Part 3: HAL 2.0
The next session, Human Being and the Machine, featured Patti Maes from the MIT Media Laboratory, who works with software agents that carry out tasks on behalf of their "owners"; and Ellen Ullman, an engineer and writer who takes the position that the complexity of life and thought goes beyond simplistic reduction. Maes discussed the state of the art as practiced in her lab, including agents that can offer contextual advice as you view e-mail and other documents based on previous similar information or e-mail recipients and senders; agents that can be attached to objects, like a coffee pot or refrigerator; a browse-along agent that scouts down paths not followed; and other kinds of agents that can assist in retaining the logic of solving a problem to be applied toward similar problems faced by other people.

Maes maintains that the right of control over agents has to be left strictly to individuals and carefully watched. Some agents they developed were too autonomous, producing unwanted results or uncomfortable users. She said that agents can build a tapestry that helps one understand oneself better by externalizing and tracking interests over time.

Ullman's talk consisted mostly of a rebuttal of Ray Kurzweil's recent book, "The Age of Spiritual Machines." In his book, Kurzweil predicts a rapid coming of age of computational power and complexity that will cause artificial intelligence to become a reality and allow people to download their mentality into computer systems. At first, Ullman's talk seemed overly specific, but after several days of reflection, her thesis stands out in stark relief separate from Kurzweil's utopian (or dystopian) predictions.

Engineers, Ullman said, reduce complexity to simplicity to solve problems. But in the process, they eliminate nuance. "Why do we mistake the systems we build" for the thing that went before, she asked. "Why do we mistake the pointing finger for the moon?"

Kurzweil reduces all aspects of human consciousness to implementation problems, she said. But memory is unlike a computer file, she maintained. "Memory will turn out more like weather" when we more fully understand it.

Ullman contended that complexity will always evade our ability to pin down the brain into bits and bytes, regardless of the underlying chemical basis.

In a very funny sequence of ideas, she noted how Kurzweil dispassionately discusses creating duplicates of oneself through downloading one's mentality into one or more places. She said, "the new cyber-being, the fearsome doppelganger, [Kurzweil] calls a simple backup" as if the doppelganger weren't one of mankind's oldest nightmares.

(It's Ullman's speech that continues to stick with me most of all and gather "moss" in the days that follow the event, as I see more and more application of her paradigm of "engineers as over-simplifiers" all around me.)