IBM Research - Zurich

Miniaturized information storage in atomic-scale antiferromagnets show the binary representation of "s" (01010011).

Twelve atoms are all that's required to store a bit of computer code – a 1 or 0, according to a new discovery that probes the limit of classical data storage.

Computer hard drives on the market today use more than a million atoms to store a single bit and more than half a billion to store a byte, which is an eight-bit-long unit of code sufficient to write the letter A, for example. 

The new technique uses just 96 atoms per byte, allowing for hard drives that store 100 times more information in the same amount of physical space, according the researchers behind the discovery.

"We can put the neighboring bits at the same atomic spacing that the atoms have inside the bit," Andreas Heinrich, a lead investigator in atomic storage at IBM Research in California, told me.

"So, we can really pack them right next to each other."

Unconventional magnetism
The storage technique is based on an unconventional form of magnetism called antiferromagnetism.

Normal magnets used in today's hard drives — and to hold your child's artwork on the refrigerator — are made of ferromagnetic materials. The spins of atoms in these magnets align with each other. 

That's "good" because it provides an overall magnetic field that we can read as a bit — a 1 or 0, explained Heinrich.

"But it is bad because the magnetic field from one  bit will interfere with the magnetic field from the neighboring bit and so you can't pack these bits too close together because they'll just talk to each other," he said.

Antiferromagnets, by contrast, cancel each other out, so there's no magnetic field emanating from them. That means they can be packed close together, allowing for the increased data storage density.

Atomic building blocks
Heinrich and his colleagues were led to antiferromagnets on an exploratory research quest to find out how small they could make a magnetic device and use it for classical data storage.

They used a scanning electron microscope, which allows researchers to see and manipulate atoms, to build a data storage system up one atom at a time.

Scientists know that single atoms exhibit funky quantum behaviors that require a different set of equations to describe. But where is the transition between quantum and classical behaviors?

At eight atoms, the team found, a bit was stable for a few seconds and "at 12 atoms it turns out that the classical concepts are so good that these magnetic structures hold their magnetic state for days," Heinrich said.

"We said that's good enough to call it storage."

The caveat is that this stability is found when the atoms are kept at a chilly minus 268 degrees Celsius, or 5 Kelvin. Stability at room temperature, Heinrich said, is thought possible at around 150 atoms.

The findings are reported today in the journal Science.

Consumer devices
The finding could lead to terabyte hard drives the size of a pinhead or thumb drives that hold every movie you've ever seen, Rick Doherty an analyst with technology consulting firm Envisioneering Group told me.

Other applications may come in medical devices such as magnetic nanobots swimming in the bloodstream that can be attached and detached to tissues electronically.

"It is going to make life better, allow us to save energy, make smaller structures, and maybe one day magnetic computer logic," he said.

While transferring some of this atomic scale technology to real world gadgets may take awhile, Heinrich said the use of antiferromagnets in traditional hard drives is likely as soon as five years now.

"If you were able to use antiferromagnets instead of ferromagnets, you … could pack these things denser and therefore you could store a lot of information on your drive."

More on atomic-scale computing and storage:

• Four-atom-wide wire may herald tiny computers
• Salt — table kind — can boost hard drive storage
• Hard drives are getting bigger, better
• New technology boosts hard drive capacity

Nothing focuses your attention on the future like a forecast, especially when it comes to the technologies that will be changing daily life in the years to come. Five years, to be exact. That's why forecasts like IBM's annual "Five in Five" are so thought-provoking, even if they're occasionally wrong.

Actually, IBM's record is pretty good: This month marks the five-year anniversary of IBM's first list of five technologies that were expected to make the most impact in five years' time. The company nailed 2006's predictions on the rise of telemedicine, location-aware mobile devices, real-time speech translation and nanotechnology. But the fifth prediction, which focused on the rise of virtual 3-D environments, hasn't worked out the way IBM expected. Sure, Second Life is still around — in fact, I'll be hosting my next "Virtually Speaking Science" show in Second Life on Jan. 4. But such virtual worlds haven't become the principal vehicle for real-world commerce ... yet.

"It's not perfect," admitted Bernie Meyerson, IBM's vice president of innovation. Sometimes the company's researchers latch onto a idea whose time has not yet come, and perhaps never will. But for the most part, "this stuff has actually panned out a lot," Meyerson said.

Is technological progress always a good thing? Not necessarily, if you're talking about key-logging software on mobile devices, or government-supported spyware. The latest predictions from IBM, issued today, have lots of potential for a dreams-vs.-nightmares debate:

1. People power will come to life: Devices will be built to capture the power generated as you jog, or ride your bike, or run water through the pipes of your home. Even the heat that builds up in your computer's circuitry could be harvested rather than going to waste. Engineers have already developed electricity-generating backpacks and shoes that could build up enough juice to power the electronic devices you carry around with you. On the other end of the scale, IBM researchers in Ireland are already working on ocean wave-power projects.

The down side? It's tricky to design devices that produce enough power to make them cost-effective — and at the same time comfortable to wear. A lot of people already feel tied down by technology. Will they be willing to pile on the extra bulk of power-generating contraptions? Will the future economics of energy justify micro-power harvesting?

2. You will never need a password again: Instead of trying to keep track of all those different passwords for your online accounts, and still worrying that someone will break in and steal your identity, we'll find ourselves actually using technologies such as iris recognition, face recognition and voice recognition to log in. "The world of biometrics is coming," Meyerson said.

The down side? It sounds a little creepy, like the world of the movie "The Minority Report," and it could be seen as another intrusion on personal privacy. Meyerson, however, argues that "you can deal with the creep-out factor" by making sure users have the freedom to opt in or opt out of biometric identity systems. The keys to your identity could be kept on your device rather than in a central repository. And using multiple methods — for example, iris plus voice — would make it astronomically unlikely that someone could crack your code. "Personally, I think the risk is far greater not doing this," Meyerson said.

3. Mind reading is no longer science fiction: This prediction isn't about psychic powers. For years, researchers have worked on ways to control robotic arms or blips on a computer by reading brain signals — and IBM thinks that technology will be ready for prime time (or drive time) within the next five years. That would be particularly good news for quadriplegics and "locked-in" patients looking for better ways to interact with the outside world. It might lead to better approaches to medical concerns ranging from autism to stroke rehabilitation. And think of the cool video games you could be playing when you just have to think something to make it so. Meyerson said companies such as Emotiv Lifesciences are already preparing the way for this brave new world.

The down side? Once you give someone direct access to your brain, wouldn't it be at least theoretically possible to eavesdrop on your innermost thoughts? "People worry about something that will interpret your brain," Meyerson said. "That's not what we're talking about here." But as long as we're talking about science fiction becoming reality, we'd better keep the dark side of the sci-fi story in mind as well.

4. The digital divide will cease to exist: IBM suggests that the cost of smartphones and online services will become so low that everyone will be plugged into the global network. "It's gotten to the point where it's cheaper to have a cell phone than to have a bank account," Meyerson said. The gap between haves and have-nots will fade away in the digital world. IBM researchers are already working to make this vision a reality. In India, they're helping to create technologies that allow even illiterate and semi-literate people to use mobile devices for basic services.

The down side? Who'll be in charge of this digital paradise for the haves? Privacy advocates might see this as a fresh cause for concern. As governments rely increasingly on digital networks to distribute services, will life become that much more difficult for those who are unable or unwilling to plug in?

5. Junk mail will become priority mail: This is the flip side of junk-mail filters. Computerized systems for filtering information will become so adept at reading your preferences that they'll become true digital assistants, presenting you with the data that you need (or want) to know while blocking the junk. In the next five years, you'll have the technology that turned the Watson supercomputer into a "Jeopardy" quiz-show champion at your fingertips. Watson might even take the initiative — for example, by putting tickets to a concert by your favorite band on hold the moment they go on sale, even before you've heard about it.

The down side? What if your personal digital assistant turns out to be a paranoid HAL 9000 instead of a helpful Watson? What if Watson goes rogue with your credit-card number? And what about the privacy concerns? Ten years from now, will the authorities be able to learn all about you by tapping into your junk-mail filter?

Are IBM's latest "Five in Five" predictions hits or misses? Visions of paradise, or another circle of hell? Feel free to weigh in with your comments below, and check out these past predictions:

• 2006: IBM issues five-tech forecast
• 2007: Five frontier technologies
• 2007: IBM follows up with a fresh 'Five in Five'
• 2008: A crystal ball for health, energy and more
• 2009: How future technologies will change cities
• 2010: Hits and misses in the five-tech forecast

Alan Boyle is msnbc.com's science editor. Connect with the Cosmic Log community by "liking" the log's Facebook page, following @b0yle on Twitter and adding the Cosmic Log page to your Google+ presence. You can also check out "The Case for Pluto," my book about the controversial dwarf planet and the search for new worlds.

A hundred years from now, will we be assimilated by the machines? Or will we assimilate them? These are the kinds of issues facing International Business Machines as the company begins its second 100 years.

Right now, most folks are thinking about the past 100 years at IBM, which is celebrating the centennial of its founding on Thursday. But for Bernard Meyerson, the company's vice president of innovation, it's all about the next century.

"That's pretty much what we think about," Meyerson told me today.

Meyerson has plenty to look back on, including his own not-so-small part in IBM's past innovations. When his cell phone dropped the connection during our telephone conversation, he called back and casually mentioned that he had a hand in creating the transistors built into that cell phone. And when I asked him to explain, he said, "I actually invented the technology known as silicon-germanium."

It turns out that IBM has played a behind-the-scenes role in all sorts of technologies, ranging from semiconductor development to barcodes to Wi-Fi. "IBM is a funny company," Meyerson said. "We don't force you to put a little sticker on anything that says, 'We're the smart guys.'"

But enough about the past: What about the future? "Going forward, you have tremendous opportunities," particularly when it comes to making sense of the huge databases that are being built up in all sorts of fields, Meyerson said. For example, imagine a system that can take medical records from the 285 million people around the world with diabetes, anonymize those records and analyze them, looking for potential new treatments or preventive measures.

"The fact is, there is no mechanism today that could do that, and the reason is that medical data is unstructured," he said. There's little consistency in how the records are kept, and medical conditions might be described in different ways by different doctors.

When you put together the volumes of data and the numbers of people that have to be covered in these massive, unstructured data sets, the figures mount up to quintillions of bytes. That's the challenge facing new types of computing tools — for example, the Watson supercomputer, which won a highly publicized "Jeopardy" quiz-show match earlier this year. Now Watson is being put to work on a tougher task: making sense of medical records, which is just the kind of job Meyerson has in mind.

Still other challenges await. Watson-style computers could digest the millions of data points involved in tracking the flow of highway traffic, then develop models to predict where the tie-ups could arise before they actually happen. The computers of the next century will have to handle a wide range of "big data" challenges, ranging from climate modeling to natural-language search engines for multimedia.

Meyerson doesn't expect Watson to answer that challenge completely. A hundred years from now, Watson will almost certainly be considered a quaint antique, much like the tabulating machines that were made back in 1911.

"Watson specifically is not the issue, as much as the combination of Watson's ability to interpret natural language, the capacity to store 'big data' and apply data analytics to come up with solutions for society," he said. "In the absence of natural language, you're going to have a short, unhappy life attempting this work. Without that key ingredient, how are you going to take the interaction of humans and machines to the next level and make it easy?"

Paramount Pictures

"Star Trek" captain Jean-Luc Picard (Patrick Stewart) is fitted with gizmos for a fictional Borg transformation. The blending of humans and hardware will probably be more artful in real life by 2111.

What will the next level be in the year 2111? "Honestly, at 100 years I'm genuinely unsure," Meyerson said. The past century has shown that the pace of technological advancement can be highly variable, depending on what kinds of breakthroughs come to the fore. But if Meyerson had to bet on one particular game-changing technology, it would be coming up with a direct interface between computing circuits and the human brain.

"If it turns out that there is a very natural way to communicate data back and forth without being obtrusive, then the whole world changes," he told me. This wouldn't be a Borg-like assimilation, in which humans look increasingly like machines. Rather, the machines would blend into the human body.

Does that sound like a grand dream for the next century? Or a nightmare? Feel free to chime in with your comments below.

More about the future:

• Hits and misses in the five-tech forecast
• Physics turns from fission to the future
• Take a test drive through the next 100 years
• Deep thinkers speculate about the next 50 years
• Futurist Ray Kurzweil reaches for immortality

You can connect with the Cosmic Log community by "liking" the log's Facebook page or following @b0yle on Twitter. Also, give a look to "The Case for Pluto," my book about the controversial dwarf planet and the search for new worlds.