There is no way to charge for content - you have to rely on the user both having flattr installed and being willing to give you something.
As a content producer you have a direct monetary incentive to not link to or recommand others who also use flattr, as it lowers your cut. The service thus have to work against network effects - the more people who use the service, the worse of I am.
At the same time there is no point in installing it on a website, so it has to fight network effects just to get accepted.
Those are smallish problems. At first, you're only getting signups and payclicks from the very generous and conscientious -- early adopters. But that starts the process of making a larger group aware of the mechanism.
At some point, you throw some switches that give those early adopters some other slight benefits -- single-page articles, skipped interstitials, early-access. That nudges a slightly larger, more self-interested group to signup.
Any ad-supported site already has disincentives to outlink -- the reader leaves! Other sites will get the impressions! But they still do outlink, because ultimately audiences value usefully-outlinked content.
But also, every Flattr site is an advertisement for the network. When 99.99% of your visitors are not yet Flattrers, your top goal is to increase signups, not slice a small pie even more finely. Sending people to other Flattr sites best achieves awareness/subscription growth.
And, if that became a real problem, the Flattr central command could incentivize in-network linking with partial backpropagation of Flattr shares.
I think you missed the point with the first comment. It's a show of appreciation, not a charge for a service.
As for your second comment, I think you're mistaken. With the same reasoning, one might think that developers have no incentive to make it easy for users to stop using their application and migrate their data to a competitor's, but this is actually a good thing to do. Google in particular is known for this. I would personally be more inclined to send cake to a site that links me to other good content.
Flattr shouldn't be evaluated purely based on hard, rational economic principles; it's more of a social effect, so it's more complicated than that.
There is no way to charge for content - you have to rely on the user both having flattr installed and being willing to give you something.
As a content producer you have a direct monetary incentive to not link to or recommand others who also use flattr, as it lowers your cut. The service thus have to work against network effects - the more people who use the service, the worse of I am. At the same time there is no point in installing it on a website, so it has to fight network effects just to get accepted.
I doubt this will be around for long.