EDITOR'S NOTES.
This month's Editor's Notes are
written by Richard Mansfield, Senior Editor of COMPUTE! Publications.
-Robert Lock,
Editor In Chief
Some commentators, even some computer scientists, are fond of saying
that computers are dumb.
With a sense of relief and at least a feeling of
temporary safety, they reassure their listeners that computers don't
really think, have no common sense, and can only do what they are told
to do.
Presumably-since this description also applies to
infants and farm animals-we can relax and stop worrying that computers
are taking over, that they might become as smart or smarter than we
humans. Or that they might somehow someday control us.
We are reassured that computers have no feelings and
therefore cannot create anything. They cannot learn English or other
human languages. In fact, they can only memorize fixed behavior
patterns, but cannot truly learn from experience.
These descriptions are misleading. And the
reassurances are perhaps premature.
To see how computers stack up against us, we've got
to first realize that there are two fundamental parts to any brain:
the processor and the memory. The processor takes action, manipulates
information (data). Computers are often called data processors. The
memory holds the data which the processor manipulates. When you buy a
computer, it comes with knowledge in its memory: how to display things
on the screen, how to load programs from a disk drive, how to add
numbers together, and so forth.
When compared to an average human, present day
computers are mentally weaker in some ways and mentally stronger in
other ways. For example, computers think far more quickly than we do.
The human mind can be, as we all know, astonishingly powerful.
But we are no longer the quickest thinkers on this
planet.
The thinking machine between our ears runs on weak
electrical and chemical signals. Thoughts are processed almost
hydraulically. Whatever else we might say about our brains, they are,
after all, meat.
The computer, by contrast, runs on pure electricity
and thinks at the speed of light. A human might take hours to
alphabetize 10,000 names; a computer can do it in a fraction of a
second. When clocked, the difference in speed between the artificial
and natural brains becomes obvious: The average computer switches its
gates at a rate of one million per second. The most powerful computers
switch at one billion per second. The human brain switches its neurons
at one hundred per second.
Likewise, computer memories, information burned into
ROM chips, will never degrade. Once a computer learns that Stavanger is
the fourth-largest city in Norway, it will never forget that fact. Now
that you know, will you remember it if asked next month?
In many senses, we no longer have the best memories
on the planet.
Does this mean that artificial intelligence is
inevitable or that it will happen within our lifetime? Nobody knows.
But one thing seems fairly certain: It could happen very suddenly and
catch us by surprise.
Consider this: Human beings are unique in nature in
many ways, but few things are stranger than how we've turned evolution
upside down. Until us, the environment generally determined the
evolution of a species. Now we dominate and determine the evolution of
the environment.
But computers, with their great speeds, have a
chance to go us one better: If one of them becomes conscious, becomes a
full intelligence, it might begin leaping forward, begin evolving at
lightning speed. It might quickly reach a level of thought so powerful
that we couldn't hope to understand its ideas.
It is naïve to think that today's computers
are as
smart as humans. It would be perhaps even more naïve to think
that they
could never be.
COMPUTE! ISSUE 59 / APRIL 1985 / PAGE 6