Moore’s Law has famously held firm for nearly 40 years, and is expected to dictate improvements in microchip design for a few years yet. An audience at the University of Auckland this week heard some predictions about the kind of future that Moore’s Law anticipates.
Intel founder Gordon Moore predicted in 1965 that the number of transistors on a chip would double each year. Ten years later Moore’s Correction stated that the number of transistors per chip would double every 18 months.
In the years since, Moore’s Law has become self-fulfilling, Professor James Goodman told an audience at the university. Goodman was giving his inaugural lecture, titled "What does Moore’s Law tell us about the future?"
Goodman says there have been periodic predictions that Moore’s Law would falter, but each time the industry has managed to maintain the exponential rate of development.
“The reason that we keep going is that whenever we come up against some fundamental limit we go around it,” he says. “We push against some other laws of physics that aren’t so constraining.”
There are numerous other examples of exponential growth, Goodman says, citing the price and capacity of computer memory, disk storage, internet growth and processor speeds. Intel’s stock price, until recently, grew exponentially. As a result, we have become “addicted” to exponential growth, he says. “A company that’s not growing exponentially is considered in trouble.”
Even so, there are physical limits that must constrain Moore’s Law. Goodman calculates that in 2009 we would produce a processor for each neuron in each human on the planet. Shortly before 2050, we would produce as many transistors in one year as there are milligrams of mass in the Earth.
“At some point this gets ridiculous, doesn’t it?”
A financial limit might strike first. Rock’s Law, named after a venture capitalist, suggests the capital cost of chip-making equipment will double every four years. “How many more years can we afford to build chip plants?” asks Goodman. “We may get to the point where we can’t even afford to build one.”
With the growth of computing power come concerns about privacy, particularly with the emergence of the internet as a single “giant network”. Goodman refers to a 2000 Wired article by Bill Joy, a co-founder of Sun. Joy raised his concerns about emerging technologies such as genetic engineering, nanotechnology and robotics in an article titled "Why the Future Doesn’t Need Us".
Goodman, formerly at the University of Wisconsin-Madison, joined the University of Auckland this year. He was attracted by New Zealand’s “less blind acceptance” of new technology, greater concern for privacy and willingness “to explore a different path”, he told the lecture audience.