This one was actually really useful, assuming the recruiters that told it to me were correct (100% of recruiters who gave me advice agreed on this, but it's a small n). I never knew how to answer the question "where do you want to be in five years." If I had a plan, it was "enjoy the fact that people like you are willing to pay me ridiculous amounts of money for something I find easy. Keep doing what you tell me until the level gets less ludicrous, at which point I'll find someone less well balanced with more money , and do what they tell me to." This seemed like a bad thing to say out loud. I assumed that this question was an oblique test of my ambition and assessment of my own capabilities.
Turns out, nope. They're asking about your plan to find out if "do good work for company X" is in it. I think the original advice I got took this too far: any manager who thinks "My only ambition is to do really well at this job you haven't even offered me." is a good answer is probably not going to support any goals you have or develop. But don't make the job on the table sound like a stepping stone. What worked really well for me was to say "I'm still gathering information on this. I know that what I want right now is to do a job remarkably like the one you're offering me, learn from it, and use that information to evaluate my next step. Possibilities include X and Y, but I want to stay flexible." Thus far I've avoided having X=starting my own company, because if I don't believe myself when I say it, they won't.
Advantages of this approach include: you're showing you really want this particular job, you're not going to run off three months from now, you're not wandering aimlessly hoping someone hires you, and you're adabtable.
The right answer probably varies on the type of position you're looking at. No small software company expects lifetime loyalty. If they limited themselves to people willing to work on one project (or related set of projects) for 20 years their pool would shrink to chlorine and dead mosquitoes. You can emphasize your desire to work with whatever their special technology is. Large companies (Microsoft, IBM, Google) have enough different kinds of work to fulfill some people's need for novelty, so play down any job hopping ambitions you might have (especially since one of those is tired of people using it as a resume booster to get a job at another). I wouldn't get overly specific what type of thing you want to work on, since you'll be expected to change every couple of years (graphics is probably broad enough that you can sell yourself as a guru. Security and perf are fine. Halo 4 Grenade Launcher renderer is not.). Emphasize growth within the organization.
For temp positions, it seems acceptable to me to admit using it as a stepping stone. How much loyalty do they want when their internal rules probably bar them from hiring you for over six months? Just make it clear that you won't leave before the six months if up. If you have a plausible reason why you prefer contracting to perm, you can slip it in here. "Full time position" is a bad answer unless it's temp-to-perm, because they'll wonder what's wrong with you that you can't find one.
I'm still at the stage where growth is the only acceptable plan: at the very least I can dream of being an SDET II. Eventually you hit a point where coasting is acceptable (not everyone gets to be an architect), but I'd be wary of admitting that. My best guess is to act excited about the inevitable treading you'll have to do as new technology arises ("I can't wait to work on the floobertygibbit"). If you're a crusty UNIX guru who hasn't done anything new in 10 years...I dunno, save money while that's still a viable option?
Anyone with better advice, feel free to chip in.
ETA: This advice is wrong. See here for an update.