This is awesome.

Here's the biggest thing that seems off to me: these indexes need to be fully populated by the programmer, right? I saw a bunch of "You could search Twitter, think about the Sunlight foundation, here's some firehoses" but then this (and the rest of) the tutorials seem to be all about me feeding it data directly.

If I wanted to search the Twitter firehose, say, I'd have to do all of the scraping and insertion myself, right? I can't just hook it up and head off? That's what makes sense, but some of the marketing seemed to imply otherwise.