On the one hand, you're arguing "the system will pick up the slack." Okay, that's easy for you to say - you designed the system. It will undoubtedly give you what you need, or you'll tweak it so that it does. It's not so easy for the rest of us to hear - after all, most of us are here because the "system" of Reddit failed us. Your vision of the way this works is "if something is popular enough, you'll see it because your friends like it." The problem is that the way you envision the system, if my friends consistently like things that I don't, they're going to cease to be my friends. Your example with thenewgreen is perfect - you know him best, but you don't follow him because he shares too much. So if he finds something cool that you might actually like, you'll be the last to know. That's unacceptable. You're basically saying "Hubski tracks influence as far as who follows who, but eventually the whole thing crumbles because once someone starts indulging interests that don't appeal to you, your best bet is to dump him entirely, good with bad."

Your "baggage" complaint is a red herring, related to your unwillingness to address "tags" the same way every other institution on the planet addresses them. If there were consensus on tags, it wouldn't be that big a deal - your problem right now is that "tags" are fundamentally "editorialization" with a mix of categorization thrown in for sport. it's the #media/#foxnews/#murdoched dilemma - these things should, systematically speaking, be one thing, not three.

Finally, you're extending dislikes in the same direction as likes. Most people "dislike" far fewer things than they "like" - while most people like ice cream, some don't like rocky road, some don't like raspberry swirl. Likewise, not a lot of people are going to block #music but some of them might very well block #dubstep.

If it really worries you, limit us to ten "blocks." You're worried about problems down the road, and I think those are things worth worrying about... but your users are asking for things now. It's scalable now. See how it works. If it gets unwieldy, deal with it when you start to see the first inklings of problems, not when they're theoretically over the hump. Again, I think that a proper approach to the "tags" issue clarifies a lot of this.