The twenty-first century will be a critical time for humanity. As if the threats of terrorism, nuclear war and disease weren't enough, we are approaching a time when technology will impose dramatic changes on us all.
A couple of years ago, I read The Age of Spiritual Machines: When Computers Exceed Human Intelligence by Ray Kurzweil. Kurzweil's book transformed my world view. Kurzweil himself is a bit of a showman, and I don't subscribe to his utopian views of the future. However, his thesis is basically sound: within the next 50 years, computers will exceed human intelligence by orders of magnitude.
Today, computer power is growing exponentially. Moore's Law originally described the number of transistors on a Silicon integrated circuit. Historically, Moore's Law has meant that every two years, we get twice the circuitry on a chip running at twice the speed for the same price. The conventional wisdom is that Moore's Law will expire in the next ten or fifteen years, when the limits of two-dimensional Silicon semiconductors are reached.
What Kurzweil shows that this exponential growth in computing power pre-dates the era of Silicon integrated circuits. If Moore's Law transcends the medium of computing technology, it's plausible that we will see continued exponential growth in computing power indefinitely. Already, optical computing, carbon nanotubes, diamond substrates, superconductivity and three-dimensional Silicon are all showing promise for uninterrupted performance gains. Kurzweil estimates that if computer performance continues on the same curve, then by about 2025, a $1000 personal computer will have the same raw computing power as the human brain. By 2048, a $1000 PC will have the same raw computing capacity as 30,000 human brains.
Of course, critics of "strong artificial intelligence" have put forth arguments, albeit weak, that claim that computers will never be truly conscious or intelligent like humans. Kurzweil has excellent rebuttals to his critics in a subsequent book. However, from a short-term (twenty-year) point of view, the strong AI question is not critical. It's reasonable to claim that, whether sentient or not, machines will be able to perform tasks currently undertaken by 99% of all employed humans.
Needless to say, this will drastically alter the economic structure of our world. While initially quite frightening, there are potential benefits. The cost of goods and services will be massively reduced, so it won't require much income to live like a king. Imagine how much a hotel stay will cost when the hotel is built by robots out of raw materials mined and refined by robots, and even staffed by robots.
Jeremy Rifkin has written a book called The End of Work that makes a the case that structural unemployment is a natural consequence of capitalism and technological advancement. Reading his book in isolation, one might be tempted dismiss it as a Malthusian prophesy of doom. I could even see myself arguing that historically technology and market forces have led to more employment and more labor. However, Kurzweil shows us that our working days are numbered. While today's machines may be a million times less intelligent than we are, we're not getting any smarter! Machines will soon have abilities that rival or exceed our own, a situation unaccounted for in traditional economic models.
What I take with me from Rifkin's book is that the End of Work will likely be more gradual than I initially feared. There's probably no critical technological innovation that will push us all off an economic cliff. Instead, automation will increasingly seep into the economic fabric, largely ending employment over the next two decades.
My prediction: at some point in the next twenty years, high "unemployment" rates will not be viewed as intrinsically negative.