The hairy assumption is that people want to do this, and that they're going to do a good job. There is some value to me if I use it myself; if you used it and I cared what you thought, it would be substantially more valuable for me.

This is the idea of the read/write Web, ... The Web is there for you and me to annotate it, share it with our friends, share it with our business associates, to add a layer of insight and personal understanding.

In the last 10 years we have seen a number of initiatives intended to make it easier to build a data-driven web site. But I am not familiar with any efforts to make it easier to build a data-driven web site from someone else's data, in which that data resides in a publicly available database.

In most cases, mobile users are searching for spur-of-the-moment data, like a local restaurant or directions.

Search streams - the logs in which the contents of users' search requests are recorded - are a wealth of data about what people want and how they express their desires. The owner of such a search stream possesses a powerful asset, once user attributes may be calculated to determine how they relate to the target market.

[Desktop search] is very valuable for corporate workers, ... Once they understand [how to use] desktop search, [workers] can't conceive of a time when they did not have it.