IN my opinion, rather than RDF or XML leading to the sort of convenient result described in the mother's medical arrangements example, I believe that an AI application will be needed to heuristically extract associations between content and add them to databases, in other words an extension of what search engines already do for e-commerce. Human labour will not be an effective or economical means of achieving the goals outlined. The application will learn about the content independently of the user and generate the kinds of relevant associations between things that we do ourselves. However, unlike our brains the system will not be affected by the boring and voluminous nature of some associations and details.
This implies the tracking of just about every action done on the user's IT system, plus as much other data (GPS, textbooks etc.) as can be accommodated. The AI core will also need to have quasi-human linguistic ability. I suspect that this aspect of AI is the key to the whole prcess, as effectively whatis needed is an intellgence capable of indefatigable research and administration based on it.
Since the forties, if not earlier, we have had several predictions of the cybernetically controlled house -- the fridge that replenishes itself by ordering fresh food supplies etc. That such things have not come to pass to a significant extent demonstrates the importance of the economic factor and the need for an ingenious programme that will be able to complete the vast data accumulation and other processing tasks required.
I believe also that imminent improvements in memory will favour this development. |
Previous Message
