I was speaking recently with Todd Dagres, the founder of VC firm Spark Capital, about the future of the video market and he had some interesting thoughts about how things will fragment. (He's an investor, along with Time Warner and Michael Eisner, in video service Veoh.)

One way he sees this happening is along the types of content.

First, he envisions companies trying to do exclusive deals for user-generated content. This won't just entail providing a business model for talented independents who attract big audiences regularly.

This means figuring out how to pinpoint the videos that are rising quickly to the top and showcase them early or pinpoint people who consistently have loyal audiences and help them break out.

Basically, he's talking about talent management, where you help people develop but on a much different scale, because the audiences can be smaller, but the costs are also smaller.

At the opposite extreme, he sees companies going after longer-form professional content. But not just the movies people are selling now. He sees people doing deals for old tv shows, pilots that never were shown, series that were shot, but weren't all shown. Or for foreign content that doesn't make it outside of their countries. So, people for instance, could pay $100 for a season pass for cricket in India.

And then, between these two extremes are content from semi-professionals, or folks who have had training or shot commercials.

Some of this bifurfaction is already happening. But there's one thing I keep having problems thinking about. As people license content, they seem to be doing it to anyone who wants it. So how does a site stand apart? With music, the answer seems to be you have to have a device, a la Apple. But maybe I am thinking too narrowly about this.

Before it's here, it's on the Bloomberg Terminal. LEARN MORE