But How Do It Know? - the Basic Principles of Computers for Everyone (32 page)
Read But How Do It Know? - the Basic Principles of Computers for Everyone Online
Authors: J Clark Scott
The point is that computers have a method of dealing with pictures based on the principles on which computers work. Using these principles alone has not yet yielded computers or software that can recognize a face with anywhere near the speed and accuracy of any ordinary person.
Voice recognition by computers is another technology that has come a long way, but has much further to go to rival what the mind does easily.
So in comparing a computer to a brain, it just doesn’t look very likely that they operate on the same principles. The brain is very slow, there isn’t any place to get the software to run it, and we don’t see the types of problems we would expect with computer software errors.
In comparing a computer to the mind, the computer is vastly better at math, but the mind is better at dealing with faces and voices, and can contemplate the entirety of some entity that it has previously experienced.
Science fiction books and movies are full of machines that read minds or implant ideas into them, space ships with built-in talking computers and lifelike robots and androids. These machines have varying capabilities and some of the plots deal with the robot wrestling with consciousness, self-realization, emotions, etc. These machines seem to feel less than complete because they are just machines, and want desperately to become fully human. It’s sort of a grown-up version of the children’s classic “Pinocchio,” the story about a marionette who wants to become a real boy.
But would it be possible to build such machines with a vastly expanded version of the technology that we used to build our simple computer?
Optimism is a great thing, and it should not be squashed, but a problem will not be susceptible to solution if you are using a methodology or technology that doesn’t measure up to that problem. In the field of medicine, some diseases have been wiped out by antibiotics, others can be prevented by inoculations, but others still plague humanity despite the best of care and decades of research. And let’s not even look into subjects like politics. Maybe more time is all that’s needed, but you also have to look at the possibility that these problems either are unsolvable, or that the research has been looking in the wrong places for the answer.
As an example, many visions of the future have included people traveling around in flying cars. Actually, several types of flying cars have been built. But they are expensive, inefficient, noisy and very dangerous. They work on the same basic principles as helicopters. If two flying cars have any sort of a minor accident, everyone will die when both cars crash to the Earth. So today’s aviation technology just won’t result in a satisfactory flying car. Unless and until someone invents a cheap and reliable anti-gravity device, there will not be a mass market for flying cars and traffic on the roads will not be relieved.
If you want to build a machine that works just like a person, certainly the best way to do it would be to find out how the person works and then build a machine that works on the same principles, has parts that do the same things, and is wired up in the same way as a person.
When Thomas Edison invented the phonograph, he was dealing with the subject of sound. Sound is a vibration of the air. So he invented an apparatus that captured the vibrations in the air and transformed them into a vibrating groove on the surface of a wax cylinder. The sound could then be recreated by transferring the vibrations in the groove back into the air. The point is, that in order to recreate sound, he found out how sound worked, and then made a machine that worked on the same principle. Sound is a vibration, the groove in a phonograph is a vibration.
A lot of research has been done on the subject of what makes people tick. A lot of research has been done on the subject of how to make computers do the things that people do. A lot of things have been discovered and a lot of things have been invented. I do not want to minimize any of the work done, or results achieved in these areas.
But there are many things that have not yet been discovered or invented.
Many dead brains have been dissected and their parts have been studied and classified. The brain does contain nerve cells which move electricity from one place to another. This is a similarity between brains and computers. But research into the actual operation of living human brains is necessarily limited. Most observations have been made during surgeries that were necessitated by accident or disease. Many observations have been made of changes to behavior after an injury or disease has disabled certain parts of the brain. From this research, it has been possible to associate certain functions with certain areas of the brain.
But no one has discovered a bus, a clock, any registers, an ALU or RAM. The exact mechanism of memory in the brain remains a mystery. It has been shown that nerves grow new connections over time, and it is assumed that this is the mechanism of learning, but no one has been able to say that this particular nerve does this exact function, as we can do with the individual wires in a computer.
Everything that goes into a computer gets turned into one code or another. The keyboard generates one byte of ASCII per keystroke, a microphone generates 44,100 binary numbers per second, a color camera generates three binary numbers per pixel, 30 times a second, and so on. No one has isolated the use of any codes like ASCII, binary numbers, fonts or an instruction code in the brain. They may be there, but they have not been isolated. No one has traced a thought or located a memory in the same way that we could follow the operation of a program in a computer.
It is widely assumed that the brain works in some much more spread out way than a single computer, that there are thousands or billions of computer elements that cooperate and share the work. But such elements have not yet been located. In the world of computing, this idea is called ‘parallel processing’ and computers with dozens or hundreds of CPUs have been built. But these computers still haven’t resulted in a human substitute.
Think of it all as a puzzle. How people work is one side of the puzzle. Making computers do things that people do is the other side of the puzzle. Pieces of the puzzle are being assembled on both sides. The problem is that as progress is being made on both sides, it looks more and more like these are two different puzzles, they are not coming together in the middle. They are not converging into a single picture.
The researchers are very aware of these developments. But when it comes to pop culture, people hear about new inventions all the time, and see the future portrayed in science fiction films, and the logical conclusion seems to be that research will continue to solve the problems one by one until in 10 or 20 or 30 years we will have our electro-mechanical friends. In the past century we conquered electricity, flight, space travel, chemistry, nuclear energy, etc. So why not the brain and/or mind? The research, however, is still at the stage where every time one new answer is found, it creates more than one more new question.
So it appears that whichever way we look at it, neither the brain nor the mind work on the same principles as computers as we know them. I say ‘as we know them’ because some other type of computer may be invented in the future. But all of the computers we have today come under the definition of ‘Stored Program Digital Computers,’ and all of the principles on which they operate have been presented in this book.
Still, none of this ‘proves’ that a synthetic human could never be built, it only means that the computer principles as presented in this book are not sufficient for the job. Some completely different type of device that operates on some completely different set of principles might be able to do it. But we can’t comment on such a device until someone invents one.
Going back to a simpler question, do you remember Joe and the Thermos bottle? He thought that the Thermos had some kind of a temperature sensor, and a heater and cooler inside. But even if it had had all of that machinery in it, it still wouldn’t “know” what to do, it would just be a mechanical device that turned on the heater or cooler depending on the temperature of the beverage placed in it.
A pair of scissors is a device that performs a function when made to do so. You put a finger and thumb in the holes and squeeze. The blades at the other end of the scissors move together and cut some paper or cloth or whatever it is that you have placed in their way. Do the scissors “know” how to cut shapes out of paper or how to make a dress out of cloth? Of course not, they just do what they’re told.
Similarly, NAND gates don’t “know” what they are doing, they just react to the electricity or lack of it placed on their inputs. If one gate doesn’t know anything, then it doesn’t matter how many of them you connect together, if one of them knows absolutely zero, a million of them will also know zero.
We use a lot of words that give human characteristics to our computers. We say that it “knows” things. We say it “remembers” things. We say that it “sees,” and “understands.” Even something as simple as a device adapter “listens” for its address to appear on the I/O bus, or a jump instruction “decides” what to do. There is nothing wrong with this as long as we know the truth of the matter.
Now that we know what is in a computer, and how it works, I think it is fairly obvious that the answer to the question “But How do it Know?” is simply “It doesn’t know anything!”
Table of Contents
The Other Half of the Computer
Doing Something Useful, Revisited
The Arithmetic or Logic Instruction
The Load and Store Instructions