Developing Artificial Intelligence

The brain is so complex that even the fastest supercomputer can't simulate it in real time.

© iStockphoto/Henrik5000

Artificial intelligence is a term that has been thrown around for decades. In the past, computers were just powerful machines that could crunch enormous numbers -- they couldn't think for themselves. A computer could only follow explicit instructions.

Today, engineers and computer scientists are trying to make the leap from computation to thinking. They've met with some progress. In 2008, scientists used the BlueGene L supercomputer to simulate a mouse's brain. That might sound simple, but a brain -- even one belonging to a virtual mouse -- is incredibly complex. So complex, in fact, that the powerful computer could only run the simulation in bursts of 10 seconds [source: BBC News].

In 2009, Cornell researchers created a computer program that was able to derive the basic laws of motion by analyzing the movements of a pendulum. The program took a series of measurements and used a genetic algorithm to extrapolate the basic laws of physics.

In the future, we may see machines capable of solving complex engineering problems. We may even reach the point where computers design even more powerful machines. How's that for deep thought?