Sure, the world's best idea, with no execution, is worthless. And sure lots of people have "great ideas" that are just pie in the sky fairytales that aren't implementable. But look at the flip-side... how many hackers are there in the world, who - in 2003/2004 or so - were talented enough to build, say, Facebook? I'd venture the number is on the order of "a metric shit-ton." Now how many had the idea of building something like Facebook? And then the flip-side to the flip-side is to ask "how many had the idea and the talent, but not the initiative?" And we could keep going back and forth... the point being, you need both the idea and the execution.
And executing a bad idea well isn't necessarily going to lead you to any success. What if I invented the world's most amazingly advanced and mondo righteous new model of buggy-whip, tomorrow? I could setup a factory, produce them, source the components at a great price, use six sigma and all sorts of fancy statistical process control and empower my employees and use the kaizen approach and churn out a constant supply of really high-quality (but inexpensive!), super advanced mondo righteous buggy-whips... which would all promptly collect dust until my factory went out of business, I went bankrupt, and I had to burn them as firewood.
Idea and execution are 2 parts of the same thing. Start with idea, then you execute, you get stuck after which you tweak the idea and then your back to execution.
You can also start with execution: do consulting for a bunch of companies in the same space. Figure out a common problem to all(ie idea) and then you execute a solution.
Idea Is Execution.
In this sense, I'd say an idea for a piece of software would be much like an idea for a novel. The idea is formed through the act of writing.