If you like reading then you should write too.

Three Things Writing Will Do For You

  1. Writing makes you smarter. Thoughts are ephemeral, fast and fleeting. Writing is frozen thought. It is easier to determine if an idea that is written down than an idea that is merely thought or spoken. Writing down your ideas shows you whether they have substance. Detecting when you're wrong is the first step to becoming right. There are many ways to record thought. Writing is the best for improving your rationality because there is nothing to hide behind. Writing is the swimsuit of ideas.
  2. Writing makes you a better communicator. The skills you learn by writing transfer to speaking. Being good at speaking makes you more persuasive.
  3. Writing scales social connections. The best job opportunities I've ever received all came from my writing. The smartest people I've ever met all came from my writing too.

What do I write? How do I write?

I assume you are proficient in English. Besides a basic mastery of language, good nonfiction has three qualities.

  1. It gets to the point.
  2. It is interesting.
  3. It is grounded in reality.

Get to the Point

Text is a static medium. The listener of a podcast must keep up with the audio stream. The reader of a blog can take as long as he/she needs to understand an idea. Text is low-density too. In terms of bits, the reader of a text document processes less raw information than the viewer of an image. Text concentrates a reader's attention on a single idea.

The reader of a blog tends to think harder than that same person watching a video or conversing in real time. The epistemic rigor is correspondingly higher.

We are products of our environment. An author's environment is the written word itself. I write differently from how I talk. I write so differently that readers of this blog are often surprised to discover I'm a kind levelheaded person in real life. Mediums overpower messages.

me before reading your website: [Lsusr] seems like a really even-keeled friendly guy who'll get along with most people

me after reading about your falling into the Dark Night of the Soul while on a hike with people you didn't like: Gosh, glad I read that AFTER cohosting an event with him

―from a conversation with Etirabys

Paul Graham recommends you write like you talk. This doesn't mean you should write a transcript complete with "um"s and "uh"s. It doesn't mean you shouldn't edit. What Paul Graham means is your writing shouldn't be pretentious. Get to the point.

It's possible to go too far in the opposite direction you end up self-deprecating. Self-deprecation is for stand-up comedy. It falls flat when presented in a two dimensions. Get to the point.

Art is often about the experience. I don't read Aaron Diaz's webcomic Dresden Codak to learn about transhumanism. I read it because it's a masterwork of visual art with a riveting story. Nonfiction writing is about the ideas, not the experience. Get to the point.

Most beginning writers weigh their text down with unnecessary crap. Superfluity doesn't just waste the reader's time. It obscures the central thesis. Obscuring logic makes sense if logic is your enemy. If you're a rationalist then logic is your ally. Expose your logic.

Practice

  1. Write a blog post at least 300 words long. Make a note of how many words it is.
  2. Edit out all pretension.
  3. Edit out all self-deprecation.
  4. Edit out anything else that talks about yourself. Don't write "As a ___". Truth is unitary. It does not depend on who you are. Don't write "I believe ". If is true then write "". If is untrue then don't write .
  5. Edit out any references to untrue statements. If is untrue then don't write " is untrue". Don't write in the first place. Write "" (which is true) instead.
  6. Identify the central thesis. Edit out anything that doesn't directly pertain to the central thesis.
  7. Remove anything that isn't necessary to prove your thesis. If you have three reasons to believe then remove the two weaker ones.
  8. Remove unnecessary words like "very". Write in absolutes.
  9. Don't use multiple words when a single word will do.
  10. Make a note of how many words are left.

Invent Interesting Ideas

If an author is bored then the reader will be too. Ideas I understand are boring. That's why I never write about anything I understand.

Writing is an exploration of ideatic space. Your initial hook is no more important than your original plan when you go on an adventure in meatspace. When I went on a cross-country motorcycle trip my original intention was to venture down Baja California. That never happened. I ended up spending half the trip at the bottom of a cliff reassembling a motorcycle engine in the sand. We barely even made it to Mexico.

Does that matter? No. If the trip went as planned then I never would have been threatened by a tweaker with a gun while pinned under my DR650 on a dark forest road in the rain. Writing is like motorcycle riding. It's not about the destination. It's about the enemies you made along the way.

You have enemies? Good. That means you've stood up for something, sometime in your life.

Good writing is though-provoking. A provoking idea is an idea that seems wrong. The best ideas to publish are those that seem wrong but are actually true.

How do you find provoking-yet-true ideas? Write down things that seem wrong. Edit out the ideas that actually are wrong.

Use precise language grounded in observable reality. Everything you write should be either true or false. Wishy-washy claims reflect muddled thought. Vague language demonstrates the absence of thought entirely. Using concrete language is like power washing your brain. When you put your ideas into concrete unambiguous language it becomes readily apparent which of them which ones are dangerous and which are garbage.

All good ideas are dangerous because their very existence is a threat to the status quo. When people complain about "infohazards" and "threats to the community's reputation" that's confirmation you're writing about something that matters. Dangerous writing is never boring.

Practice

What's an article you wish someone else would write for you? Write that.

Write From Personal Experience

The best authors all write from personal experience. For this to work, you need to have experience doing interesting things. George Orwell fought for antifa in the Spanish Civil War before writing Nineteen Eighty-Four. Karen Pryor trained dogs and dolphins before writing Don't Shoot the Dog.

With my tongue I can feel the silencer holes we drilled into the barrel of the gun. Most of the noise a gunshot makes is expanding gases, and there's the tiny sonic boom a bullet makes because it travels so fast. To make a silencer, you just drill holes in the barrel of the gun, a lot of holes. This lets the gas escape and slows the bullet to below the speed of sound.

You drill the holes wrong and the gun will blow off your hand.

The building we're standing on won't be here in ten minutes. You take a 98-percent concentration of fuming nitric acid and add the acid to three times that amount of sulfuric acid. Do this in an ice bath. Then add glycerin drop-by-drop with an eye dropper. You have nitroglycerin.

Mix the nitro with sawdust, and you have a nice plastic explosive. A lot of folks mix their nitro with cotton and add Epsom salts as a sulfate. This works too. Some folks, they use paraffin mixed with nitro. Paraffin has never, ever worked for me.

Fight Club by Chuck Palahniuk

It doesn't matter whether you're writing about AI alignment or how to paint one's nails. The best writing comes from personal experience. The best history books are autobiographies. It is only by testing our ideas against reality that we can differentiate between right and wrong. Nothing is more empirical than direct personal experience. A randomized controlled trial conducted by someone else is mere hearsay.

The best writing is derivative. We copy each others' techniques. We tell the same stories over and over again. But the best writing is never purely derivative. Memories are non-transferrable. You can't steal another's heart and soul.

Practice

What is something everyone in your in-group believes but which you have directly observed to be false? Write an essay explaining what you observed.

32

4 comments, sorted by Click to highlight new comments since: Today at 1:18 PM
New Comment

So, I think your message has two contradictory parts:

  1. Write from your own experience, in your personal voice, and so on

  2. Write like this! Remove X, remove Y, etc

This took me a while to realize, but (1) is the more valuable advice. Different people's personalities are tuned to different ways of writing, and any trick you learned elsewhere and try to layer "on top" (yes, even brevity) can get in the way. The goal of writing is partly to discover how you write. Not how you write at current skill level - that probably just sucks - but something more like an idealization of how you read.

There was some book that a friend once mentioned to me, I don't know the author or how the book is called, but an anecdote stuck with me. It's about a guy who's trying to learn opera singing, and keeps trying to find some imagined teacher, "the man with the voice of a red bull", a phrase that came to him in a dream. Quite an image for an idealized opera singer, don't you think? Well, he spends years and never realizes that the thing he was dreaming about was his own voice - in a potential future where he chased the art in exactly the right way.

To me that story sums up what art should be about. You've got to have a dream; it has to be your dream, and it will suggest ways to chase it. To some people the dream tells to omit unnecessary words, to Nabokov it suggested something quite different. And to you personally, I guess the question is: are short sentences getting straight to the point really the most enjoyable thing about writing to you? It's certainly the thing that Paul Graham likes, so more power to him, but you're allowed to like other things too.

I was never a fan of this advice to remove all reference to the self when making a statement. If you think everything is broken or complicated and you don't think you have strong reasons to think you're doing any better than average, why pretend that everything is fine and we can just be authorities on the way that things are rather than how they impressed us as being?

My English teacher took off grades every time I explained things as though from a perspective as humble and precarious as honest, good epistemics require me to report; using terms like "I think" and "it may be the case". 

Now, I could understand if the idea was that no one knew anything and we were all just roleplaying and that school was there to teach me to roleplay. But her defense against my skepticism towards non-subjective reporting was, and I quote, "there's a system for how these things work; it'll be explained in later grades."

It was at that time I was getting truly fed up with my educators. I will not lie about my confidence in my authority.

I don't read Aaron Diaz's webcomic Dresden Codak to learn about transhumanism. I read it because it's a masterwork of visual art with a riveting story. Nonfiction writing is about the ideas, not the experience. Get to the point.

That's a little strong. Nonfiction is about ideas, but we generally care about ideas because they are connected to experiences that matter to us (positively or negatively), and it's hard to convey ideas without conveying any experiences. In fact, this very post occasionally stops to do things that I would call conveying experiences, such as when you relay Etirabys's experience of you at different times. Arguably even you mentioning Dresden Codak in the previous sentences is evoking an experience. :)

Nonfiction conveys information. 

Fiction evokes emotion. [...]

Though the ostensible purpose of nonfiction is the conveyance of information, if that information is in a raw state, the writing seems pedestrian, black-and-white facts in a colorful world. The reader, soon bored, yearns for the images, anecdotes, characterization, and writerly precision that make informational writing come alive on the page. That is where the techniques of fiction can be so helpful to the nonfiction writer. [...]

TRADITIONAL NONFICTION: New York City has more than 1,400 homeless people.

BETTER NONFICTION: The man who has laid claim to the bench on the corner of 88th Street and Park Avenue is one of New York City’s 1,400 homeless people.

(Sol Stein, Stein on Writing)

There are some points here that are tried and true wisdom, passed down through the great tomes on writing, such as how to cut the fat. I especially like your thought experiment angle for coming up with ideas. I might give that a shot soon!

There are also a few points I'd be inclined to tell a new writer to take with a grain of salt. Limiting oneself to absolutes may damage credibility. “The best history books are autobiographies” is a highly subjective statement, and I’d challenge it. In this way, one can aim for “concrete” and land on shaky ground.

Writing precisely what you don't know is great for one particular style, i.e. the "this is what I learned today" style of blogging. And it can be great for staving off your boredom, but there's also the reader to consider. Some subjects and styles - and especially if the authoritative tone you call for is used - demand a writer who really knows their subject. A veteran who has seen it all and can comment on the history of things. Of course even that writer at the frontier of their field should always continue learning, both forwards and sideways to combine their knowledge with adjacent fields.

I'm largely with you in terms of Getting To The Point, though again, this is about style and subject matter. Nonfiction can entertain as well as inform or persuade, and as long as the reader is engaged, going on a few fulfilling detours can be fun/humorous/insightful. Some of the best stuff barely even has a point, or the point might be elusive (Hunter S. Thompson, Charles Bukowski). Intro paragraphs/sections from new writers can often fail to hook, but there's a place for flowery prose and humour in nonfiction.

New to LessWrong?