Philip Guo (Phil Guo, Philip J. Guo, Philip Jia Guo, pgbovine)

Writing for Work

Writing for work is always hard. The most sustainable way for me to do it is to write for 2 hours every day, and to maintain that level of dedication for a 3-week burst before a deadline.

As an academic researcher, writing for work is a big part of my job. I've gotten better at it over the years, but it never feels any easier. Technical writing is still the most arduous, most mentally taxing, most emotionally draining kind of work that I do. And I like writing, too!

I also write a lot for fun, but that's much easier since I do it only when inspired. In those rare sublime moments, words just flow out nearly as fast as I can type. But writing for work is an integral part of my job, so I can't just wait for inspiration to strike. If I don't write enough, then I get fired. Unfortunately, the process of explaining dense technical arguments in a concise and convincing way is far harder than, say, writing this article. It's always a slog.

Real writing

I can do at most two hours of real writing per day. That's shockingly little. If you have less writing experience than I do, then you probably can't do more than two hours per day either.

So what counts as real writing? Only one thing: producing new text that is nearly finalized. Real writing means writing polished sentences that need to only be lightly edited for style, consistency, and accuracy. It's like writing production-quality code that only needs some tests and patches, not hacking up throwaway prototypes in a caffeine-fueled frenzy.

Outlining isn't real writing. Brainstorming on the whiteboard isn't real writing. Organizing notes isn't real writing. Scribbling sentence fragments and brain dumps into your document isn't real writing. Talking about writing isn't real writing.

These activities are all important precursors to real writing, though. And they are far easier to do because you can bullshit yourself into believing that you've done a good job even when your thoughts are sloppy. But when you write for real and then show it to others, there's nowhere to hide.

Two hours, three weeks

Okay, given that I can write for only two hours a day, how do I ever make any progress? I structure my writing around deadlines, either real or artificial. For my job, the two main kinds of deadlines are conference paper submissions and, more recently, grant proposal submissions. I will usually start writing three weeks before a deadline. I can't sprint for any longer, and anything less than three weeks isn't enough time.

For those three weeks, I make sure to always get in my two hours of real writing per day – no matter what. My preferred time to write is first thing in the morning, when my mind is freshest. On some days, it's excruciatingly painful to squeeze in those two hours, especially when I'm traveling, hosting out-of-town visitors, or feeling sick. But as Jerry Seinfeld says, “Don't break the chain.” If I skip a day, then that means I'll have to do four hours of writing the next day (or three hours over the next two days), which is nearly impossible. So don't break the chain.

Also, listen to Taylor Swift. You can't demand the perfect conditions for writing. You need to be able to write anywhere, anytime:

One positive consequence of being able to write for only two hours a day is that even if I have a deadline, it's not like I'm writing 24/7. I still have a ton of time to do other stuff at work. However, the main caveat is that my other work can't be as strenuous, since my brain is already fried from two hours of writing each morning. Another cool consequence is that once I get my two hours of writing done, I can even slack off for the rest of the day without feeling guilty.

Prep work

Prep work ensures that I can actually get down to real writing when I sit my butt in the chair each morning. Examples include:

  • outlining
  • brainstorming on the whiteboard
  • getting feedback from others on high-level pitches
  • starting the LaTeX file with the proper paper template
  • creating a title, subtitle, and section headings to structure the document
  • finding and importing bibliographic citations into the document
  • pulling in snippets of notes and external sources into the document so that they are ready to integrate into the text
  • making graphs, figures, and tables, and then adding captions

Read Intense Single-Tasking for a recent personal example.

Prep work reduces the transaction costs of writing, which makes it easier for me to kick-start my two morning hours. If I sit down in front of a well-formatted document with all the figures, references, and notes already in place, then I can immediately start cranking out new text. But if I have to face a blank screen in the morning, then that's not very motivating. So it's very important to prime the pump.

Fortunately, prep work is far easier – and often more fun – than real writing. It doesn't feel mentally taxing, so I can do it anytime, even when tired.


After exhausting my two-hour daily writing quota, I usually take a break and then go back later in the day to do some editing. Editing old text is much easier than writing new text, so this isn't much of a mental burden either.

However, it's important to resist the urge to start editing when you're supposed to be writing new text. Since editing feels productive, it's very tempting to do it in lieu of writing. But that doesn't move you forward toward your end goal; it just polishes what you've already written. Editing doesn't count toward my two hours.

Rough drafts

I don't believe in rough drafts. Whoa, blasphemy! I see two major problems with rough drafts:

  1. People can't give you meaningful feedback because the writing is so rough and hard to understand. If I'm forced to critique a draft that supremely sucks, then I'll just offer some half-ass comments to be polite and not try super hard. Not good for either party.

  2. Even if you do get good feedback, you're tempted to just make minor incremental changes to your draft rather than rewriting large chunks of it. Thus, spending time on a rough draft not only takes time away from real writing but also gets you stuck on a not-so-great local maxima.

Instead, I use outlines, whiteboard sketches, paper notes, scribbled sentence fragments, and other pieces of prep work as my rough drafts. These artifacts are low-fidelity prototypes that I can revise, redo, and throw away without any pain. I get lots of early-stage feedback on these prototypes, but when I start writing, I write for real. The mid- to late-stage feedback I get is all on the final draft.

Grinding versus writing

This final section is most relevant for people such as grad students who need to both do a lot of grinding and most of the writing on their research papers.

Research is never done; there's always more work that you can do before starting to write a paper. Thus, you need to stop grinding three weeks before your chosen deadline and switch into writing mode. (This estimate is for 10–15 page technical papers in my field. A longer or denser paper needs more lead time.)

As you're writing, you can figure out how to refine your experiments, implement new tweaks on your system, create new graphs, and add technical details on-demand for your specific paper. Remember, you're writing for only two hours per day, so you still have the rest of the day to grind, albeit with less energy. Focusing your efforts on the upcoming paper will help you direct your grinding efforts. Of all the possible paths you can take with your research, what should you implement, test, or measure right now? Whatever is necessary to make your paper stronger.

Seriously, trust me on this one. Don't wait until one or two weeks before a paper deadline to start writing. A heroic sprint will only lead to burnout and a badly-written paper that will likely be rejected. (Some experienced researchers can pull off uber-writing-sprints, but they don't need my advice!)

Created: 2014-01-29
Last modified: 2014-01-29
Related pages tagged as assistant professor life: