Reading, writing, and talking all implement the same interface: they act as a valve through which information can enter and exit your brain. But they're not the same. They're great at very different things.
Writing is great because you get time and space to think carefully about how to present an idea in into its best clothes. Also, once you pay this cost, lots of people can get value from it.
Reading is great because you get the product of thousands of hours of research and thought distilled down to 6 or 8 hours. You don't spend time meandering down dead ends.
Talking is great for getting to know people, and letting other people get to know you.  It took me a long time to even realize that this was a thing you could explicitly do in conversation. Before, I thought you just talked about ideas and heard the other person's ideas, and after several years you slowly got to know each other naturally.
This helped me understand why I dislike abstract philosophical conversations: they feel better suited to reading or writing. It also explains why I love talking with people about their past: it's a route to understanding someone's present inner world. 
I'm not saying that you should never talk about ideas. If you're a domain expert looking for advice on a research project, for example, then you should absolutely call over your friend to ask her what she thinks of something. But in many cases where we casually talk about ideas, reading or writing would be better suited. 
If you find yourself in a boring conversation, ask yourself whether you'd be better off reading or writing about the subject. If so, you might enjoy steering towards something that conversation is uniquely good at: getting to know your conversation partner's inner world.
 Should you offload "letting other people get to know you" into writing, and use talking only to ask about other people? I don't think so. I tried it freshman year and experienced it as lonely. Telling other people about yourself seems like an essential part of making friends and feeling known. Many people seem to try this after reading Carnegie, and I'm curious whether anyone's gotten it to work sustainably.
 Even after I realized that I liked these conversations, it took me a while longer to consistently produce them. This topic deserves its own post.
 There are more exceptions. For example, asking questions (talking) after reading a textbook chapter is usually a good use of time.