The Future of Computing: Ten Competing Visions

"It's a bug that cars were invented before computers," soon-to-be-replaced Google CEO Eric Schmidt said in September. The Google chief was looking ahead to a day when cars drive themselves. The surprising thing is that this Knight Rider-like future isn't that far off. Google made waves in October when motorists spotted the company's driverless cars shuttling company engineers around the highways of California. Google soon reported that its driverless cars had logged 140,000 miles as of October, guided by an array of "video cameras, radar sensors and a laser range finder." Google admits that driverless cars remain experimental and are hardly ready for prime time, but the project "provides a glimpse of what transportation might look like in the future, thanks to advanced computer science."

Futurist Ray Kurzweil and his team at KurzweilAI spend a lot of time developing future-of-technology scenarios--eventualities such as an Internet in 2050 that is a trillion times faster than what he have now, artificial intelligence before 2100, and space warfare. But Kurzweil's most famous prediction is something called the Singularity: the moment when humans will physically merge with machines. For Kurzweil, the Singularity will be a technological flashpoint at which humans become nearly immortal cybernetic beings. In addition, our minds will become faster, have greater capacity, and be able to share knowledge with others in much the same way that we transfer files between computers today. Everything will be great in this new cybernetic future...until a superhacker taps into our brains and turns the lot of us into an army of cyberzombies. The good news: Cyberzombies don't at all mind being cyberzombies.

While Kurzweil and his associates make prognostications on the basis of digital tea leaves, the folks over at IBM's research labs make the predicted future our reality. On February 14, an IBM super computer named Watson will show up on the TV game show Jeopardy to challenge two human Jeopardy champions, Ken Jennings and Brad Rutter. IBM's Watson already won a Jeopardy practice round in mid-January, so the computer's prospects for Valentine's Day look good.

Watson is seen as a giant leap forward for artificial intelligence, because the supercomputer has to understand and react to everyday human speech. Most computers that react to human speech do so based on clearly defined keywords such as "call," "play" or "search." Not Watson. IBM's latest supercomputer must hear the Jeopardy answers just as humans do, understand the meaning of each statement, and then decide how best to formulate a response. Watson is powered by 15TB of RAM and about 2880 processor cores that can perform 80 trillion operations per second.

Technologists have been dreaming of the day when people's computers won't be laptop or desktop machines, but handheld devices that users fit into a base station equipped with a monitor and keyboard. Motorola's Atrix 4G, introduced at the Consumer Electronics Show in January, is the closest device yet to a PC that you can carry in your pocket. The new dual-core Android-based smartphone can slot into a base station to access a virtualized Windows 7 desktop that is hosted online via the Citrix XenApp. When you're finished working, you remove the device and it functions as a smartphone running Android 2.2 (Froyo).

The Atrix isn't quite like having a full-fledged Windows machine embedded in a smartphone, since the virtualization servers do all the heavy lifting--but it's awfully close. The Atrix is due out on AT&T before the end of March. The downside: Citrix's virtualization client is for enterprise users only.

Imagine a world in which you didn't have to swipe a credit card to charge something, click a button to open your garage door, or undergo a thumbprint scan to reveal your identity. Instead, you'd have your body scanned by a machine or proximity sensor. That's the idea behind embedding a Radio Frequency Identification (RFID) in the human body. The FDA has already approved one RFID chip for human use: the VeriChip, owned by PositiveID. The company plans to use the technology to detect glucose levels in the body in real time; meanwhile, it has ditched its implantable medical history identity chip, VeriMed. But PositiveID's reversal hasn't prevented enthusiasts from implanting RFID chips in their bodies. Of course, the idea of people being tagged with microchips has prompted some critics to interpret human tagging as a sign that a Revelations-style apocalypse is imminent.

For precision work, surgeons of the future may rely on nanorobots--tiny devices made of synthetic or biological material that measure just one-billionth of a meter or 3.93700787 × 10^-8 of an inch. Nanotechnology could be used to repair or replicate human tissue, to seek out and treat cancer cells, or to perform other tasks at a microscopic level. Already, scientists from the Institute of Robotics and Intelligent Systems in Switzerland are experimenting with nanorobots injected through a patient's nose to help remove blood clots in from the patient's eyes. From there it's just a short leap to Star Trek: Voyager's Seven of Nine, right?

Kinect isn't the only motion-sensing technology that Microsoft is working on these days. Leaked documents unearthed in June 2010 claim to show future plans for something called "Your PC, Your Way" as part of the forthcoming Windows 8 operating system. One possible feature of the new OS is eye-tracking technology built into the computer's Webcam that can pause a video if you look away from the screen. Windows 8 may also use facial recognition instead of a password to log you into your user account, and proximity sensors to power the PC up or down when you enter or leave the room. A public beta of Windows 8 should arrive later this year, but it's not clear whether the new tracking features will be part of it.

It may not be part of the forthcoming Windows 8, but technology already exists that would enable you to control a PC by thought alone. Scientists at New York's Wadsworth Center have created a system that can detect electrical patterns from your brain and translate them into computer commands. The downside is that this brain-powered technology requires special electrodes and a gel spread all over your head, though scientists are beginning to experiment with electrodes implanted in the human skull. The hope is this technology can help quadriplegics interface with computers. Brain wave computing is a long way from being ready for everyday use, but it's a cool concept. Check out this 60 Minutes report about exercising mind control over PC matter.

The day when robots perform all of your house chores is getting closer, thanks to start-ups like Willow Garage. Based in Menlo Park, California, Willow Garage aims to make robots the next great personal productivity tool. From getting you a beer to folding laundry, Willow Garage's robots will help people carry out tasks in everyday life. So far, however, Willow Garage sells just one type of robot: the PR2, which is designed for research projects at universities and other institutions. The PR2's current price tag is a $400,000 (not including tax and shipping), so don't expect to see a Rosie the robot maid in your neighborhood anytime soon.

If you want to use multiple processor cores in your computer at home, the most power you're likely to find today is six cores on a single chip. But Intel 48-core experimental chips do exist, which prompted Intel scientists to say that a chip with 1000 cores is theoretically possible. It's unclear as yet how fast 1000-core chips would be; members of the current crop of 48-core chips aren't much faster than a 1.83GHz Atom processor. The biggest challenge of a 1000-core chip isn't speed, but the necessity to rethink software design to take advantage of so many cores at once--Windows 7 64-bit can handle just 256 cores. However, Intel's scientists may not be the first to hit 1000 cores on a single chip. In December, Scottish scientists claimed that they had built a working prototype of a 1000-core chip that is 20 times faster than any chip on the market today.

Jetpacks, Androids, and the Future That Never Came