The usual geek mode is ListenWithYourAnswerRunning. This is probably the biggest social problem for geeks. We're just dying to share our understanding; we really want to help; so in any conversation we want to tell everyone the most important idea we have.
We're bright enough to know there is a time and a place where our ideas can be received, and just after "hello, how are you?" is not that place. But ListenWithYourAnswerRunning works so well when talking with our fellow geeks, we rationalize that the reception problem is with the non-geeks. Call 'em suits if you like, or something else derogatory. 7 will get you 2 you're married to one of them ... and treating em like furniture is costing you ...
There's also a scaling limit on ListenWithYourAnswerRunning. Building your answer while simultaneously tracking the context of the ongoing conversation causes you to lose track. And then you are a dolt or a boor - sitting dumbly while everyone else follows the focus of shared thoughts, or blurting out something that's irrelevant, stale, or tangential.
Therefore,
ListenWithQuestionRunning. Instead of spending your mental cycles on trying to think of the most important thing to say next, spend your mental cycles trying to find the most important question to ask next. Not just the most important question based on the conversation stated so far, but the most important question based on your understanding. When you have an opening - and only then - ask just that question. And nothing more.
This has various semi-miraculous effects. First, even if you have rudely interrupted someone else, they will hear that you're interested in creating a space for them, not for yourself. They're pleased that you're actively assisting them to explain themselves. Or someone else in the conversation may be pleased you've given them a chance to explain themselves. This aggregation of pleasure in the conversation improves the quality of everyones' contributions to it.
Moreover, if you are truly concerned with asking the most important question you'll find yourself doing ActiveListening without even trying. What you're trying to do in conversation is hone shared understanding. With geeks you can do this by popping out ideas and watching your brethren/sistren scamper around trying to catch 'em, process 'em, and bat 'em back to you. But expecting folks to do this, even when they are geeks and even when you have a generous intent, invariably comes across as selfish/childish. Not cool.
ListenWithQuestionRunning comes across as wise. It fits the rhythm of non-geek conversations without disabling your natural mental overdrive. It gives you a way to get boring/stupid ideas out of the way without running 'em down like dogs on the freeway. And it's contagious - when you help make the conversation an open, hospitable place, folk feel obligated to be open and hospitable right back.
Used consistently, ListenWithQuestionRunning enables a geek to run circles around the non-geeks. You keep doing it until no one else can answer the most important question. Then and only then you pop out an answer. Short and sweet because the conversation has prepared everyone for it. And because your backburners have had at least 10 times their usual time allotment to refine it.
Now don't flip over into ListenWithYourAnswerRunning again. You know the most important question usually changes as the conversation develops. So you go right back into ListenWithQuestionRunning after contributing something. Don't defend your answer. Let the others come to terms with it in their own sweet time. If they have difficulty, ListenWithQuestionRunning! That'll tell you the most important question that will assist them.
Part of the ConversationPatternLanguage