I reluctantly signed up for a journalism major, thinking I needed a fall-back way to make money should my career as a novelist fail to take off. As I started to try on journalism, including doing internships and working at the campus paper, I found I actually liked it. So I started to want to be a journalist.

From about ninth grade on, I knew I was a writer at heart. I had fantasies of being a great novelist, but I thought that seemed like an iffy way to try to make a living. So I tried journalism while in college, and really liked it. But even in journalism, I've always pursued ways to be somewhat literary, whether writing a column or writing books.

Fidelity is the total quality of an experience, including a sense of exclusiveness and aura. Convenience is simply how easy something is to get, which often means a low price and ubiquitousness. A super-fidelity product or service would lose its luster and quality if it's pushed too hard toward convenience. A super-convenient product or service would start to get expensive and exclusive if it moved toward higher fidelity, which would naturally undermine its convenience.

Artificial intelligence uses a complex set of rules - algorithms - to get to a conclusion. A computer has to calculate its way through all those rules, and that takes a lot of processing. So AI works best when a small computer is using it on a small problem - your car's anti-lock brakes are based on AI. Or you need to use a giant computer on a big problem - like IBM using a room-size machine to compete against humans on Jeopardy in 2011.

If you spot a market where the only choices are at either one end or the other - high fidelity or high convenience - there's probably a big opportunity at the other end. That was the opening for Federal Express, for instance. When it started, there was only one mail service in America - the US Postal Service, which was high convenience. Fred Smith created a high-fidelity mail service.

Eventually, we need to have computers that work differently from the way they do today and have for the past 60-plus years. We're capturing and generating increasingly massive amounts of data, but we can't make computers that keep up with it. One of the most promising solutions is to make computers that work more the way brains work.

Who "wins" and who "loses" depends on when you measure it.