True information management: the next big thing

Because the technology market is in the doldrums, people aren’t upgrading their IT infrastructures – and the applications they’re using today are by and large the same ones they acquired three years ago. What happened?

I would like to suggest that we have effectively come to the end of the first wave of the information age. We have essentially reached the frontier of data management and are looking for true information management.

What I mean is that there is a fundamental difference between information and data. We have been acting as though they are synonymous, but information is really “data in context.”

Our systems have been optimized to manage data. Technological advances have been devoted to storing and accessing data. Information generation, such as it is, is still left to the minds of the users.

This is why Web browsers are so unsatisfying to anyone who uses them regularly. They are mindless miners of data. There is no context generated. No exclusion of clearly unrelated Web sites from those that might be pertinent. Most importantly, there is no summary that simplifies the delivery of important points.

Most of our technology now is very difficult for non-IT people or children to use. We would like to think a little old lady in tennis shoes could use a Web-enabled PC to communicate with grandchildren, but in fact she won’t, because the technology is not as easy to use as a TV set nor is it reliable enough. What happens to the PC the first time it blue-screens? Back to the knitting, presumably.

I would like to suggest that the next big thing for IT will be true information engines. I am not talking about databases that support Boolean searches, and I am not talking about ever more complex, interconnected office suites. Nor am I talking about data delivery in a medium-independent way. These are all old paradigm notions that don’t further the “utility” of IT technology.

I am suggesting that what we need is a truly intelligent agent that can accept context-independent, natural-language queries and translate those queries into actionable information. These intelligent agents would depend upon a new kind of interface and would be much easier to use than today’s. As a front end for traditional systems, they would make IT both more reliable and much more user-friendly than has been the case.

Interestingly, such agents would also consume tremendous amounts of computing power and memory. (There will finally be a reason for buying that new P8 processor with 10G bytes of RAM, 2 terabytes of disk and a clock speed of 10 GHz.) They would be supported by high-bandwidth connections to data sources. Initially they would be expensive enough that only businesses would be able to afford them, but everyone would want the capability.

When they arrive, we will really be in the information age. And network management will be essential for making the technology practical.

Intelligent agents would make very bursty demands on the network. Some hypothetical architectures for intelligent agents postulate a kind of intelligence-on-demand, where local workstations subscribe to intelligence as a kind of utility. For anything approaching natural interaction, the network will need to manage these sessions dynamically.

Is this vision unrealistic? I don’t think so. Sites such as http://www.kurzweilai.net or https://www.kurzweilai.net/index.html?flash=1 (for the nonflash version) make it clear that this kind of technology is very close to delivering real applications. For those who are ready, this technology will drive a huge increase in demand for IT. Managing such an environment will be challenging, but exciting as well.