AI & UI - Thinking beyond...

Can we build code that can think on its own?
For example, can we build a component that can learn from the inputs that it gets?
I was trying to build a component that can handle additional events based on user inputs. This component can adapt to user's requirement by changing from a 2d to a 3d component on its own.
UI has come a long way and today's user's expect the UI to hold the data within it and expect the UI to "analyze" this data and display appropriate responses through data visualizations to tell the users what action to take!
Talking about 2d to 3d and vise-versa, I believe the marriage between flare and papervision could create the best of the 2d and 3d worlds!
Building a new component that takes ideas from flare and papervision and has Artificial Intelligence embedded makes it mouth watering!

Original = Light weight = High Performance

The fact of the matter is we as developers look for great open source frameworks and libraries that can make our lives easier.
The vision should always be on scalability and performance but for some reason all these free libraries are making us lose focus on the plain fact that we can build ultimate stuff that can be much better than all these that exist for free.
The problem is using these libraries limit our ability think beyond what exists!
The idea is that by creating a component on our own using simple logic creates unbelievably great stuff! All the actionscript code for 2d, 3d and data visualization are great but there is much more than we see.
All I suggest is enjoy creating stuff than just re-using the "elephants"!!!