Comment by hindsightbias

The population of innovation is a sliver of subset of the engineering/CS/AI community. UI-Urbana (Andreesen), Stanford (Page/Brin), Harvard (Zuck), etc, etc. That gets you down to a couple of 1000 people or so. The size of an American High School.

You don't have to see the future, you just have to know where those working on it are and meet enough of them. They're in places like a Supercompute booth or MIT poster session trying to tell you. They may not have The Big Idea yet, but talent tends to stand out.

Andreesen didn't randomly pick an IBM group. He picked one doing Unix distributed system management and virtualization 25 years ago. I didn't have foresight to make the same choice, it just seemed obviously interesting and there were only three major players in that space.