As IDC, a leading research firm, has observed, a big part of the net's near-term future is platform-independent 'smart content' that informs all business and consumer processes. This will only grow in importance with the explosive growth of delivery mechanisms (e.g. wireless, PC, Internet appliance.) Hoping to catch this wave, many current firms are marking digital data with XML, thus embedding knowledge in text in a simple, pragmatic way. But XML, like relational databases, is just a general framework for structuring knowledge; it doesn't in itself impart any meaningful structure to data. It's only as useful as the humans who've spent their time figuring out how to apply it in each particular case. Webmind understands the content of text files and databases, and outputs this knowledge into XML files or databases, or uses this knowledge to provide information services or answer human questions directly. Furthermore, as time goes on, it evolves more and more meaningful understandings of the data i t digests. What makes content smart is not XML or any other representational formalism, it's the fact that the content has been processed and structured by a mind human or WebMind.

Similarly, the current information retrieval space is dominated by simplistic statistics-based tools which simulate text understanding. The performance of these systems, though good compared to standard key word search engines and SQL database queries, remains well below the standard necessary to adequately mediate information transfer in the transactions-based Internet economy of the near future. For example, many billions of dollars have already been lost due to the failure of technology to adequately deliver the information that secures and supports web-based transactions . With the Gartner Group predicting that fully 30% of the US economy will be transacted on the Internet by 2005, the problem will worsen dramatically. For the next generation Internet, the critical bottleneck is no longer computing power or data transmission. It is the extraction and delivery of appropriate data and, more importantly, the ability to infer and act on information from both structured and unstructured data.