Thursday, April 15, 2010

Review: "Rapture for the Geeks: When AI Outsmarts IQ", by Richard Dooling

The cover of Rapture harkens to the days of green text on a grainy black monitor: the highlight is an electric smiley winking back at you. What’s behind that smiley? Does the almighty computer know more than it’s letting on? Not likely, you say…but will it, and if so, when? And what’s going to happen then?

These are the questions Richard Dooling attempts to explore in this engaging essay on the singularity: the point in history, predicted to happen soon, when the capacities of computing machines surpass the capacities of us.

Not a tame subject, and Dooling knows it. He isn’t afraid to go places with it, though, places that are sure to offend some, amuse others, and give everyone in-between pause for thought. The book is stuffed with technological tidbits…did you know the capacity of the human brain is 10 petaflops? That IBM created a machine three years ago that runs at 3 petaflops? That NASA’s newest is currently receiving upgrades that will bring it by 2012 to 10 petaflops—the same as the human brain?

That computer hardware will match us is not in question. What Dooling concerns himself with is more esoteric matters of the singularity—for example, will the software ever be written that can mimic the human condition? That bit is not so clear, and he cites the opinions of many experts in the various affected fields to argue yay or nay. There are religious or spiritual considerations as well, and these are delved into also. After reading some chapters of this book, I was beginning to see reality itself in a different light. Just what is reality, anyway?

And what about consciousness? Let’s assume we invent a machine that can match us in every way we can think (pun intended). It acts like us, dreams like us, has wishes and desires like us. Maybe even pack it into an android body. Okay…is it sentient? Does it, or should it, have rights the same as us?

Does it have a soul? Do we?

The mind reels at all these considerations—and Dooling delights in reminding us that we may need to be making these considerations in the not so distant future. (I did find it disappointing that he didn’t mention the very latest in experimental computing, specifically, the use of organic brain tissue. That’s a whole other can of digital worms there. Fifty years from now when the time comes to throw out your old toaster, is that toaster going to want to go?) Aside of all that, however, this book is just fun reading for your inner geek. Consider one prediction for the future in which computers, having surpassed our intellectual abilities (the logical follow-up to the singularity), decide that man isn’t actually needed. Didn’t Asimov write about that? Or Bradbury? This time around, though, the story may be for real.

The book rambles into other areas as well. For example, the fragility of modern-day documents. Why is it we can read documents carved in stone 3000 years ago, but ones placed in Starwriter format (or whatever) on big reels of tape (or whatever) 20 years ago may be gone for good? Not only does the hardware media fade away into obsolescence, so do the proprietary file formats we store them in.

Lastly, the book is peppered with informative, ironic, or hilarious quotes from near and afar. If nothing else, I’m glad I read it just to collect a few of these gems.

All in all, a thought-provoking, interesting read.