When you sit down and think about it, the speed at which computers do things boggles the mind. It may not seem like it while you’re waiting for that big file to save, but computers operate extremely quickly. The problem is that while they can do millions of tasks a second, each task is very, very simple.
The big difference between computers and humans
Oh, I should say that there are a lot of differences between computers and humans. And we can argue about what the most important one is. You could say that computers are incapable of emotion, or that they lack a soul, or that they can’t spontaneously decide something on a whim. All of those things are true, I suppose. But I think the part of the whole comparison that matters right now is the way computers process information as opposed to the way that we do.
Human information processing
We don’t really understand a lot about how humans process information. What we do know is that humans can work with large chunks of data as if they were single, monolithic things. What I mean is, you can look outside and say, “yes that is outside.” You can look closer and see a house. You can look at that house and see a person. Each of these things: outside, house, person, are single things in your mind. Your mind does not need to understand the number of beams or studs in a house, or understand the basic workings of someone else’s digestive system. You look at a person and you say “person.” You may even recognize the person.
Computer information processing
Computers, at their hearts, understand only ones and zeroes. That means that everything needs to be broken down to its component parts in order to be understood. Today’s computers can deal with strings of ones and zeroes very quickly and most use some form of parallel processing so that more than one string of ones and zeroes can be dealt with at one time. But it’s still a very simple process of counting ones and zeroes and comparing them to the codes stored in the computer in order to do the next thing.
So, these extremely fast streams of data go in, and the computer reacts in a predictable way. It’s impossible for our human brains to even think about the speed at which that data flows, but we can understand that it’s such simple data that it takes a lot of it to make something happen.
Yes, even in computers with “neural networks.”
We’re starting to see a lot of devices, large and small, with some sort of ability to recognize things. At their most basic level, sensors can tell you if it’s dark or light, if your door is open or closed. But there’s more than that. Chances are your camera can find the faces in a picture and adjust exposure so they look good. Your car may be able to sense other cars well enough to warn you to avoid them. There’s also the sophisticated software that governments use to find specific people in security footage. That stuff’s kind of creepy.
You’d be tempted to think that this kind of thing is proof that computers “understand” real life the way that we do. It isn’t. It’s proof that there’s a lot of sophisticated simulation going on. The computer in your car may be able to see other cars, in a manner of speaking, but it doesn’t really understand that a car is a machine and the people in it aren’t. It can’t even really understand the concept of a journey. It can give you directions but it doesn’t understand why you’re going.
Will we really ever get artificial intelligence?
I believe we will. And I believe it won’t be an imitation of human intelligence, as today’s neural networks are. It will be something totally different, and as impossible for us to understand as our own brains. I believe it will happen. The only hope is that when it does, we all find a way to coexist.