This student makes some excellent points, and it's a bit scary how some things reinforce each other. Some journals, Nature being the most notable, chase impact factor above all else. The only thing they truly care about is that the papers they publish get cited. This means highly original work, which frequently doesn't get cited much for a long time, is less desirable than a turn-the-crank paper from a respected name…
That's interesting, because I've also seen the opposite problem in high impact journals, as in the Arsenic-based-life paper published in Science last year. The journal was keen to get a big dramatic story, and must have let it in with fairly minor peer review. Then it got roundly criticised after publication. I hate the way we have to 'sell' research to compete for funding and prestigious publications. But my impress…
In many cases, the paper is not the best method of conveying information, but it works well enough for most cases that no one has managed to usurp it. Editor reviewed journals might have some flaws in their process, but it is effective enough at picking out useful papers that scientists clearly chose to purchase them and use them as a source. So I wouldn't hate the publishing cycle per say, just the fact that in those cases where it is not exactly optimal, the tyranny of the majority is ruining the day of a small set of research types.