Distributed Ecs of Scale
Jeff makes the point that peer-production mechanisms are hardly perfect - Wikipedia's easily gamed. I think this is an important point to discuss.
It's not that peer production is perfect, it's that it's good enough. We could call this satisficing vs optimizing just to be nerdy. In other words, in many domains, the cost (or price, if you like)/value ratio is as good or better than pro production.
So Wikipedia's not perfect - it can be gamed easily, as Jeff points out. But on average, the cost/value ratio is superior to, for example, the Brittanica. Similarly, a Next Big Thing will be peer production of micromedia. Blogs were the first wave. Podcasts, vlogs, etc are the second. Movies and music will be the third. In each case, the cost/value ratio is superior to pro produced alternatives in certain domains.
Take Jeff's blog - I read it because I get a huge amount of info not only more cheaply, but also higher value info, than I would from reading a bunch of publications about cool new things.