In high school, I was an insatiable consumer of science fiction. In hindsight, it is unbelievable how much time I spent reading instead of doing the other things usually considered to be typical for teenagers (going out, doing homework, etc.). I was the classic science fiction nerd. My live revolved around the words of Asimov, Weir, Clarke, and probably every other science fiction author you could name.
It was not until relatively late in my academic career when I realized my thirst for the written word had given me a superpower. All of those years slurping up science fiction had accidentally given me an unusual ability (in this day and age). An ability that, once upon a time, had been commonplace.
Fortunately, this ability can be trained quite quickly. What had taken me years to learn by accident can be developed in just a few weeks (or days, if you're committed).
This blog post will serve as documentation of:
In the interest of avoiding clickbait, I'll just come out and say it plainly: this "superpower" is the ability to read full books from start to finish.
Sounds boring, right?
I call it a superpower because it is becoming rarer by the day. Fewer professionals read anything longer than a short blog post or article today than at any point in the last century. I won't get into why, since there is already an abundance of documentation explaining why people have been reading less. But in order for something to be a superpower, it needs to be more than just rare. Why should reading count as one?
Reading qualifies because it allows one to do something that is otherwise extremely hard: understand a topic or idea at a level approaching an expert.
When I was in school, I remember sitting in lectures surrounded by students who hated our courses' textbooks. My colleagues called the infernal blocks of paper "boring", "repetitive", and "heavy". Indeed, they were, but they also contained most of the information the professor would lecture about, and presented it in a format I could consume in a fraction of the time.
I realized that I could save myself hours of tedious homework and lecture time by simply reading ahead in a course's textbook.
Now that I've spent a decent chunk of my life removed from my studies, I've come to realize that this same philosophy can be applied professionally in our work as engineers. Hundreds of hours of intellectual floundering can be saved with just a little bit of reading.
To illustrate why the kind of deep understanding that can be obtained through literature is relevant, I'll bring an example from my own recent work.
For context, I work on Harper, a grammar checker for those who care for their privacy. We've recently been working to expand our offering to compete with Grammarly Premium. This is a complex, open-ended problem with no "right" answer. I can imagine two approaches to it, one good and another bad.
I was pretty well versed in the kinds of offerings that Grammarly provides, and it was pretty obvious how many of them work. Grammarly takes some LLM provider with an API (like OpenAI or Anthropic), provides a bunch of style guides as context and wraps it all in a pretty UI and sells that at a premium. We could do that same thing, right?
In the bad case, I could just copy what Grammarly is already doing and call it day. Good idea, right?
I didn't think so, so I sat down and read through some of the prime-time literature coming out of labs written by experts in natural language and the arts. They pointed me in a completely different direction.
By reading deeply, I was able to come to the realization that many of the features that Grammarly hides behind their paywall are relatively simple to implement, if you know what you're doing. Not only that, but combined with the existing infrastructure Harper already owns, we could do it cheaply without disrespecting our customer's privacy or copyright. In other words, we could differentiate ourselves from the competition by tackling the problem deeply and approaching it from a unique angle.
This story has happened over and over again, and I can imagine that it applies to most of the people working as engineers today. A big open-ended problem needs to be solved. Rather than approach it in the obvious manner, you can learn from the experts yourself and produce something better as a result of their insight.
On occasion, I will find myself in a dry spell. I will go several weeks or even months without reading anything of much significance. In doing so, I will gradually lose my ability to read with depth for extended periods of time. My mind's muscle for grappling the tough ideas proposed in the words of experts will weaken.
When that happens, here are two things that I do to beef it back up.
Often, the biggest barriers to depth are the simple distractions of life. You have to pick up the kids, cook dinner, or go shopping. There is always something on the docket that gets in the way.
My solution: schedule a session of deep reading so that it looks like any other chore on your calendar. If I don't do this and, instead, try to fit it inside the gaps, I find that my mind will come up with excuses for doing anything other than read.
My mind says, "But Elijah, you need to do your taxes right now. You should do that instead, because reading is boring." Your mind might say something else. I happen to enjoy doing my taxes.
Schedule time for deep reading and you'll be able to say "no" to that procrastination monkey in your head.
I love my job, and I get a huge thrill when I find out a way that I can do it better. So if I need to find something to read, I often start by asking myself, "what is the hardest part of my job right now?"
The answer can be:
From there, I plainly ask ChatGPT: "What is a good book that covers X in depth for a person with Y level of preexisting knowledge of the subject."
To develop the skill of reading deeply, your book doesn't need to be about a subject related to your job. That just happens to work well for me.
A journey of a thousand miles begins with a single step. A book of a thousand pages begins with a single paragraph. The best thing you can do is start.
So, I implore you. If you're stuck on a problem or an idea, try reading a book about it. I have personally found an immeasurable amount of meaning through literature, and it's unblocked me on a number of problems in my professional and personal life. I suspect it can do the same for you.
This post was proofread by Harper.
I have been seeing an increasingly prevalent trend of people showing up in online spaces flaunting that they are writing with the assistance of AI. They seem to be proud of this. They shouldn't be.
The title of this post is somewhat misleading. Local-first software rarely needs to be scaled at all.
It didn't work for me, and if you reading this, it probably won't work for you either.