The problem with showing your entire idea spreadsheet to the mob isn't that someone is going to steal one of the ideas. The problem is the signal-to-noise of the responses.
When you sit with one person, in person, and go through the list, you get to focus on one idea at a time, you both get to control the agenda, you can both react directly to each other's social signals, and you've got a rapid-fire exchange of sentences. If the exercise proves to be a waste of time for either of you, you can cut it short.
None of this applies when you post a list of N things on the net and ask for "feedback". An arbitrary number of people will respond to an arbitrary number of the things, in arbitrary order, with arbitrary levels of engagement. Most of your visitors will, at best, be drive-bys: They might type out their thoughts on one of the points, but they're unlikely to stick around for ten posts' worth of back-and-forth, unless they're already your friends. (And, if they're your friends, why not run the ideas by them in person, over drinks?) And there's a good chance that the ensuing thread will be dominated by critiques of your spelling, cheap jokes about the funniest idea on your list ("Blog posts with a 140-character limit? Don't you know how to type? lol."), or an exhaustive discussion of one of your points -- whichever one happens to be the best at generating free-form conversation, which is a metric completely unrelated to how good an idea it is.
And, of course, the responses will probably be largely context-free. You won't necessarily know if that insightful-sounding response is from a potential customer, a VC, an employee at a big company, an entrepreneur, a con artist, or the Vice President. As the original article takes pains to point out, knowing who is critiquing your idea is really helpful.
Free-format brainstorming meetings do not scale well at all, even in person. On the web they're even worse.