I don't know, I disagree that they're not solving really interesting problems. My first jobs out of college were at hedge funds, pretty much by accident, and I got the chance to work with a ton of really smart people -- many of whom were math or engineering PhDs -- in a very challenging environment.

One two-person team I got to work with was doing something resembling stat-arb, and was pushing about $250 million of trades in the last 10 minutes of the market. Beyond getting to be hands on with some crazy math, building a system to make that happen isn't exactly trivial.