Are you too optimistic? Maybe. Let's consider some data, bearing on the question of Red States vs. Blue States.
1. For many years, the United States has been conducting a remarkable natural experiment, involving the random assignment of judges to three-judge panels. Here's one thing we know. On DDD panels (all-Democratic panels), Democratic appointees show VERY liberal voting patterns in ideologically contested cases -- much more liberal than on DDR panels. On RRR panels (all-Republican panels), Republican appointees show VERY conservative voting patterns -- much more conservative than on RRD panels.
One way to think about this is that if you create what might be considered "Red State" federal panels, you'll get much different votes, from "Red State" judges, than on "Red State/Blue State" panels. The same applies to what might be considered "Blue State" federal panels. A disabled person, challenging certain conduct as discrimination, has a much better chance before DDD than before RRR.
For present purposes, here's the kicker: For both individual Ds and individual Rs, the voting patterns are far more moderate on RRD and DRR panels.
2. A few years ago, I was involved in a little experiment in Colorado. Liberals in Boulder were assembled to discuss climate change, affirmative action, and civil unions for same-sex couples. Conservatives in Colorado Springs were assembled to discuss exactly the same issues. The liberals spoke only with the liberals; the conservatives spoke only with the conservatives.
The most important result of the discussion was to make the liberals more liberal and more extreme on all three issues, and to make the conservatives more conservative and more extreme on all three issues. Another result was significantly to decrease diversity among liberals -- and also among conservatives. (Note: I'm speaking here of the difference between people's anonymous predeliberation statements and their anonymous postdeliberation statements.)
In short, deliberation among like-minded people increased extremism at the same time that it squelched diversity of view.
For the blogosphere, the implications are not obscure. The blogosphere has many "Red State" sites and many "Blue State" sites (often created by polarization entrepreneurs). These sites look like RRR or DDD; they are similar to our Colorado experiment. To that extent, the results should be a) greater extremism and b) less internal diversity. A third likely result is greater contempt for those on "the other side."
True, you're right, exposure to competing views can deepen division (especially if we're called names). True, many people read diverse sites and do not live in echo chambers. True, the blogosphere is creating communities of interest that overlap. True, many bloggers lack any kind of simple ideological orientation.
But some helpful empirical evidence has already been compiled, and even at this early stage, one thing seems true: Numerous people are listening only or mostly to people who already agree with them. If you're too optimistic, that's why.