Will big computers outthink big brains?
Published: June 2, 2009
The first electronic computer was tested in 1942. Since that time, computers have changed a great deal, but my own brain has not changed much. While computers have gotten faster and more powerful, I'm still using the first brain installed in my skull.
Sometime soon, a computer will be able to simulate my brain. Later, there will be a computer that can literally read my mind.
We will soon live in a world in which the computing machines are more intellectually powerful than all the human brains on the planet.
It is fun to track the growth in the power of big computers. The Web site Top500.org does this, and posts its findings every six months. The next results are due this month. Last November, its rankings showed "Roadrunner" slightly ahead of "Jaguar."
Roadrunner is an IBM-made computer delivered on 21 tractor-trailer trucks to the Los Alamos National Laboratory. Jaguar, made by Cray Research, is at the Oak Ridge National Laboratory.
Roadrunner does calculations at 1.1 petaflops. A "flop" is a floating point (decimal number) calculation. "Flops" are the number of floating point calculations per second that a computer can perform. "Peta" is the abbreviation for 1,000 trillion.
These truly fast computers can perform 1,100 trillion calculations per second. They get faster every year.
The speed of the Intel chips used in many personal computers has increased approximately 10-fold every four years -- under 1 million instructions per second in the 1970s to more than 100 billion instructions per second today.
Comparing the power of a computer with the power of a brain, computers still lag. Some estimate that it will be another seven years before a computer can hope to simulate a human brain. Not until about 2025-2030 can we expect that the largest, fastest computer will be able to think human thoughts with information read from a human brain.
We should also expect the kinds of ethical and moral challenges that arose with nuclear weapons and stem-cell research. These technologies outpaced the capacity of human culture to cope with their moral challenges.
Science-fiction writer Isaac Asimov often dealt with these ethical problems in his stories about robots. His story "Runaround" is famous for introducing his three laws of robotics:
- A robot may not injure a human being or, through inaction, allow a human being to come to harm.
- A robot must obey orders given to it by human beings, except where such orders would conflict with the First Law.
- A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
The question today is, "How do we incorporate safeguards into the computers we are building?" In a world in which the minds of the machines are far more capable than ours, will machines find humans irrelevant?
Virginia's computer technology Standards of Learning contain several sections on the serious moral and ethical issues of computer use, beginning in third grade (CT 3-5.3).
Walter R.T. Witschey is professor of anthropology and science education at Longwood University.
Advertisement
Reader Reactions
To the poster above:
What you’re missing is that computer technology both grows more powerful and shrinks in size exponentially, essentially doubling in power and shrinking in half every 1.5-2 years (following Moore’s law).
Given that your number is accurate, this would mean that computers would equal the size and power of the human brain in approximately 25.5 years, and then a couple of years later, be half as large and twice as fast.
That would would make the estimate in the article only slightly optimistic, but I think your estimate of 1/100000 may be off a smidge.
No, I think it’s safe to say we’ll be seeing computers capable of simulating the human brain within the next 20 years, and 20 years after that, computers that make us look less intelligent by comparison than your average insect.
Consider the hardware estimate of 36.8 petaflops of speed to equal a human brain, and that IBM’s “Sequoia” supercomputer will be capable of 20 petaflops and scheduled for delivery in 2011. Now consider that a modern desktop PC is more powerful than a supercomputer from 10 years ago. With history proving that processing power is growing exponentially and the exponential trend is growing exponentially after each new system comes out, I don’t see size as an issue. Economies of scale show that electrical needs go down as size goes down, so that won’t be an issue either. As for robots, I see my previous statements on size/electrical requirements not being needed at all though…instead of fitting all that in a robot, just give it a wireless modem to connect to whatever-sized supercomputer…that takes size and electrical needs out of the equation.
One fact that is mormally not included is a size comparison and power comparison.
The human brain is about 1/1000000 the size and takes 1/1000000 of the energy of the so-called super computers.
I give it hundreds of years before they can meet that.
Unless you are talking about robots larger than the average home, no one alive today will see a thinking robot.
Post a Comment(Requires free registration)
- Please avoid offensive, vulgar, or hateful language.
- Respect others.
- Use the "Flag Comment" link when necessary.
- See the Terms and Conditions for details.
Advertisement