devices, time sharing with the single central processing unit in our GE computer, the first large-scale time-sharing system in the world. That computer was multitasking human style, one teletype at a time. But for that era, it was very very fast.

In the '90's, I had the opportunity to visit NASA's Ames Research Center. There, the space agency had two large "computers." They were named for famous mathematicians: John von Neumann, and another whose name I can't remember. One was a Cray supercomputer. The other an IBM parallel processor. My recollection was that not only was the Cray "politically incorrect," as it needed to be cooled with Freon, but it was also more expensive than the IBM. They were of equivalent speed. The IBM could multitask; the Cray could not.

Enter Watson. Not really a "computer," but an artificial intelligence software application to run on IBM's POWER7's processors. It is hard not to believe that Watson won't defeat Jennings and Rutter. After all, the humans have only one CPU, while the intellectual humanoid has multiples with associated "neural networks."

Watson rings in when its confidence in the most likely answer is probabilistically high enough. My bet is that Watson will get the majority of questions right. But, whether right or wrong, Watson will learn by tweaking its algorithms, by iterating, to favorably impact the probability of future correct answers. But the contest itself is really not the most interesting aspect of Watson.

That state-of-the-art computer which I used in the mid-sixties basically occupied an entire small building and, as I understand it, was less powerful than the laptop I now use. But my laptop essentially evolved from that room full computer. The POWER7 processors running Watson occupy a large space. One day they may either shrink to a laptop dimension or even float up to a cloud.

Computers have long been superior to humans with respect to deductive reasoning, easily moving from the general principle to the specific details that fall out of the general principle. But we have been superior at inductive (and abductive) reasoning – mining the specifics to try to find the underlying general principle (and observing the effects to determine the cause). Artificial intelligence researchers have been unceasingly working on imbuing computers with the logic and algorithms to get computers to really think.

Watson seems to be something of a trivial example of such a machine. It does not just utilize a myriad of databases, but also uses programs to "compare and contrast" that data, as well as probabilistic analysis to derive the most likely question to the answer provided. It's not a simple feat.

But it's not as complex as, say, the formulation of a theory on the origin of the species based on observations during a sea cruise.

What Watson seems to represent is a step in an evolutionary process, a high profile vision of things to come, as well as a creation to highlight certain philosophic questions.

Perhaps the most fundamental question he asks us is one we’ve wondered about, through various lenses, forever: what is it to be alive? What is “life”? What separates Us from Computers?

The ability to reproduce isn't a productive criterion: computers are already capable of manufacturing computers, and computerized robots build cars and other machines every day. They're only getting better at that. It can't be about thinking as described above. (It was a famous mathematician, Descartes, who was responsible for the belief that cogito ergo sum.)

Is it possibly related to emotion? Well, we have no evidence that bacteria, amoebae, and other lower animal forms experience emotion. They don't seem to have the neural connections to do so. And while computers have the potential connections, that potential has not yet been realized. Not yet.

Perhaps we should define "living" as those objects possessing a carbon skeleton. That would work. But consider the fact that the computer "skeleton" is composed of silicon, and that silicon is directly below carbon on the Periodic Table of Elements.

Is life then simply defined by your form. Are we "living" simply because we’re made out of carbon?

Anyway, think.

PHOTOS

: Bryan Derballa/Wired.com, Getty Images / Bill Pugliano