In my own lifetime, I’ve seen computer technology really explode. My first computer was a Commodore 16, with no hard drive, and 16k of memory. There was no graphical operating system, and I couldn’t do much more than type at a prompt. The PC came with a monthly magazine subscription that had sample programs listed in the back, and I can remember my dad and me sitting for hours typing in code for a Space-Invaders type of program. It never worked. We tried entering the program three times, but it would crash after about 30 seconds—must have had a bug.
Since that time hard-drive size and processing speeds have increased exponentially—Moore’s law describes a trend in hardware manufacturing where the number of transistors that would fit on an integrated circuit doubled about every 24 months. It also seems a new PC becomes outdated minuets after removing it from the box. However, in spite of all these computing advancements, human language programs have lagged far behind.
Computers can process large amounts of data very quickly—billions, and even trillions of instructions per second. A computer can search an encyclopedia for a phrase like “history of computers”, and return all the results in a list in just a few seconds. There’s just no way a person could do that kind of rote processing—it would take us months or years to do the same thing.
Computers tend to do certain types of tasks efficiently, like searching through a list or adding numbers. But there are certain tasks that are tough for a computer. For example, a person can look at an image of a friend, and within seconds recognize them. It’s a bit more difficult for computers-or at least used to be.
Many forums and blogs on the internet have a CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) defense built in. When reply to someone’s thread or blog, it requires you to look at an image and then type it into a verification box. This is to keep computer programs from posting advertisements and other nasties to the site. This is effective because it is difficult for a program to look at an image and determine what it is.
Another difficult task for a computer is to process language. Human languages are ambiguous—take a look at this sentence: Time flies when you’re having fun, but fruit flies like bananas.
Does fruit fly? How does time take flight? Is “flies” a noun or verb, an action or insect?
Also stressing a word in speech can change the meaning of the sentence. For example:
“I never said anything about you”
– Maybe someone else did, but I didn’t.
“I never said anything about you”
– I never said anything, but may have written something.
Imagine a future without the mouse or keyboard. You get home from work, open the door, and immediately speak to your house computer, “Jake! Please turn on the TV.” It responds in a pleasant voice and tells you it is now set to channel 23 for the evening news. Later on that evening you’re sitting on the couch, “Jake can you message my sister, and invite her to the cookout Saturday?” It replies a few seconds later and asks, “Would you like me to ask her to bring something?” A program that can understand and follow a conversation would be very useful.
It could also become a very personal possession, and be passed on to your children and grandchildren, allowing them to ask questions about your life like “Jake. What was my grandfather’s favorite food?” or “Jake. What was my dad’s first job?”
For now, computers that can communicate intelligently with us lies within the realm of sci-fi, but science fiction often drives scientific discovery.