Times Online had an article about computers, currently, can’t beat humans at Go. I hate the tone of the article; the subtle arrogance. After Gary Kasparov lost to Deep Blue, it marked an achievement. Not one man over a machine, but a team of men’s ingenious designs against one man. The machine was just an instrument of no escape. It calculated with mathematical precision, each possible move, considered every possibility, until it used the rules that the programmers decided to conclude what move had the best chance of victory. I’m sure another team of people sat down and decided to come up with a program that could defeat Deep Blue, they could.
The point is, that the attempt to defeat one of the greatest chess minds alive with a device, is a societal achievement, not the machine’s. The fact that the machines could toss aside amateurs years before, seems of no consequence. If the achievement was solely for the machine, we would remove silly constraints on time. I mean, a computer could play a single game of Go or chess for 100’s of years, where as I doubt a human would care to. The other fact is $1,000,000 is not a lot of money. That would probably buy, at most, one Ph.D. in Computer Science for 10 years. How long do people learn Go for, before they become competent players? How much do the professional Go players earn collectively?
Invariably, the problem isn’t with computer, and it is certainly not because it lacks
adaptation to uncertainty, intuition, wisdom, the ability to understand the thoughts and feelings of others, and a sense of mortality
Pattern recognition is fundamentally a hard problem. But, our brains are adept at handling it. Unfortunately, our system is also far from perfect, in that we are hard wired to see faces and patterns, which is why we get potato chips that apparently look like Elvis to Jesus. Teaching a computer to recognize this garbage as a face would be deemed unacceptable, however we seem to be perfectly happy with human imperfections.
As another example, there was a time when I was in a 3-D modeling program, and I was looking at a head model. I was inside the model. Something I can usually easily discern. But, the problem was that this was a face. So my brain was flipping it for me, causing me to be utterly confused. Optical illusions are a collection of all types of tricks to get the brain to see things that aren’t really there. Do we blame the computer for not having these flaws?
If we ultimately are going to create an Artificial Intelligence, it would sad to handicap it with all of our imperfections. Nor should we applaud the fact that there are a few things, that by evolution, we are good at doing but unable to explain. Our mechanisms for detecting faces and general pattern recognition took thousands if not millions of years to develop. To fault our ability to design a machine in the last 50 that can do a better job than the most sophisticated parts of the brain is not giving our evolution enough credit and placing too much heed to our collective intellect.
That said, to applaud the inability of a computer to do something is to applaud our ignorance, our impatience, our superstition, and our hubris. It is revel in the Dark Ages of humanity. It is to rail against progress, achievement and the collective human intellect. It is for these reasons I find the undertones of the article so despicable.