Start-ups should be based on radical ideas. There should be a high failure rate for start-ups, because if there isn't their ideas aren't bold enough.

An awful lot of successful technology companies ended up being in a slightly different market than they started out in. Microsoft started with programming tools, but came out with an operating system. Oracle started doing contracts for the CIA. AOL started out as an online video gaming network.

I would say the consumer Internet companies - in a lot of ways, if you go inside the consumer Internet companies and you see how they run, it's how all their businesses are going to run.

One of the advantages of moving quickly is if you do something wrong you can change it.

Whatever you're selling, storage or networking or security, you're going head to head with the incumbent players.

I always had the old-school model that I'm going to work for as long as I'm relevant and focus on for-profit activities and someday when I retire I'm going to learn about philanthropy.

Around '93, '94, the conventional wisdom about the Internet was that it was a toy for academics and researchers. So it was very, very underestimated for about two years.