Singularity and Modern Man
Ray Kurzweil writes books that you realize, as you read them, are beyond the thinking you have indulged in - ever. He has written books on health and the impact of technology on growth. His last book, written in 2005, takes everything he has written before and adds to it in crucial ways. Gordon Moore has Moore's Law. Encapsulated in Kurzweil's the law of accelerating returns (Kurzweil, 35), Moore's law says that the amount of computing power that can be crammed into an integrated circuit doubles every year. Lately, the doubling is even accelerating. Kurzweil takes that doubling and thinks, not about the technology involved in the doubling, but in the result of the doubling. We all learned about arithmetic progressions way back when when some teacher said something like, "Which would you rather have, two tons of gold every year, or, an ounce of gold that doubles every year." Everybody wants the two tons of gold, and, of course, they picked the bad choice. Arithmetic doubling is a wonderful way to get rich, especially if you're doubling gold. So who cares? In ends up that, while most people don't give it much thought, all this doubling effects all sorts of things. DNA was discovered in the fifties. In the early nineties, they started work on the human genome, expecting to take fifteen years to decipher it. One team started late - way late - but still ended up in the same place, only much faster, because they re-arranged their thinking and methodology. Ask a scientist when genomics are likely to significantly effect man-kind (as if they weren't doing so already) and she's likely to say, "fifty years from now, give or take." Kurzweil says that nonsense - we'll be using genomic solutions much quicker than that, because of the doubling effect. It took us fifty years to get the genome, yes, but it won't take us fifty years to put it to work. Focusing on shapes of paradigm curves, Kurzweil describes a three step curve of slow growth, rapid growth, and, then finally, a leveling off (Kurzweil, 43). What we forget is that, when stacked end-on-end, all those curves really represent an exponential sequence, not a doubling sequence. Once things get rolling with a new technology and new iterations are created as a next chapter, so to speak, those chapter sum to rapid growth, to exponential growth. He plots out the telephone industry, the cell phone industry, dynamic ram, dynamic ram price, and on and on.
The point of all this is interesting. Allow yourself to be boggled at this point because you're not going to believe what I am going to say next. Eventually, and it is not going to take too long, computer's computing - thinking - processes will catch up with man's. And it is going to take less time than we think. Kurzweil's Theory of Technology Evolution says that there is a migration occurring toward everything becoming information - everything. Nanotech manufacturing processes will allow embedding of information everywhere. So the question becomes not when will artificial intelligence become useful, but when. Then we have to decide what high level artificial intelligence means for us. Now, consider what I just said. AI is going to make significant contributions to thinking. Is that a threat? By now you probably realize that I don't always see threats, but opportunities. And I think this AI will be useful, and perhaps, the ultimate opportunity. HAL, yes. But a useful HAL. Wait and see. Or profit now. Interesting.
Kurzweil, Ray. The Singularity is Near. Viking. 2005.