Kurzweil recently gave a presentation at the Business4Site conference in Los Angeles. While his scope was broad, several of the points he made were about the future of technology (and the expandingor is that contracting?digital lifestyle). In particular, he focused on exponential "leaps" in technology evolution.
By Kurzweil's account, each technological paradigm shift accelerates subsequent innovation logarithmically. Other futurists such as Alvin Tofler have held similar views, but Kurzweil demonstrates this exponential rate of change quantitatively, in terms of both the nature of innovation and the time of innovation. His math presents a compelling picture that allows him to make some interesting forecasts. This view of change happening in explosive leaps has been coined Kurzweil's Law.
Kurzweil forecasts that, by the year 2020, computer technology will surpass the human brain in the capability to "think." Computers are good doing complex calculations involving a number of variables very quickly. However, pattern recognition is still a difficult task. Kurzweil reasons that it is only a matter of time before we can create computers with as many "synapses" as the human brain. Kurzweil's Law predicts it. In 1968, a transistor cost $1. Today 10,000,000 transistors cost $1. According to Kurzweil's quantified schedule, by 2020, we will create computers with more power than the human brain. And given the current rate of miniaturization technology, which is exponential also, we should be able to fit all that computing power into an apparatus that is considerably smaller than the current size of Big Blue, the famous chess-playing computer.
Another point that Kurzweil makes is about the changing nature of the computer architecture overall. At the present time, a computer main processor is electronic, with bits turning on and off as the result of a change in electrical charge. Kurzweil theorizes that recent innovations in biotechnology will allow us to create technologies that will transform computer chips from electronic components to molecular nanotech components. (See http://news.nationalgeographic.com/news/2003/02/0224_030224_DNAcomputer.html.) Whereas in the past we flipped bits to make a processor work, in the future, microprocessors will be a collection of programmable molecules.
The implications of this theory are profound. If things go as Kurzweil says they will, someday Intel or one of it biotech subsidiaries might be in the brain-creation business, making microscopic molecule processors that have more power and wherewithal than today's silicon-based chips.