The standard story of history (including the one in Wikipedia) says that James Watt invented the steam engine, or at least a significant part of it. If you look carefully, however, you'll find that he merely improved it somewhat and patented it, used the patent to build a monopoly, used the monopoly to drive the real inventor into bankruptcy, used his power to stifle innovation in steam engines for years and years. Some tried to fight this but lost in court. After Watt's patent expired, improvement started again, and the Industrial Revolution begun.
In 1811 Cornish miners began publishing Lean's Reporter which contained the "source code" for building new steam engines -- plans with enough detail that others could copy and improve. They refrained from getting patents on their new inventions. 186 years later, the same thing happened for web servers and other programs.
Two charts: graph of the increasing power of steam engines, graph of the market share of Apache.
Quick poll: all of you put something in to Wikipedia and all of you get something out. On balance, in terms of benefit and effort, do you think you put more in or get more out? (Only a couple people say they put more in, the whole room says they get more out. Studies find the same in the free software community. Therefore, there's a significant degree of self-interest. Asked the same question about others, the only half think others take more than they give. Of course this disparity isn't possible in reality -- unless something modifies the valuable in between...
Imagine a typical barter system: Alice gives up a fish because she doesn't really want one, Bob gives up the potatoes because he has too many -- so they're both beter off. Another possibility is a price market. Another possibility is to combine the two into a tribal cooking pot, creating a soup that's worth more than what's put in. How do you divide the soup? Tribes have a concept of reciprocity -- you take in proportion to what you put.
Economic goods are non-rivalrous, so instead of dividing the pot, everyone gets all of it. That's why it's possible for others to be so much more generous: because you get so much out despite the comparitvely small amount you put in, so you imagine there must be someone doing all the rest and not getting any.
If you look at the money created through such processes, it's enormous. Debian created around $14B in value. Many people criticize this by saying they just copy commercial companies, just like people say James Watt was a grand inventor, but that doesn't seem consistent with the facts -- in many areas the free software is the source of innovation. Plus it's discovered a new organizational structure for innovation.
The project is run through self-organization. If you look at projects on SourceForge with more than 10 developers, especially the projects with more than 50, there's an enormous amount of concentration: the top 10% of developers write 75% of the code. This power law is the same thing in every self-organized network, except it's much sharper in these projects because you lead by doing.
So why don't these top 10% capture all their value? Well, they might make 75% of the code -- but that'd be a three-legged chair. You need the bottom 98% of users to build the last leg. That's what makes the network sustainable.
The concept of collaborative knowledge isn't new, what's new is the scale and the tech.