As follow-up to this post, I read a great piece by Richard Price. It describes the failings and costs of the traditional scientific publishing model, and why these are driven the development of new metrics to track academic success.
To break out of the tragedy of the commons, new reputation metrics, developed by a number of startups, have been developed that incentivize scientists to share their research openly, rather than incentivizing them to put their research behind a paywall. Scientists are adopting them to better stand out from the crowd when applying for jobs. Examples of these new reputation metrics include inbound citation counts, readership metrics and follower counts.
He predicts that most scientific journals will eventually disappear, as scientists' reputations depend less on where they publish, and more on their public profile.
I'm not as sure of this last statement as Richard is. I feel most Journals will move to an open access model and will gradually lower their up-front publication charges.