You should write. You (probably) already know that[1].
I started internalizing it more when I came to a simple (high-school math related) realization: an unlikely combination of rare skills is unintuitively unlikely.
The probability that two independent events, P(A) and P(B), occur is P(A) times P(B). If the probability of events A and B are 10% and 10%, the probability that both occur is 1%. Easy-peasy. But of course, in reality, finding completely unrelated events is rare, but let's put that aside for a moment.
However, we can analogize events with skills and probabilities with percentiles. That is, the percentile of two independent skills P(A) and P(B) is roughly P(A) times P(B). So, finding a person in the top 1% implies finding someone within the top 10% of two independent skills[2].
Given that writing is a communication and thinking tool, it's safe to assume that if you do not write well, you (probably) cannot be a great hacker. Smarter hackers than I have already said this many times.
To become a top one percent hacker, you can either be a top one percent hacker or a top one percent that happens to be a hacker. The former is about becoming insanely great[3] at programming. The latter is about becoming great at the hacking craft while also being great at something else.
If how well you communicate with computers is independent of how you do so with people: we have a jackpot.
Learn to talk to/write to/read computers and people alike.
After all, the probability that you can do both may be rare squared.
Footnotes
[1]: This post is, in small part, a response to a post I saw on HackerNews titled “Engineers should focus on writing”.
[2]: Or any other solution to the equation, I know, I know...
[3]: I still wonder about the relationship between The Macintosh and Insanely Great.