This is a fantastic post by Scott. I think the great idea fallacy continues because it appeals to people's fantasy of becoming successful without much work. Anyone can imagine themselves having a great idea and then being successful. It's a lot harder (and less pleasant) to imagine taking an unremarkable idea and spending years of hard work to make it succeed.
I think part of it is due to academia. In computer science classes a lot of questions have a "best" answer. In the real world customers don't buy software because the architecture started with the "best" ideas. They buy software because it fits their need. The best idea could contribute a superior customer experience but it is peripheral to what it takes to be successful.
Only if your conception of "academia" doesn't extend beyond first-year undergraduate coursework. Real academia is entirely about tackling questions for which answers are unavailable or incomplete, and finding better answers. Intellectual astronauts don't make it out of grad school, because they don't publish papers (for that matter, they tend to barely make it out of undergrad, because they're the guys most likely to be complaining about "jumping through pointless hoops").
Academics also tend to love to talk about their ideas with other people, and aren't very secretive (with the exception of their closest competitors, perhaps). My first real intro to the attitude that ideas are inherently valuable was when I moved into the tech world. If I had a nickel for every goofy entrepreneur I've met who is in "stealth mode"...well, I'd make more money than I did in academia.