I just finished Emergence by Steven Johnson, the cofounder of the now-departed online magazine FEED. The book jogged my brain, helping me to slip back into a world of interesting theories of information and software (a world that had seemed to be fading in the aftermath of the dot-com crash). Johnson aims to explain how up-and-coming software programs will be built upon models of emergence, where organization and coherence emerge from the bottom, wiggling their way upward instead of being commanded from above.
My hope was that Emergence might give me some language for explaining how search engines are making knowledge out of the spray of loose information on the Web. And there were moments in which I thought to myself, hmmm, think of crawlers as software agents that learn as they crawl, using one link to glean information about the next link, and so on, without a central commander telling them what types of information to pick up. Search engine filters take the next "emergent" step, by ranking and indexing those links without any human being providing a blueprint for the kinds of pages that should receive top billing.
But although the concept is fun to play with intellectually, I'm not convinced that it applies as neatly as I'd like. At what point are the rules (written by humans) doing more of the guiding than the agents (the software robots) themselves? Worth pondering as I consider how much artificial intelligence is making its way into search-engine territory.
3:48:22 PM 
|
|
