When you try to do something ten per cent better, you tend to work from where you are: if I ask you to make a car that goes 50 miles a gallon, you can just retool the engine you already have.
Most of us have to spend a lot of energy to learn how to drive a car. Then we have to spend the rest of our lives over-concentrating as we drive and text and eat a burrito and put on makeup. As a result, 30,000 people die every year in a car accident in the U.S.
We don't take on Google Glass or the self-driving car project or Project Loon unless we think that on a risk-adjusted basis, it's worth Google's money to do it.
When we try to make a car that drives itself, we believe - whether we're right or not - we believe that there would be strong net positive benefit to the world if cars could drive themselves safer than people could.
People text when they're meant to actually be driving. So imagine what they do when they think the car's got it under control.
It's crazy that you have to tell your phone or your computer or your house or your car 'It's me!' hundreds of times a day. Wearables will solve that problem.