Tangled in the ThreadsJon Udell, December 21, 1999
Where We Go From HereWaxing philosophical at the end of the millenium
After next Saturday, most of us don't figure we'll see the dawn of another century, let alone another millenium. But one of this century's most accomplished technologists -- Ray Kurzweil -- isn't so sure. In a radical book entitled The Age of the Spiritual Machines, Kurzweil argues that computational substrates to which we can "port" human minds will exist within a matter of decades. Mind transplants are routine science fiction fare, but Kurzweil's deadly serious when he says that if Moore's law holds until 2020 it will produce a platform on which we can run the application we call consciousness.
Kurzweil likens the process to the old story in which the inventor of chess asked the emperor of China for a grain of rice, doubled for each square on the board. Halfway through the process, the emperor is doing OK: he only owes the emperor a few dozen bags of rice. Then it gets out of hand. We, says Kurzweil, are at that halfway point in a game that doubles computational power like those grains of rice. At square 31, you get discontinuous speech recognition. A square 32, another doubling, you get continuous speech recognition. At this point in the game, though, the emperor never guessed that square 64 would cost him more grains of sand than the universe has atoms. We're at the tipping point, says Kurzweil. Even the acceleration will accelerate, because machine intelligence doesn't let the end of the Moore's Law era slow it down, it just jumps into quantum-computing and other substrates.
There are three ways to evaluate Kurzweil's thesis:
- He's completely wrong. The mind isn't software, the brain isn't a computer, and it will never be possible to transplant brains to computers. This is a perfectly reasonable position, but I don't buy it. I think consciousness is computation, and I don't think the soggy meat between our ears is the only place it can ever happen.
- He's half-right, half-wrong. The mind is software, but the reverse engineering of that software is not the sort of problem that will yield to the brute computational force that we'll have available in the foreseeable future. This is another reasonable position, and it's the one I do buy into. Back in the 50s, the first generation of AI researchers tacked some hard problems, including chess and natural language recognition. Few would have guessed that forty years later brute force computation would have cracked chess wide open, yet made hardly a dent in the problem of natural language understanding. Language is central to our mental lives, and we are still clueless about how it really works.
- He's just plain right. Although I don't buy that, I don't completely rule it out either. Kurzweil's no crackpot visionary, after all. He's a prolific inventor whose ideas have become seminal products -- music synthesizers, text readers, dictation systems. A guy who solves engineering problems like these, leaving whole new industries in his wake, deserves to be taken seriously.
Before Moore's law runs out of gas, Kurzweil thinks, we'll have the storage to represent all the states of the brain, and the cycles to emulate all of its state transitions. If we can't figure out what all those states and transitions are, then we'll brute-force the problem with "brain scanners." One way or another, the brain's slow but richly parallel processing power will be matched, then exceeded, by alternative mechanisms.
If the network is the computer, what is the Web?
We humans have another trick up our sleeves, of course. It's not only our individual brains that exploit parallelism. It's our culture, too. Although the Internet is in principle a massively parallel network computer, projects like Seti@Home and distributed.net haven't yet begun to transform computing in anything like the way the Web has begun to transform human culture. The dynamics of this transformation are explored in another recent book, Susan Blackmore's The Meme Machine. Blackmore is a disciple of Richard Dawkins, who in his book The Selfish Gene argued that the true engine of evolution isn't the reproductive behavior of species, but the replicative behavior of genes. These replicators drove biological evolution for millions of years but -- Dawkins noted in passing -- DNA is not the only possible substrate for replication. With the advent of human culture there appeared a new medium for replication -- not of genes, but of ideas, or images, or tunes, or techniques. Dawkins suggested that competition among these "memes" kicked evolution into a higher gear. Blackmore elaborates on this idea: memetic replication drives cultural evolution far more efficiently than genetic replication drives biological evolution.
From this perspective it's the Internet's ability to link minds, not merely computers, that emerges as its most important innovation. Computers may be marching relentlessly up the curve defined by Moore's law but we, as a species, aren't standing still either. We're augmenting the massive parallelism within our own brains with a new kind of massive parallelism that can link many minds into collaborative relationships.
Of course it's true that science, literature, art, music, and all other cultural activities have always been processes of meme replication. The printing press was a major catalyst enabling these processes to run faster. The Web is an even more powerful enzyme, whose effects we have not yet begun to fully appreciate.
Big ideas, modest goals
I'll come right out and say it: I think that the Web, or more precisely the human collaboration that emerges through the medium of the Web, is a new phase in the evolution of our species. That's the big idea that prompted me to write a book about Internet groupware. (If you're going to spend a year writing a book, you might as well pick a big idea to write about!)
Now, with the exception of millenial columns like this one, I don't spend a whole lot of time thinking about brain transplants and meme replication. But I do marvel, every day, at the ways in which the Web enables groups of people to form, exchange ideas, and work together to solve problems. I'm amazed that we can not only read but also write hypertextual documents, and that we can find these documents when we search for them. And I'm privileged to have been able to explore this new culture of online community with those of you who have, for many years, congregated in BYTE's comfortable and ever-enlightening corner of cyberspace.
So where do we go from here? Well, for starters, I can hardly wait for the broadband revolution to make telepresence as routine as email. As I write this words, I'm descending into Pittsburgh on the next-to-last leg of a grueling 28-hour business trip wrapped around one meeting: three delays, two cancelled flights, one huge headache, and a crater in my work schedule. Will I be able to upload my mind to new hardware in 2020 and write another millenial column in 3000? At the moment, I couldn't care less. My goals are a lot more modest than that. I'd settle for being able to project myself across the country in 2001 without having to suffer the slings and arrows of US Airways!
Jon Udell (http://udell.roninhouse.com/) was BYTE Magazine's executive editor for new media, the architect of the original www.byte.com, and author of BYTE's Web Project column. He's now an independent Web/Internet consultant, and is the author of Practical Internet Groupware, from O'Reilly and Associates.
This work is licensed under a Creative Commons License.