However, I was thinking about the TiVo Super Bowl report, and I had a brainstorm. Why not use the data collected from TiVo to provide an interesting service back to its subscribers?
Here's the antecedents to the idea. Back in the 1990's, people that read USENET had an idea to setup group filtering where the collective votes of the readers of a newsgroup could be used to highlight good posts and hide ones that were flames or spam. I'd also seen some of the post-debate coverage where they did studies with an audience that was able to continually say if a candidate was favorable or unfavorable during their responses.
Software running at the TiVo HQ could use the data collected from people watching a program to determine what are the highlights and what parts are people skipping over. I'd expect for most programs, the highlights would be the actual program, while the skipped items would be the ads. Then, when watching a show, your device could pull down that aggregated data and use it to provide chapter points, letting you skip over boring sections and get right to the good stuff. While useful for regular shows, I think this could be really interesting for things like awards shows and sporting events. If you were willing to delay watching something for a couple of days, they could probably get enough data to build a pretty good navigation map of the program, letting you easily skip to the next play or the next presenter.
Of course, I can see why TiVo wouldn't want to implement this, as it could really tweak the content providers. A few years ago, Sonic Blue was sued over their commercial skip feature that used blackouts to determine when a program had gone to commercial. TiVo still hides their 30-second skip feature behind a back door code. The studios are pushing for copyright law changes that would make TiVo and related services outside the realm of fair use. Still, it points at some ways that understanding the collective behavior of a population can make our own interactions smarter.