I read an interesting article in IHT about new technology developed by Xerox subsidiary Powerset. It works on natural language technology that will make computers understand human language. This could be applicable to search engines and to other types of software, IHT article states that new “natural language” search engine will challenge Google.

I have no idea how likely it is that new technology will be available soon. However if this happens, it would imply a major productivity acceleration felt in the global economy. Imagine that instead of typind a letter you quickly say few key frazes to your computer and add “please do it in a very formal style”, computer does the rest, formats it, prints it or even automatically sends it, while you are on your way to the meeting with next client. Or instead of spending hours to search for data, you say to you computer, find data on global sales of widgets, put it into Excel format and make three charts in Powerpoint presentation (or whatever will be out there to replace Excel and Powerpoint).

It takes me back to speed limit discussion again. I have argued that economists predicting that global economy speed limit can go down in next few years (or that speed limit in US has been on decline recently) are wrong, wrong big time. Global economy empowered by ICT is getting very good at knowledge sharing, and knowledge multiplies by sharing. It means that the pace of innovation generation will increase massively in the next few years. Speed limit will go up, simply our statistical offices are not able to measure growth and inflation properly. For example in a recent paper Robert Gordon estimates that the upward inflation bias in USA has recently been at least 1 percent (and that Boskin Commission also underestimated the bias in 1995-96). With further rapid ICT deepening and production and services globalization this bias is likely to rise even further.