As a few people sent in, famed author Seth Godin is doing an interesting experiment with Kickstarter, where he has teamed up with a publisher who essentially wanted to use the platform to prove there's significant demand for Godin's next book. Basically, if he could effectively sell pre-orders for the project to raise $40,000, then the publisher would invest in the project as well and support getting it into bookstores and putting a promotional campaign behind it. It took Godin less than three hours to surpass that goal (and then go way, way beyond it as well).
Godin makes a strong point about how the traditional process, of investing a ton of money upfront, without knowing if there's really demand, is inherently risky for traditional publishers (and studios and labels). This is one area where a platform like Kickstarter is quite interesting beyond just the "fundraising" side of things. It can also be a tool for gauging demand for a project.

In fact, some others have been recognizing exactly that. Andy Baio recently wrote a column at Wired, in which he talks about using Kickstarter as a way to judge demand for something without having to put forth that initial capital expenditure.

Of course, once you realize that it can be a demand platform, rather than purely a funding platform, interesting possibilities open up:

As far as I can tell, nobody’s flipped it around and tried to commission a musician to play for fans. Most bands already play corporate events and private parties. If fans collectively raise the same amount of money, why not play a house show for them instead? For fans, it’d be a once-in-a-lifetime experience to see an artist they love in an intimate setting. For musicians, it’d pay well without the malaise that comes from playing the Intel holiday party.

In other words, as a demand platform, Kickstarter (or others) can be used to demonstrate demand (and actual money) for something that people want to come into existence, and then people can figure out how to make it happen. That's pretty powerful just for being different than how things have been done before.

That's not to say that this makes sense for everything, or that there aren't risks associated with it. Execution matters, and paying up at the demand stage can lead to disappointment if the eventual product doesn't live up to expectations (or, worse, never actually gets made). So there's a different kind of risk there, though one that is likely to be more distributed. There is also the risk of "failure." A good idea that may not be explained well at this stage may not come to fruition. But, in the end, these are really just flipsides to the traditional risk taken by gatekeepers. It's just that it's getting moved around, and done in a way that actually decreases the overall burden of the risk, which makes it possible to create more with less overhead (a good thing for everyone!)

And, remember, we're really only in the first few years of these types of efforts. Kickstarter, which gets most of the attention, is just three years old. Imagine where things will be 10 years from now.