The Lamisphere

When I think about the future of the internet and “Web 3.0″, I generally don’t dwell on haptic displays or fully-immersive virtual worlds. Rather I think about the Do Button.  I’ve given a series of talks this year on emerging technologies that will change the world, and one of the main topics covered has been looking at how semantic tools will morph into stronger AI as they become process oriented.

In my world of 2012, the Google home page of the future looks amazingly similar to the home page of today.  The Google search button is relabeled “Do It”, and the logo now sports a techno-color “Agent” appendage.  While the number of pixels changed is relatively small, the implications to a web-engaged society is enormous.

Behind the “Do It” click is a semantic processor that takes the natural language command (e.g. Schedule a date with Kathryn for tomorrow night) and determines the context of the sentence:

• Who is Kathryn?
• What is a “date”?
• When is tomorrow night?
• Is Kathryn available?
• Where is this request to take place?
• What kinds of dates do Jonas and Kathryn like to conduct?
• What needs to be scheduled in order for them to have a date?
• Is a baby sitter available?
• Is a table available?

From there, the process engine goes into action.  Each task is executed through the complex decision-tree of “date scheduling”, ultimately resulting in a text message to my iPhone “Date with Kathryn scheduled, click here for details”.

How the tasks occur may be based on training done to the system on a per user basis, but more likely is based on a collaborative training scheme where you can have your agent execute tasks that were learned by someone elses’ agent in your trusted network.

I’ve pegged the emergence of the Do Button at 2012.  But recent events make me think I have over shot by 3 years.  I’m keeping an eye on a certain Stealth Company that might be scheduling dates for my wife and I within the next 12 months.

In: Artificial Intelligence, Google, Just Plain Cool, Semantic Web, Siri | | #