 |
|
If Microsoft's demented "Office Assistant" is any indication, the prospects for things starting to think are pretty dismal. The amount of time, mouseclicks, and aggravation required to get work done in Microsoft Word has exploded since Monsieur Paper Clip arrived on the scene. Redmond's concept of "things that think" reminds me more than anything else of the proto-bird Archaeopteryx falling out of Jurassic fern trees in barely-controlled crashes. Personally, I only want to be cornered into interacting with things that think after they have become really, really good at it. Luckily, in spite of its title, this book has little to do with artificial intelligence (or artificial stupidity, as the case may be) -- it should have been called "When Things Start to Compute."
As computing power, memory, and bandwidth have become cheaper and cheaper, the incorporation of embedded microcontrollers into mass-market artifacts has become the norm, and the networking of these objects is following closely behind. Over time, this will have the effect of removing many of the hard edges from the world around us. When everyday objects can compute, store information, communicate with each other, and adjust their behavior to the user and the situation, the surface effect will be that the environment will adapt to our needs and support our goals, instead of us always accommodating to the environment.
I just bought a new car for the first time in 15 years, and was astonished at the hundreds of ways designers have taken advantage of embedded processors to make automobiles more tolerant of human frailties and more pleasant to use. For example, if you accidentally leave the dome light on, but lock the car and walk away, the car will eventually turn the dome light off anyway, preventing the battery from running down. Or if you have the radio turned up much higher than normal when you shut off the ignition, the radio will come back to life at a midpoint volume setting when the car is next used to avoid startling the driver. And anti-skid brakes are phenomenal!
The absorption of computing capabilities into the environment also allows us to push information "into the world" that we would have previously been forced to carry around in our heads or on stacks of paper. Smart devices can then deliver this information back to us whenever and wherever we need it, remind us of tasks when we are in the appropriate context, and take care of routine chores on our behalf. The use of email and calendaring servers by many folks as a sort of "filing cabinet" that can be accessed from wherever they happen to find network connectivity is only the first and most basic instantiation of this trend.
For example, I feel certain that future generations will be incredulous at the concept that someone would have to periodically look in their refrigerator, try and deduce what isn't there, write down a list, drive to the store, pick those items off the shelf, and tote the groceries home again. How much more sensible it will be for the refrigerator to keep its own inventory and communicate with our scheduling programs so that it is aware of our work hours and travel plans. The refrigerator can then send refill orders for milk and eggs to the market via the network as needed, and schedule the deliveries for a time when someone is at home to accept them.
Neil Gershenfeld works at the MIT Media Lab, a hotbed of user interface innovation since the 1980's, and in many ways the spiritual descendant of Xerox's Palo Alto Research Center (PARC). Interesting projects described in When Things Start to Think include Gershenfeld's collaboration with Yo-Yo Ma to build a digital cello, wearable computers that are sensitive to location and objects around the user, "smart paper," printable computing devices, quantum computers, and the evolution of digital currency. Although the book is parochial in nature, and intended for nontechnical readers, it is a pleasant read and presents an essentially optimistic outlook on the sweeping changes that are already infiltrating our lives.
-- Ray Duncan (duncan@cerfnet.com)
|
What are things that think? Books that can change into other books. Musical instruments that help beginners engage and virtuosi do more. Shoes that communicate through body networks. Printers that output working things instead of static objects. Money that contains behavior as well as value. |
Why should things think? The rights of people are routinely infringed by things, and vice versa. Dumb computers can't be fixed by smart descriptions alone. Useful machine intelligence requires experience as well as reasoning. We need to be able to use all of our senses to make sense of the world. |
How will things that think be developed? By taking advantage of what nature already knows how to do. Through research that maximizes contact, not constraints. With education that happens as it is needed, rather than in advance. By interconnecting systems of people and things to solve hard problems. |
| Readability |
|
| Originality |
|
| Organization |
|
| Accuracy |
|
| Consistency |
|
| Depth |
|
| Timeliness |
|
| Editing |
|
| Design |
|
| Overall Value |
|
Explanation of ERCB rating scale: No stars = unacceptable, 1 Star = marginal, 2 Stars = average, 3 Stars = above average, 4 Stars = exceptional.