Steve Gillmor thinks that linking has been gamed to death, making it worse than useless. I see how he might get tired of seeing the same link in multiple places, but he's reading the situation with just a bit too much cynicism.
The links themselves haven't been gamed at all. The gaming comes when the links get decoupled from their ends, and aggregated. Technorati, Digg, Google, and all the others divorce the link from its context, which is how the opening for gaming got created.
If Steve, Doc, or someone else makes a link to a site, it can't get gamed between the author and the person directly reading the page (or RSS thereof). It's only in the value subtractive process of Silage that we open the door for rot (aka Gaming the system).
The point of Gesturebank seems to be to preserve the context, and try to avoid undifferentiated bulk accumulation of data, which is defintely the way forward. I'm still waiting to be able to see the data I'm already giving to the cause there... (I don't get to see the anonymous pile of data yet, for some reason) because I think it'll be VERY interesting, and useful.
I'd like to know what Doc reads, and be able to make it as cheap as possible for him to point out something interesting. I imagine what we really need is the thing that Vanevar Bush originally described back in 1945... a system for leaving a trail of hyperlinks and commentary about a vast collection of documents.
The threads are the value, not just the weight of them, or their bulk counts. We need to preserve the value of the ends by building and using tools which keep his in mind.