While I can't say that I think your suggestion is a hard and fast rule (people vary too much for this to be the case), I think part of the reason I landed my current job is because I was practically drooling over it in the job interview. The company is large enough to handle setbacks, they have a decent-sized tech department that does test-driven development, pair programming and is almost pure Perl. What finally made me excited as heck about taking the job is when they started describing some of the technical hurdles they had to deal with. At one point, an interviewer was talking about how they deal with data entry people who type "Terminator" instead of "Terminator II", thus causing problems figuring out which movie the data is for. I immediately started asking questions about how they were solving the problem and started discussing possible alternatives. I'm sure that enthusiasm is part of why I am here.
Interestingly, now that I get to see more of what they do, I've turned to a different problem. They have a wide variety of problems where they need to infer Y from X, but despite having an obvious correlation, there is no clear algorithm to determine the answer. As a result, in my spare time, I'm working on a neural network in C (and discovering how bad my C is) to solve these problems. That's what makes enthusiasm pay off. The problems are so so much fun that I'm charging ahead on my own initiative on something that might benefit the company.
New address of my CGI Course.