The intelligence gap between man and machine will largely close by the year 2050, according to Intel's chief technology officer, who yesterday reiterated that point during a keynote address at the Intel Developer Forum.

At the IDF event in San Francisco, Intel CTO Justin Rattner said that the chip maker's research labs are working on human-machine interfaces and looking to foster big changes in robotics and in the ability of computers to interact with humans. He specifically pointed to work that Intel is doing on wireless power and on developing tiny robots that can be programmed to take on the shape of anything from a cell phone to a shoe or even a human.

"The industry has taken much greater strides than anyone ever imagined 40 years ago," Rattner said. "There is speculation that we may be approaching an inflection point where the rate of technology advancements is accelerating at an exponential rate, and machines could even overtake humans in their ability to reason in the not-so-distant future."

Just last month, Rattner, who also is a senior fellow at Intel, made similar comments in an interview with Computerworld, saying that perhaps as early as 2012, the lines between human and machine intelligence will begin to blur. The intelligence gap should become awfully narrow within the next 40 years, he added, predicting that by 2050, computing will be less about launching applications and more about using systems that are inextricably woven into our daily activities.

In that same vein, Rattner talked about programmable matter during his IDF speech. He explained that Intel researchers are working to figure out how to harness millions of miniature robots, called catoms, so they could function as shape-shifting swarms.

"What if those machines had a small amount of intelligence, and they could assemble themselves into various shapes and were capable of movement or locomotion?" he said. "If you had enough of them, you could create arbitrary shapes and have the assembly of machines that could take on any form and move in arbitrary ways."

The basic idea is that the catoms, which one day should be about the size of a grain of sand, could be manipulated with electromagnetic forces to cling together in . Rattner said that Intel has been expanding on research work done by Seth Goldstein, an associate professor at Carnegie Mellon University.

"We're actually doing it for real," Rattner said. He added that Intel started "at the macro scale," with catoms that were "inches across." The robots had microprocessors associated with them and could attract or repel one another via electromagnetism or the use of electrostatic charges, according to Rattner. "It's programmable matter," he said.

During his speech, Rattner showed off millimeter-scale 3D catoms and said that electronics could be embedded inside the miniature robotic spheres.

Jason Campbell, a senior researcher at Intel, said in an interview that the development and use of catoms will change the way people interact with computers and other devices in significant ways.

"Think of a mobile device," Campbell said. "My cell phone is too big to fit comfortably in my pocket and too small for my fingers. It's worse if I try to watch movies or do my e-mail. But if I had 200 to 300 milliliters of catoms, I could have it take on the shape of the device that I need at that moment." For example, the catoms could be manipulated to create a larger keypad for text messaging. And when the device wasn't being used, Campbell said, he could command it "to form its smallest shape or even be a little squishy, so I can just drop it in my pocket."

Campbell envisions that each catom would have a computer processor and some form of memory. Four years ago, he thought it would take 30 to 50 years for this kind of technology to be realized. Now, though, he estimates that the time it will take is much closer to 10 years.

Both Campbell and Rattner said the biggest obstacle will be figuring out how to make the micro-bots think like a swarm. Instead of sending individual directions to each catom, one set of instructions will have to be sent to make the catoms work together, so each one takes the correct position to create the desired 3D shape. But both were optimistic that it will happen, eventually.

"Sometime over the next 40 years, this will become everyday technology," Rattner said in an interview before his speech. And could catoms actually take human form? "Sure," he said. "Why not? It's an interesting thing to speculate on."