Execution is important.
But I think many people have the skills to create what ends up being successful.
Take Groupon for example. It made something a lot of people wanted. Could most people on Hacker News have made it? Sure.
But Groupon created it before anyone else (I might be wrong about this?).
Another example is HN. Pg has said that what he thinks users want most are quality articles and discussion.
While it's "just a forum", I think Pg is right for not adding a lot of features, or spending the limited time he has working on the UI, and instead focusing on how people can discover quality articles and make/read quality comments.
"Make something people want." It's simple, but it sure isn't easy.