I've heard some Amazon employees say something similar (but maybe not as enthusiastically) about Amazon. One person described their structure to me as "like terrorist cells": it's understood (and encouraged) that 3 different teams are all trying to implement the same thing, in completely different languages and with completely different interfaces. They might not even know the others exist, until they ship. Hopefully at least one will survive.
In some cases, more than one does. Amazon has at least 3 different data stores you can use, last I saw.
Theres a view in economics that you 'can not play market' within the boundaries of the firm. That is to say if there is an overseeing committee that can end a project prematurely then you are not really having market forces operate 'within' the firm and are doomed to failure. There have been a couple of companies that have tried to do this. It is documented in the following book by Nicolai Foss:
http://www.amazon.com/Strategy-Economic-Organization-Knowled...
Isn't it ironic that you say Amazon's strategy is "doomed to failure" at the same time that you link to a book for sale on Amazon?
The company(s) mentioned in the book still exist, but they no longer try to 'play market'. The more incentive you give people/subgroups to compete with one another, the more incentive they have to not share information(best practices) with one another. Managers see this and reallocate resources. Groups feel alienated that their resources are taken away from their idea to fund someone else's idea and then they go back to keeping their real ideas to themselves and/or starting a separate business outside of the parent firm. Innovation in the mentioned companies increased dramatically when the 'play market' strategy was initially adopted though.
I also heard Enron was much like this (amazon + the defense co).
There's a certain 4 trillion dollar financial services firm I know. They operate like this at a country level: each country is graded against the others on an elaborate score card. The idea is that when a country finds an innovative way to "win" one year, althe others can use the idea the next year.
Only, the winning countries try (nicely) to sabotage the sharing process. For example, Country "U" outsourced part of their self-service web applications to a nice fellow I know living in Country "C." Although it is not in writing, the understanding is that if he so much as has a coffee with anyone from that company in Country "C," his home town, he is out of work.
Furthermore, the Country "C" guys don't really want to hire him, they're afraid he'll report what they're doing back to Country "U," so when they ask him out for coffee, all they really want to do is dangle contracts in front of him while pumping him for information under the guise of establishing his experience.
It sounds like a cheap spy novel, but it's business as usual when some bright person at the top decides that a little competition is healthy :-)
Amazon has multiple paradox resolution teams competing to negate each other's efforts.
Entirely possible that my previous company would have never lasted had they not been bought. Perhaps those opposing internal forces would be too much to bear. But it did work for a while.
That actually sounds like fun. Friendly competition indeed :D