It goes the other way. Importance determines population. If the capital is not populous, that reflects the fact that the capital isn't especially important.
But assuming you're in a more or less steady state, that effect has already happened, so it's perfectly valid to just assume that importance and population are tightly correlated. The directionality doesn't really matter to the result.
Does it though? You still have provided no supporting arguments for that. People live in places for reasons beyond importance. In fact, "importance" really isn't a factor people consider when they decide to live somewhere at all really in my experience.
> and either San Francisco or Los Angeles (or both) also rank above Chicago
Wouldn't it be Sacramento in this case, as the capital but not most populous city? The most important city is a combination of multiple factors: size, economic or political power, and probably some more (like cultural).
There's no evidence for that anywhere, nor is this the author. This is a quote from a friend of someone in the elevator and there's no given criteria, so we can only speculate.
That said, I'm basically questioning an off-the-cuff quote made by a friend of someone who was in the elevator. The speaker should be forgiven for being imprecise…