But How Do It Know? - the Basic Principles of Computers for Everyone (31 page)

With most of the computers that actually get built, while the individual bytes in RAM remain 8 bits, everything else is expanded to 16 bits, 32 bits or 64 bits or a combination of these in different parts of the machine.

Our RAM only has 256 bytes, which is ridiculously small, but that’s all you can have with an eight-bit Memory Address Register. If you use 16 bits, you can have 65,536 bytes of RAM (that’s 64kb), if you use 24 bits you can have 16mb, if you use 32 bits you can have 4 gigabytes of RAM.

Real computers have things that this one does not, but they are not capable of doing things that this computer cannot do.

In our computer, if you want to shift a byte three bits to the left, you would put three shift left instructions in your program. In most real computers, they have shifters that will shift any number of bits in one instruction. But the result is the same, your byte ends up looking the same in either case, the real computer just gets the job done faster.

In our computer, the adder can add two eight-bit numbers. If you want to add 16 bit numbers, you have to employ some software to do it. In most computers, the adder can add 16 or 32 bit numbers in one instruction. Again, the results are the same, one is just faster than the other.

The stepper in our computer is a simplification of something that most computers have, called a ‘state machine.’ State machines provide steps, but start the next instruction as soon as possible, do what is necessary for an interrupt system, can create more complex instructions, etc. Since all we needed was six consecutive steps, we built a simpler thing and just made up the term ‘stepper.’

So yes, our computer is a simple, small, relatively slow computer, but it can do everything that more complicated machines can do. The things that make a bigger machine bigger, are designed to get the job done faster, do it in fewer clock cycles, do the same task with fewer instructions, operate on several bytes at the same time. But the nature of what the machines do is exactly the same. Every task they can do comes down to shifting, ANDing, ORing, XORing, ADDing and NOTing bytes. There are no other fancy types of operations that have been left out of this book.

In a bigger machine, you can do addition, subtraction, multiplication and division in a single instruction. That is because they have huge numbers of gates arranged into things like a ‘hardware multiplier.’ There is no reason to show you the details of how you construct one of these, it is a very complicated job for the few people who need to build one. It is understandable, and it all ultimately comes down to NAND gates just like everything else. But we have seen how to do all the math operations there are with just an adder, shifter, NOT gates and some software. The hardware multiplier gets there faster, but the results are exactly the same.

Bigger machines have more registers, the registers are each multiple bytes, they have adders that can add three numbers at the same time, but still the instructions come down to the same simple operations. Your understanding of computers is not small because we have looked at a small computer.

Philosophy

Why do we have a chapter called “Philosophy” in a book about computers? The only thing in this book that even comes close to being a philosophical question is its title, “But How do it Know?” We will attempt to answer this question a little later on.

This book has been about the computers that we have today. But what about the future? As computers and software continue to advance, how soon if ever, will the day come when there are walking talking computerized robots that look and act just like people? Will the day come when we have to decide whether or not to give these robots the same legal rights as people? Will computers eventually take over the world and replace people altogether?

To answer these sorts of questions, people often refer to a major question that has been outstanding in the field of philosophy for many years.

The question is, whether man is composed solely of the structural body that we can see and dissect, or whether there is an integral spiritual component to every human being which accounts for the qualities of consciousness, love, honor, happiness, pain, etc.

That question is far beyond the scope of this book, and it remains unconvincingly answered despite many books arguing each viewpoint. There are people in the sciences who say that we are on track to building conscious computers, and it will happen. There are people in the humanities who say that it is impossible because you can’t manufacture a spirit. Each side has been unable to sway the other.

If we define the brain as that funny looking chunk of gray meat enclosed by the skull, and define the mind as whatever it is that is responsible for consciousness, memory, creativity, thinking, and everything else that we notice going on in our heads, then we can restate the big philosophic question as: “Are the brain and the mind one and the same thing?”

Then when it comes to the question about building a convincing human robot, there would be two possibilities.

If the brain and the mind are the same thing, you might not be able to build a synthetic person today, but as time went on, eventually you could understand every structure and function in the brain, and build something of equal complexity that would generate true consciousness, and that really should act just like any other person.

If the brain and the mind are not the same thing, then building a robot buddy will always be about simulating humanity, not building something of equal quality and value.

Restating the question doesn’t make it any easier to answer, but this idea of separating what we know about minds from what we know about brains may be useful. Early on, we said that we were going to show how computers work so that we could see what they were capable of doing, and also what they were not capable of doing. We are going to take what we know about brains and what we know about minds and compare each individually to our new knowledge about computers. In doing so we can look for differences and similarities, and we may be able to answer a few less controversial questions.

Computers do certain things with great ease, such as adding up columns of numbers. A computer can do millions of additions in a single second. The mind can barely remember two numbers at the same time, never mind adding them up without a pencil and paper.

The mind seems to have the ability to look at and consider relatively large amounts of data at the same time. When I think of my favorite cat, I can re-experience seeing what he looks like, hearing the sounds of his purring and mewing, feeling the softness of his fur and his weight when picked up. These are some of the ways that I know my pet.

What would it mean for our computer to think about a cat? It could have pictures of the cat and sounds of the cat encoded in files on a spinning disk or in RAM. Is that thinking? If you ran the bytes of these files one by one through the ALU, would that be thinking? If you put the picture on the screen, would that be thinking? If you played the sounds to the speakers, would that be thinking?

The sounds and pictures encoded in the computer are just byte patterns sitting where they are. They don’t look like anything or sound like anything unless they are sent to the peripherals for which they were designed. And if they are sent to the screen and speakers, the computer doesn’t see them or hear them. Of course, your computer could have a camera pointing at the screen, and a microphone listening to the sounds, but the computer still wouldn’t see a picture or hear a sound, it would just collect more strings of bytes very similar to the ones sent to the screen and speakers in the first place.

There could be programs that perform mathematical operations on the picture files in order to discover patterns, and store the results of these calculations in other files. There could be files that relate one picture file to other similar picture files, and pictures to sounds, etc., creating more files.

But no matter how much programming is applied to the picture files, there is something that the mind can do that the computer simply doesn’t have any facility for.

The mind can consider the whole of some thing all at the same time. You can think of the whole of the cat all at once. Its sort of like the difference between the movie film and the TV screen. The movie film has whole pictures, the TV screen only has one pixel at a time. You could say that your mind works so quickly that you don’t notice the details, it gets integrated into a whole just like the pixels get integrated into an entire picture. But what does the integrating? And when it’s integrated, what is it and where is it? And what looks at the integrated whole?

We’ve just seen everything that’s in a computer. The computer moves one byte at a time over the bus. The fanciest thing it does is to add two bytes into one. Everything else it ‘does’ amounts to nothing more than the simple warehousing of bytes. A stored byte doesn’t do anything beyond maintaining its own current setting. A computer just doesn’t have any facilities that integrate the elements of a picture into anything else, nowhere to store that something else, and nothing with which to look at it.

I’m not saying that something couldn’t be built that would perform these functions, I’m just saying that computers as we know them today don’t currently include any such device.

Here is another question. If a brain works like a computer, then it needs to have a program for the CPU to run. Where would this program come from?

Although the brain has trillions of cells, the entire human body starts with one fertilized egg cell. So any program that the brain has, would have to be present in this single cell, presumably in the DNA.

Scientists have now decoded the entire DNA sequence of humans. DNA is interesting in that it is a long string of only four types of things. It’s digital! A lot of the pieces of this string are used for making chemical reactions take place to make proteins, etc. but the majority of it is called ‘junk DNA’ because no one knows what its purpose is. But even if you consider that the entirety of the DNA is devoted to computer software, then there could be about a billion instructions in this program. Now that’s a lot of software, but the average home computer probably has that much software loaded onto it’s hard drive, and that wouldn’t be anywhere near enough to run a human being.

Some have said that the human computer programs itself. As a programmer myself, I just can’t imagine how this would work. While it’s true that a program can accumulate data and modify the way it works based on the collected data, this is not the same thing as writing a new program. If someone ever writes this program that can write any new needed program, there will be a huge number of computer programmers put out of work forever.

Then there are the kinds of errors that computers make versus the kind that people make. If a computer gets stuck in a loop, it appears to have stopped completely. Have you ever seen a person walking down the street suddenly stop working? All functions just cease. The person would just fall down until somehow his computer re-booted. People do collapse from time to time, but it is usually because some other part broke, like having a heart attack, and you can see the person recognize the pain as it takes them down. But if the human computer got stuck in a loop, there would be an instant loss of consciousness and the body would just fall completely limp with no struggle. I have never seen that, but if the brain operated just like a computer, you would expect to see it on a fairly regular basis.

Then there is the matter of speed. As we have seen, a simple computer can do a billion things in a second. When it comes to the brain, it has nerves that have some similarity to the wires in computers. Nerves can also carry electricity from place to place. In a computer, wires come out of gates and go into other gates. In the brain, nerves are connected together by “synapses.” These synapses are spaces between nerves where the electricity in one nerve creates a chemical reaction, which then causes the next nerve to create its own electricity. These chemical reactions are painfully slow.

No one has shown that these nerves are connected up anything like the wires in a computer, but their lack of speed makes it very unlikely that it would do much good even if the connections were similar. After the electricity travels quickly through the nerve cell, it reaches the synapse, where the chemical reaction takes about one five hundredth of a second to complete. That means that our simple computer built out of NAND gates could do two million things in the same time that only one thing could be done by a computer built out of nerves and synapses.

Another area where the difference between the mind and computers is quite obvious, is in the area of recognizing faces. The mind is very good at it. If you walk into a party with fifty people present, you will know in a matter of seconds whether you are among a group of friends or of strangers. A lot of research has been done into how people accomplish this feat, and a lot of very interesting information has been uncovered. There is also a lot of speculation, and there are many fascinating theories about the underlying principles and mechanisms. But the complete and exact structures and functions have not been uncovered.

If you give a computer a picture file of a person, and then give it the same file again, it can compare the two files byte by byte and see that each byte in one file is exactly equal to the corresponding byte in the other file. But if you give the computer two pictures of the same person that were taken at different times, or from different angles, or with different lighting, or at different ages, then the bytes of the two files will not match up byte by byte. For the computer to determine that these two files represent the same person is a huge task. It has to run very complex programs that perform advanced mathematical functions on the files to find patterns in them, then figure out what those patterns might look like from different angles, then compare those things to every other face it has ever stored on its disk, pick the closest match, then determine if it’s close enough to be the person or just someone that looks similar.