We’re all used to technology changing quickly, and computers seem to change faster than most things. For example, flat LCD monitors and DVD/RW drives are now the norm, yet more changes are coming and very fast.
These will - somewhat strangely - result in your computer virtually disappearing. In the previous article in this series Duncan Campbell introduced Cloud Computing; if you already use Gmail, Hotmail or Yahoo mail to access your email while you are away from your regular computer, then your email is already in the Cloud.
Webmail services like Google and Yahoo promise unlimited storage. Google also offers many Web-based applications which you access through your Web browser. In the future, Google wants to be the place where you store all your email, documents, photos, videos, and every other conceivable type of digital file. Google will use your data to try to target advertising at you.
You’d think Google must have an enormously powerful computer to deliver all its services, but in fact, it consists of tens of thousands of relatively inexpensive PCs, all housed in a large air-conditioned warehouse called a server farm.
The PCs in the farm are used to store data, but also to provide processing power for computation through a Grid Computer. A Grid is simply a network of inexpensive PCs connected together so all their processing power can be combined to create a very powerful virtual computer. Typically, for a problem to be able to be handled by a Grid Computer, the problem itself must be able to be divided into small sub-problems that can each be allocated to a separate PC. Fortunately, editing a document online or providing you with answers to a Web search are just such tasks. Indeed, many large companies now use the spare processing power of all the computers in their offices as Grid Computers after office hours. So if you work in a large office, the PC on your desk may be processing the payroll, managing stock control or predicting the cost of oil in the future, all while you sleep.
You can even set up your home PC to work on a grid when you’re not using it. Perhaps the best example of this is the SETI@home project (setiathome.berkeley.edu), which uses your home PC to analyse radio signals obtained from radio telescopes to search for extraterrestrial intelligences. There are currently over 1.3 million computers connected to the SETI grid, and you can add yours and help find ET.
The Cloud and the Grid mean that we are moving into a future where all our data is stored in the cloud, along with most of our processing power. This future is called “ubiquitous computing”. What this means is that wherever you go, provided you are connected wirelessly to the Internet, you will have access to all your data files and all the processing power you may need on demand. This will have profound effects on the sort of hardware that we’ll all be using. However, for once this doesn’t mean more complex, harder to learn, or more expensive. In fact, we can start to see this process happening in the very latest computers. Apple’s ultra-thin MacBook Air notebook is so thin not because of what it has, but what it doesn’t have. The MacBook Air doesn’t have a CD or DVD/RW drive. Apple believes permanent disc based-storage is a thing of the past. The Air doesn’t need an Ethernet port, since you’ll use wi-fi. You can even buy it with a 64GB solid state hard drive that consumes less power than a mechanical hard drive and is more reliable in a portable product. You see, as all your data is stored online you’ll no longer need hundreds of gigabytes of storage on your computer; in fact you’ll only need enough to cache your data before it is sent to the cloud or as it is received from the cloud. Since you’ll be able to access grid processing power when you want to edit that holiday video or play the latest game, you won’t need the latest powerful chips from Intel and lots of RAM in your computer either. You’ll only need enough processing power to drive your video display and Web browser.
Computers will therefore get smaller and smaller, until they’ll eventually become so small they’ll virtually disappear. Academics are already experimenting with wearable computers. For example, the famous MIT Media Lab has a Web site dedicated to its work (www. media.mit.edu/wearables). The neat thing about computing becoming ubiquitous is that it opens up a new range of ways that we will interact with computers and that they will interact with us.
One of the key elements of ubiquitous computing is that the computer is locationsensitive and always knows where you are. You may remember a scene from the Tom Cruise sci-fi movie Minority Report, in which he walks into a shopping mall. As he does so, a large screen greets him personally, and as he moves around the shopping centre, advertising billboards change to show products and special offers that he might be interested in. This is not science fiction. In the near future as you walk around town (hopefully if you’ve agreed to the service), you’ll be made aware of special offers and discounts at businesses you are walking past. Initially this will be by SMS to your GPS-enabled mobile phone, but eventually it may be directly to a heads-up display in your sunglasses or contact lenses.
Social networking services (like Facebook and Bebo) will also be ubiquitous; as you look at a menu in a restaurant you ay be informed that your best friend liked #43, the stir-fried chicken with cashew nuts, last Tuesday. This echnology, called “collaborative filtering”, you can use now to help you choose which movies to see at MovieLens (movielens.org/html/tour/index.html). As more and more people use ubiquitious computing, these services will become more useful. Your car, which will also be connected to the ubiquitous computer, will know where empty carparks near your destination are and when you get to the bar, the barman will have your“usual” drink waiting for you - your car will ave told him you’re almost there. Many of these applications seem trivial, but everything will be designed to make your life easy. Imagine a personal assistant working for you 24x7, knowing your habits, likes and dislikes, and anticipating your every need and smoothing over any unforeseen problems. This is one of the aims of Artificial Intelligence research.
Artificial Intelligence (or AI) always makes you think of the computer HAL in 2001, Commander Data in Star Trek or maybe even the Terminator. While creating a fully conscious intelligence is a long-term goal of AI, most researchers work towards much more attainable goals. In fact, many useful problems were solved by AI decades ago, but it’s only been recently that other technologies have enabled them to be deployed. A great example is car navigation systems. The basic algorithm (called A*) that enables the unit to find the shortest route between two places was invented in 1968. However, it wasn’t until a few years ago that a useful LCD screen, a GPS and enough processing power could be put into a small box for under $500. The military have been using hugely expensive precursors of these for decades.
Fighter pilots have been using heads-up computer displays and voice commands to control most on-board systems for years, and now similar heads-up displays are appearing in luxury cars. Further into the future, AI researchers are creating systems that can recognise words from your brain activity alone. Tom Mitchell, a leading Machine Learning scientist (www.cs.cmu.edu/~tom), is teaching computers to recognise words based on brain activity. The idea is that in the future you’ll be able to just think the words to an email and they will be written and sent, all without you actually touching a computer. Professor David Leake, Editor of AI Magazine, the official magazine of the Association for the Advancement of Artificial Intelligence (www.aaai.org), says: “The next generation of intelligent user interfaces will know our preferences, understand our context, anticipate our needs to actively pursue them and will learn from experience to do better. They’ll know how to be partners, keeping us in control. The smarter they are, the more they’ll vanish into the fabric of our lives.”
Of course, there are good reasons why AI researchers want to build robots. The principal one is that our environment (our homes and workplaces) is designed to be usable by human-shaped objects (us). If something is to help us around the home - vacuum, cook dinner, take out the rubbish or walk the dog - then it’s likely to be easier if it is reasonably human-like (do you remember the old cartoon of a group of Daleks looking at a flight of stairs while one of them says: “Well, there goes our plan to rule the Universe!”?). For the last decade, researchers have been competing in an annual global competition called RoboCup (www.robocup.org). Every year, teams of students and academics from universities all over the world compete in robot soccer. The ultimate goal of RoboCup is that by 2050, a team of fully autonomous humanoid robots will play and beat the human world champion team in soccer. Again, while this may seem frivolous, such technology would be useful for a huge range of other activities beyond soccer.
Of particular interest to researchers in Japan and Korea is the concept of robots as companions and medical assistants for the elderly. As some industrial societies age, there will simply not be enough young people to provide care and companionship for all the elderly. A robot companion for an elderly person could perform basic household chores, monitor the person’s medical condition, administer medications and, most importantly for many, be a round-theclock companion. Professor Jim Warren of Auckland University says: “There are two really different things the healthcare robots of the near future can do. They can serve as a telepresence for a remote caregiver (usually a monitor is placed as the robot’s ‘head’ and displays the face of the remote caregiver) or they can be autonomous and engage in voice interaction in their own right (often in combination with a touchscreen).
At the University of Auckland, we’re just starting a collaboration with Korea’s Robotics Institute to look at a low-cost robot aimed at providing a combination of companionship and nursing assistance for the elderly. Our vision is an intelligent medicine chest that can follow you around and chat with you.” The collaboration, supported by $1.8 million in government funding, will establish the New Zealand/South Korea Robotics Centre.
However, not all robots will be humanoid; some researchers are working on developing nanorobots. These machines operate at the scale of molecules. A group at Rice University, in the US, has recently developed a nano car that has a chassis, pivoting suspension and freely rotating axles made of organic molecules. The wheels are buckyballs; spheres of pure carbon containing only 60 atoms each (see tinyurl.com/4t2pk4). The nano car measures about four nanometres across (a nanometre equals one millionth of a millimetre). Scientists are envisaging medical nanorobots that could move through your body, killing viruses and cancer cells, repairing damaged organs, and making you young again. Such machines could revolutionise the health sciences.
But the science of nanotechnology is still in its infancy and there are significant hurdles to overcome; the sheer mechanics of size, for one. Conventional tools simply can’t make components that small. The concept of nanotech development is known as Molecular Simulation. Part of its task is developing tools to predict how materials and fluids will behave at nanometre size. Then there’s the question of shrinking computer chips to nano size. Current technology allows chipmakers to print circuits that are 1000 times thinner than a human hair, but nanotech researchers have developed carbon tubes that are 50,000 times smaller than a hair. These are touted as the future conductors of information. One snag: longer nanotube threads have been found to mimic the toxic qualities of asbestos. They’ve produced inflammations and lesions in rats – no way are you going to inject something like that into a human body. There are decades of research ahead in this field. Learn more about nanotechnology at science.howstuffworks. com/nanotechnology.htm and www.nanotechproject.org
Building robots at a molecular level has parallels in computing, where researchers are now considering computers that use the physical properties of atoms to perform computation. Quantum computers have the potential to perform certain calculations significantly faster than current silicon-based computers - perhaps a billion times faster. Although quantum computers are still largely theoretical (the largest quantum computer built so far contains just 16 qubits, the basic processing unit of a quantum computer, but can play sudoku), the ideas have already had an impact on cryptography (the secure encoding and decoding of sensitive private information). Because of their theoretical massive processing power, quantum computers could decode information that could take centuries to crack using the fastest current computers. Fortunately, quantum computers can also create secure codes that could never be cracked by current computers, enabling very sensitive information to travel securely across the Internet. Professor Cris Calude, a theoretical computer scientist at Auckland University, believes that: “Quantum computers can outclass classical computers for specific problems - encryption, teleportation, searching unsorted databases; but for most jobs their advantages over their classical counterparts are modest or simply inexistent.”
As we said in our last issue, these predictions could be overtaken by scientific research already in progress. At the time of writing, the Large Hadron Collider, the gigantic particle accelerator built in Switzerland, was due to be switched on (see News Sites, pg 77). The computer network developed to process information from the collider is more than 10,000 times faster than a typical broadband connection. One day, such computing power will be available to everyone – you could download a movie or an entire music catalogue in seconds. British researchers have developed a technique called “the trapped rainbow”, which could allow lightbased data processing and storage, rather than the much slower current electron-based method
Of course, many of us won’t be alive by the time such ideas become reality. But as the old song goes, if you don’t have a dream, how are you going to have a dream come true?