I hear people say this often and never quite understood it. The idea is very important. Having a bad idea vs. a good idea is like the difference between trying to sell snow to an Eskimo and trying to sell a glass of water to a man dying of thirst. Given the right salesperson, both can be theoretically done and both can be screwed up, but that doesn't mean that execution is everything.
Absolutely, you need both. I'm not sure how to quantify which one is more powerful or important, especially since "idea" can mean many things, from a very vague concept to a fully fleshed out and vetted business plan. I was just taking issue with the "ideas are worth nothing, execution is everything" mantra.
That should go without saying though. I think a better way of asking the question is "what's better: a brilliant idea executed so-so or a so-so idea executed brilliantly?"
That to me is a more interesting and logical question.
While there are clearly bad ideas, Facebook for robots is probably one, there are lots and lots of great ideas. I always tell people my 'real' startup idea is time travel, but I'm still working on the implementation.
Often these great ideas are hard. PG has an essay on this (http://paulgraham.com/ambitious.html). Better search is a clear example. Google is amazing, but it isn't perfect. Yet being better than google is really, really hard.
I think one of the tricks of being a founder is to pick the greatest idea that you are capable of implementing, with the caveat being that picking something that you are capable of implementing right now is probably too conservative.
