People have been predicting the death of television for 20 years now, and so far it's been entirely wrong. But it does seem viewership habits are starting to change.

The content defines the platform, so whereas when I was working at ABC from '66 to '76, people said it was the "great wasteland." It was the least-objectionable program that succeeded. It was, if you could get behind "All in the Family," you were successful.

It's very hard to be good. It is self-destructive thinking to think that there is too much good.

Wrapping rubber bands around a watermelon is not journalism. It is entertainment. But the key to success in media has always been a broad mix of serious reporting and entertainment. The New York Times does not make its money on reports about Iraq and Syria. It makes money on its gardening section, food and, yes, stories about cats. "The Today Show" is a very successful program because it is a mix of the celebrity chef and the crazy pet who does the rolls and serious news and interviews.

Successful publication is all about the mix. What Buzzfeed discovered was that people like cat pictures. We can pretend to be embarrassed about that as a species, but it is actually a truth. So Buzzfeed publishes a lot of cat pictures. But they use the money from cat pictures to build an exceptional newsroom that publishes stories that far fewer people want to read, but it is very important journalism.

Some people have made a fortune by being employed. Jerry Bruckheimer does not own his content. Warner Bros. owns his shows. They are on CBS, and he makes a fortune.

When CNN launched in the early 1980s, everybody said: A 24-hour news network won't work. They launched, they did ok, CNN went almost bankrupt because of the risks they had taken, they got bailed out, and 25 years later CNN is a huge global brand. I think the same is going to happen in digital. If you look at the younger generation, there is a huge consumption of digital media and almost no consumption of print or traditional television. Eventually money will follow that. It is just a question of which companies win, how long it takes to get there and what kind of model you need to apply.