The Lifecycle of Software Objects by Ted Chiang

The Lifecycle of Software Objects
Ted Chiang
Subterranean Press, 2010
U.S. trade hardcover, first edition
ISBN 978-1-59606-317-4
150 pages; $25.00

Ted Chiang is one of my favorite writers. He only writes one short story, novelet or novella a year, it seems, but every one is a masterwork. A year in which Chiang’s name does not appear on every award ballot means that he’s skipped writing for a year. If you haven’t yet read Stories of Your Life: and Others, I strongly urge you to do so at once. This is what brilliance looks like.

The Lifecycle of Software Objects is the first work Chiang has written that fails to live up to the high standard he has set for himself. He has hold of an idea that deserves a full novel: what duties do humans owe to the artificial intelligences they create? But rather than write that novel, he has forced his work into the confines of a novella. It shows every time he tries to demonstrate that time has passed with a clumsy transition: “The year following Blue Gamma’s closure involves many changes for Derek.” “Another year passes.” “Two more years pass. Life goes on.” The seamless writing Chiang’s audience has come to expect has gone missing.

Chiang posits that humans have developed a software program that gives buyers a pet – one that never requires walking in the rain and doesn’t cry to be held the way a baby would. But the program requires a significant investment of time and energy, as this software has the ability to learn. Each digient, as these creatures are called, develops its own personality based upon its interaction with its owner. They are able to speak, and their ability to do so develops much the way a human child’s would, that is, through interaction with humans. Digients are great fun for a lot of people.

Who then, as it turns out, tire of them. Digients become yesterday’s fad. They are turned off, the same way you can turn off any other computer program, and left in stasis. Fortunately for the digients, they don’t experience anything while they’re turned off, and the only way they notice is because they do not share experiences with other digients who have been kept running continuously. Nonetheless, the fading of the craze for these creatures becomes more and more troublesome, particularly as the original platform on which they were built to run becomes obsolete, and there is nowhere left for them to play.

Worse, some humans have taken to abusing digients for sexual pleasure. This bears an uncomfortable resemblance to child pornography and/or prostitution; but do those concepts make the slightest bit of sense when applied to software programs? If you program the digients to enjoy their sexual exploitation, does that excuse using them in such a way?

Chiang has hold of a great many questions about artificial life. Steven Spielberg’s treatment of the subject in the movie, “A.I.: Artificial Intelligence,” is cursory in comparison. As usual, Chiang is playing with deep philosophical concepts.

The problem is that the concepts are too deep to be convincingly explored in such a short space. Many of his fans have long wished that Chiang would write a novel; this idea would have been the perfect vehicle for him to attempt his first. As it is, The Lifecycle of Software Objects is the longest work Chiang has ever written. It should have been longer yet.