(This is Part 8 of a series. Go back to Part 7.)
2a. The rise of machine common sense
The evolution of the world wide web is going to be one of the most fascinating processes ever. It's going to go places that we can scarcely even dream of today.
We're now at what's called "Web 2.0," which as far as I can tell entails using the web to foster social intelligence about various things.
The most famous example is Wikipedia, which by allowing people anywhere to contribute and edit articles has built up a very impressive encyclopedia, one that I use frequently to obtain information about various subjects. Just the other day I was delighted to learn about the "Five Good Emperors" of Roman times.
Another possible example is the prediction shops, in which people using the web can make bets about various potential events, and which have proven to be quite good predictors of what will happen—who will get elected, who will win the Oscar, things like that. Again, it's a use of collective intelligence to achieve a goal.
The Open Source software movement is a first cousin of this, because it involves groups of programmers contributing their time and effort to develop and improve a piece of software, the most famous example being the widely-used Linux operating system.
Another representative for Web 2.0 is social networking, embodied by sites such as MySpace, YouTube and Flickr, where people come together to socialize, share photos and videos and so on. Yet another representative is the "mash-up," where the contents of two or more websites are combined to create a third entity—real estate listings combined with Google, say, to create maps for each listed house.
Okay. Now we appear to be on the brink of what's being called Web 3.0. If Web 2.0 involved the collective intelligence and networking of human beings, Web 3.0 will be about the first tentative approaches to having machines being able to have some limited understanding of what it is that they're providing.
The current web could be described as billions of websites connected by all kinds of links. In other words, lots of data is connected or linked together but the computers that hold the data or connect it together have no "understanding" of what they're holding or connecting.
That's going to change—and Web 3.0 is the first step in that direction. To see how, let's contrast how Web 2.0 and Web 3.0 would handle the solution to a typical problem: designing a vacation package.
In a wonderful example provided by John Markoff of the New York Times, let's imagine that we want to go someplace warm, we have a budget of $3,000 and an 11-year-old daughter. In the current internet we'd probably be going for hours through search engine results, sifting through lists of airline flights and hotels and times and prices and accomodations and availability and so on.
In Web 3.0, in contrast, we might give the same details given above and receive back within a few seconds a complete vacation package that satisfies our criteria. in order to do that, the relevant computers would have to "know" what it means to have an 11-year-old on vacation with us, "know" what it means to want to go to a warm place and so on.
A good description for this is "common sense," and the term illustrates something that we humans do easily every day but which computers have had a really hard time with so far. For instance, because of our interaction with the world we know that the horizon is horizontal, that the sky is above the horizon, and that a cup placed upon a table will tend to stay there—all things that we don't even think about but which we have knowledge of.
In fact there are thousands, perhaps millions of such everyday facts which help us make decisions and form judgments, things which are part of our background knowledge but which computers don't yet "know." In Web 3.0 computers will begin to "know" and link up some of this background material—material which we humans take almost completely for granted but which is so vital to our rational process.
This new "common sense" knowledge by web computers will make it possible for them to offer much more sophisticated solutions to our challenges, such as helping us pick a college or design a financial plan or construct a home. Instead of lists of documents we will begin to get intelligent answers from the web, answers that will incorporate some background understanding of the world we live in.
(This is the end of Part 8. Go to Part 9.)
—jim sloman, for 1.2.07
|
