By all accounts, Google is a company that values creativity. They hire the smartest people and insist that they spend 20 percent of their time working on things they want to work on, their own pet projects. The idea is that some of these project ideas will be brilliant and Google can use them to make even more money.

Makes sense, right? Bob Cringely doesn't think so. Too many brilliant people working on too many brilliant projects, it turns out, is a bad thing because, well, there are just too many. From 400 to 4000 each year, by his calculations.

Google quite properly will pursue 10 projects per year and five of those will fail both because they are expected to and also because they were never worth pursuing in the first place. This leaves 390 orphaned projects of which 35 are absolutely stunning but unrecognized and 355 are pretty darned good. What happens to THOSE ideas?

You can't keep a good idea down, he argues, and eventually the Google employees will take those orphaned ideas away and generate new companies that will compete with and beat Google at their own game.

It's an interesting argument, but when that starts to happen, I would counter-argue, Google will just buy them up.