Here is the Overriding Principle: what major research universities care about is research. That’s all. Nothing else.
Don’t be too well known outside the field. ... I hate to say this, but the evidence is there: if you have too high of a public profile, people look at you suspiciously. Actual quote: “I’m glad we didn’t hire Dr. X; he spends too much time in the New York Times and not enough time in the lab.” And that’s the point — it’s not that people are jealous that you are popular, it’s that they are suspicious you care about publicity more than you do about research.
Don’t write a book. ... Why? Because while you were writing that book, you weren’t doing research.
Don’t worry about teaching, leadership, organizing, etc. ... Why? Because you’re spending time on something that isn’t research. But generally it won’t hurt, it just won’t help. ... Everyone agreed my case was very close, and my teaching was among the best in the department; it didn’t help. The point is simple: this stuff is not research.
Choose your hobbies wisely. ... But here is the paradox: you are better off if your hobbies are nothing like your work. Permissible hobbies include skydiving, playing guitar, or cooking. Suspicious hobbies include writing of any sort (novels, magazine articles, blogs), programming or web stuff, starting a business, etc. Why? Because there’s a feeling that this kind of activity represents time that could be spent on research.
Don’t dabble. ... it will be taken as evidence that your interests may wander over time — so that, whereas you were hired to be an expert in area A, maybe in a few years you won’t be doing that at all. Kiss of death.
I’m generally in favor of the tenure system; like democracy, it’s the worst system out there, except for all the other ones that have ever been invented.