by Flemming Funch
AndroidTech, Web 3.0 - You Ain't Seen Nothing Yet!. Isn't it a little early to try to say what Web3.0 is, when we haven't entirely agreed what Web2.0 is? I guess not. I feel that Web 3.0 will be characterized and fueled by the successful marriage of artificial intelligence and the web. Artificial Intelligence? Isn't that the kool-aid that the Semantic Web community is drinking? Yes and no. The technologies considered pivotal in the Semantic Web are indeed considered by many to have their underpinnings in artificial intelligence. But, most of the Semantic Web projects I've seen are focused squarely on the creation of, and communication between, intelligent agents that do the natural language and topical matching work in a transparent manner, behind the scenes, without requiring human intervention.
This approach may eventually be viable but I feel that it misses a key ingredient of Web 3.0 that will finally bring artificial intelligence to the forefront. Currently the vast majority of artificial intelligence is embedded in various niche areas of commerce such as credit card fraud detection, or the speech recognition application that converts your voice to text as you dictate a document, etc. The reason for this of course is that we are still decades away from computers that will have the incredible and flexible pattern recognition capabilities of the human brain. Well, right, yeah. Artificial Intelligence looked very promising 30 years ago, and it maybe still does, but I haven't noticed much progress, other than that faster computers can make something look a bit intelligent by brute force. But what he's thinking about is not really AI, but this kind of thing: The reason Web 3.0 will lift artificial intelligence into the limelight is it will fill in the technological gaps that currently hamper the key uses for artificial intelligence. It will do so by shunting out the parts of the problem that require a human being to human beings with the help of the web. But, it will do so in a manner that is transparent, massively parallel, and distributed.
Amazon has taken a unique and innovative step into this area with their Mechanical Turk web service. Yes I know this is the second time I've written glowingly about Amazon in regards to Web 3.0, but as a web service junkie you have to love what they are doing. The Turk service allows developers to shunt out the parts of their applications that require human intervention to a paid participating group of volunteer workers, in a manner that mimics a standard web service call. This creates a standardized platform for utilizing human pattern recognition capacity in a modular manner. Google is another company experimenting with something similar with their Google Image Labeler game. From the game page:
"You'll be randomly paired with a partner who's online and using the feature. Over a 90-second period, you and your partner will be shown the same set of images and asked to provide as many labels as possible to describe each image you see."
The players have fun and Google gets thousands of images tagged with relevant text labels. Wow, I have played that, and I didn't even get the point. That's brilliant, of course. Use human intelligence for something productive, have fun, or get paid for it.
That might somehow become an important trend. But that's essentially a Web2.0 kind of thing, to individually tag stuff and to collectively produce a bunch of ordered stuff.
I'd hope Web3.0 would be a cyberspace kind of thing. I mean cyberspace as flying around in multi-dimensional space and all the data you need and want being layed out visually in a way that's useful to you. "Computer, show me all the people I know, sorted by hair color!" But we're kind of far off from that, as computers still operate with a desktop metaphor, with filing cabinets and documents. And even 3D environment mostly just try to look like the regular 3D world with added improvements. Looking at a webpage on a virtual screen in a virtual world doesn't make it any better at all. We need something with more dimensions, and where you actually can look at enormous amounts of data in a useful and intuitive way. But maybe that's too much to ask, and maybe that'll be Web9.0 or something.
|
|