When I was growing up in the Deccan—a land of black soil, a black magic called bhanamati and great backwardness—my dreams invariably revolved around a globe-girdling computer called Multivac.
Through sweltering summer days and nights without electricity in the late 1970s, I immersed myself in the fantastic world created by the great science-fiction writer Isaac Asimov, who wrote many stories of a world run by a sentient supercomputer. Multivac lay buried deep underground, controlled all human existence (tired of its burdens, it once planned its own demise), struggled to understand jokes, and even ran the democratic vote by choosing one person to represent the entire electorate of the US.
It all seemed so faraway, so otherworldly, so impossible.
Today, how quickly the world of Multivac (Asimov meant it as an acronym for multiple vacuum tubes, the bulky predecessors to transistors and silicon chips, the building blocks of electronic devices) is emerging. The Internet may not be self-aware, but its web is worldwide. So-called artificial or machine intelligence is not self-aware, but it runs more of the world than we realize, including car brakes, power plants, microwave ovens, cellphone circuits.
For a while in the 1980s, it seemed the advance of computers based on silicon chips would face physical limits. Moore’s law, postulated in 1965 by Intel co-founder Gordon Moore, was readjusted to predict that the number of transistors in a computer chip would double every 18 months, down from Moore’s count of once every 12 months.
Moore himself has been amazed at how scientists have kept the silicon-chip going. Your cellphone’s brain, a silicon chip called a semiconductor flipping billions of switches as you make calls or surf the net, has more computing power than Apollo 11, the US spaceship that first put man on the moon in 1969. Your iPod’s semiconductor is more powerful and more than 100 times smaller than those used when I was growing up. The best computer circuits being created today are 40 nanometers, each packed with one million transistors on less than one-thousandth the width of a human hair.
The advances flow in with such regularity that they are barely recognized in the popular media. For instance, two days before 2010 ended, researchers at the University of Glasgow announced they had figured out how to divide transistors on a chip into smaller groups, each doing a different task and collectively processing, at speeds 20 times faster than current desktops, an algorithm central to playing movies.
Apart from making smaller, faster computers, there are two other ways of pushing machines to new frontiers: First, writing better algorithms, or formulae that let computers solve a problem; and second, getting more computers to collaborate, regardless of where they are physically located—that’s what the latest wave of cloud computing is all about, freeing computers from a hard drive and logging in to this worldwide gathering of computers to access and store information.
If this is beginning to resemble the world of Multivac, that’s because the world is indeed moving towards a globe-girdling net of computing. It’s been variously called the Evernet, the Supranet and a host of other names that essentially mean a newer Internet.
The Internet today is a precursor to what US techno-evangelist George Gilder once called the Telecosm (he hosts an annual conference of the same name, drawing tech leaders, engineers and inventors), a world where human communications will be universal, instantaneous, unlimited in capacity, a world of unlimited bandwidth—the unseen pipes and switches that funnel the Internet’s data. Gilder once said that the Internet today serves largely the same purpose and is the same point in history today as the rutted horse-ready roads of early 19th century America down which the first Model T cars trundled.
Millions across the world, and certainly in India, are at present working towards pulling the Internet into the 21st century. Most don’t see the big picture, occupied as they are in building myriad building blocks: new chips, new algorithms, new architectures, new technologies. You will see many parts of this new Internet show up in the new computers, communication and entertainment systems you may buy this year.
As for the end of silicon, over the last six months of 2010, scientists reported progress on at least four possible replacements: diamond, graphene (both varieties of carbon), DNA and light beams. None of this is easy. On 23 December, Dutch scientists announced they had succeeded in (only) controlling the building blocks for a future ultra-fast quantum computer, which would not be built with transistors and silicon but with as yet-unknown materials. A working quantum computer— very much the esoteric, even bizarre, stuff of alternative universes—could be 50 years away. Or could it?
In less than a lifetime, the Internet has moved rapidly along the scale of progress Asimov envisaged. There is much to look forward to in 2011.
This is a fortnightly column that explores the cutting edge of science and technology. Samar Halarnkar is editor-at-large, Hindustan Times. Comment at firstname.lastname@example.org