We start to try to live in tomorrow and the future, and start to think about what we build today as a stepping stone to graduate users.

The thinking needs to shift from being less about the actual things, to being about the individual user.

You sort of have to separate what are questions you can ask that are going to help make you smarter about your thesis versus... trying to get somebody to validate it for you.

You have to take those concepts and prove them, a little bit like you do a PhD thesis.

No one is going to tell you what to build. If they did, then they should do it, not you.

You have to have the definition of what you're trying to do across the system.

It turned out that building mobile software was a lot more like building hardware... where you had 1 shot and you had to get it right, right out of the gate.