The most straightforward path would be if we could bring the cost of solar electric and wind down by another factor of say, three, and then have some miraculous storage solution, so that not only over the 24-hour day but over long periods of time where the wind doesn't blow, you have reliable energy. That's a path. But energy storage is hard. That's not a guaranteed path.

I'll get to see many disease eradications. and we're seeing a lot of progress.

I really do think cancer will largely be a solved problem. I think most of the infectious diseases like malaria - our foundation is very involved - once we're finishing polio eradication, then starting up this malaria eradication, and getting that done as fast as we can.

It's pretty amazing to go from a world where computers were unheard of and very complex to where they're a tool of everyday life.

We're still missing about a dozen vaccines that will make a huge difference. For adults, we've got HIV and TB are still huge; for kids malaria is still killing a half million kids a year out of that 6 million. We probably need some vaccines, but we need a little more data to make sure we're getting the vaccines that will save the most lives.

Any version of Windows is going to have lots of great new things that people use and things that are tough.

The trick generally is to break programs into pieces and have those pieces be individually testable and so then when you move on to the other pieces you treat it as a black box knowing that it either works or doesn't work.