What Does "Watch TV" Mean?
There are three different usage paradigms for television: active interaction, passive consumption, and background consumption.
Active Interaction
People have been working on cracking this nut for a long time with little consumer success. I was working for an interactive TV start up in 1990; we did classified ads, sports scores, photo albums, and weather. Recently TV widgets have become a hot new meme -- they're just another take on the same interactive TV ideas that consumers have already rejected. Television watching is in its essence a passive medium. All the wishing in the world is not going to change that. The only real success in active interaction with the TV has been the console-based video game business. From a real world perspective, this consumption paradigm is de minimis and I expect it to remain so.
Passive Consumption
Passive consumption has some brief interaction at the beginning of the session to start or select a video, followed by a long period of passivity. Over the years, this interaction has changed from choosing a channel, to putting a tape in the VCR, to starting a DVD, now it's becoming manipulating the VOD interface (either a local DVR or a streaming service). This brief interaction is followed by a long period of passive, albeit engaged, consumption. If the capability exists in the delivery channel, users often skip commercials. This is what is generally considered "watching TV," but it does not constitute the majority of the TV consumption in the US. This is the consumption paradigm that most "TV start-ups" are focused on. This business has been disrupted many times since 1975, when Sony introduced the Betamax. Each time, the programming selection method changed, and each of these changes was a disruption that was predicted to bring down TV. None of them did. TV consumption continues to grow. It will be disrupted further but this disruption, like the previous ones, will not change the majority of television consumption.
Background Consumption
The vast majority of television consumption in the US (better than 75%) is background consumption. Background consumption is what drove the average daily household television viewership in 2009 to 8 hours 21 minutes. Since 1997, despite a fracturing media landscape, despite dozens of new consumption channels, the average American household consumes ~20% more TV now then it did before the internet revolution. Americans are NOT watching less TV because of the internet and online video: they are watching more. Many Americans leave the TV on in the background at home during all their waking hours. They briefly engage with it and go back to other tasks. It's companionship. It's serialized, interrupt-driven content snacking. They leave the TV on while the cook, or eat, or clean the house, or surf the internet, or tweet and play on Facebook, or play World of Warcraft, or have sex. Commercials are not skipped in this use case. Attention is shifted away, attention is shifted back. The influence is subliminal but very real.
Finally, the expected default behavior of television is not to wait for user input. You turn it on. It makes noise and shows you pictures. It does not start an interrogative. If you like what it's showing you, you leave it alone. If you don't, you change the channel.