• Goto NETFUTURE main page
  •                                 NETFUTURE
                       Technology and Human Responsibility
    Issue #53       Copyright 1997 Bridge Communications         July 16, 1997
                Editor:  Stephen L. Talbott (stevet@netfuture.org)
                         On the Web: http://netfuture.org
         You may redistribute this newsletter for noncommercial purposes.
    *** Quotes and Provocations
          Does Your Computer Have You in Its Grip?
          Intelligent Agents and the Economics of Meaninglessness
          The Net As Womb
          The Phantom Pursuit of Computer Literacy
    *** Toward the Great Singularity (Part 3) (Stephen L. Talbott)
          Are the mystics of the Net yearning for the past?
    *** About this newsletter

    *** Quotes and Provocations

    Does Your Computer Have You in Its Grip?

    If you're studying the human-computer interface and are looking for a research topic, I have a suggestion for you. Latch onto a few computers and fiddle with the operating system internals so that you can collect data about the timing of certain user actions in relation to various internal states of the computer. For example, you could have the system record the internal clock time of each keystroke or mouse click that follows a substantial pause (say, five seconds).

    You might expect that the collected data would be randomly distributed; that is, initiatory keystrokes recorded close to an even second or an even minute, according to the internal clock, would be no more numerous than those recorded close, say, to twenty-three seconds past the minute. (This assumes, of course, that the clock time is not displayed on the user's screen.)

    My prediction: with some people you will find some clusterings that deviate too far from a random pattern to be easily explained. In other words, you will find that human actions we assume to be freely initiated can in fact be "entrained" by the computer in ways not currently easy to understand.

    The key requirement of the experiment is to record human actions whose initiation is independent of any identifiable system prompting, and to correlate those actions with various invisible transitions in the internal logical states of the computer, looking for more-than-random correlations. The experiment should not be run in a laboratory or any other special setting, but rather should track users in the course of their normal jobs. Many individuals with different character traits should be checked. Also, the experimental subjects should not know that the timing of their actions is at issue.

    A lot of possible correlations may need to be explored, and getting good data may require some ingenuity. Needless to say, if my prediction should turn out to be correct, there would be profound implications.

    Intelligent Agents and the Economics of Meaninglessness

    The Economist recently pointed out the bind that developers of intelligent shopping agents are in. When comparison shopping becomes widespread, prices will be driven down to the uniform levels typical of commodities,