Link post

This post is Part 5 of the se­quence on Bab­ble. After writ­ing Ham­mers and Nails, I figured out that my fa­vorite Ham­mer is writ­ing.

Write about ev­ery­thing. Write it im­me­di­ately. Edit af­ter­wards.

In this post, I tell two sto­ries about the magic of putting words on pa­per. Make your own con­clu­sions about the brain-en­hanc­ing effects of writ­ing.

Then, I share my re­cur­sive blog­ging pro­cess which avoids the ick­i­ness of plan­ning and gets di­rectly to the thick of things. The re­cur­sion makes it easy to ex­pand sec­tions with­out break­ing into struc­ture­less ram­bles. My pro­cess is mostly in­spired by Non­fic­tion Writ­ing Ad­vice.

Fi­nally, I ex­plain why dis­claimers suck.

Mean­ing Injection

In mid­dle school, I car­ried a sin­gle enor­mous binder to ev­ery class, filled with col­ored tabs and lined pa­per. In that binder, I took notes re­li­giously, try­ing var­i­ous tech­niques of or­ga­ni­za­tion and high­light­ing.

I had no good no­tion of why I took notes. After all, I never looked at the notes I took—text­books were always eas­ier on the eyes than my chicken scrawl. I did no­tice, how­ever, that I re­mem­bered things bet­ter when I wrote them down, so for a time my plan was to sim­ply take the notes and for­get about them in my binder.

I even­tu­ally no­ticed what was ac­tu­ally work­ing, and dis­carded the su­per­sti­tion. I left my binder to gather dust and started bring­ing lined pa­per to class to take the notes. When the bell tol­led, I dropped the notes in the re­cy­cling be­fore leav­ing the class­room. I wrote down only the key words and spent the rest of class star­ing at those scat­tered words, in­ject­ing them with the mean­ing. Imag­ine lonely crim­son ink-drops fal­ling into a glass of wa­ter, burst­ing into deli­cious, velvety ten­drils at con­tact. Plop, Plop. That’s how mean­ing in­jec­tion feels.

Sev­eral iter­a­tions later, my study strat­egy is perfectly stream­lined.

I no longer take notes in class. The night be­fore each test, I skim the text­book with a blank sheet of pa­per and write down all the im­por­tant terms—with­out defi­ni­tions; the sheet of pa­per serves as a kind of cor­po­real mind palace. I read through the list once, in­ject­ing the words with mean­ing. Plop, Plop. And that would be enough.

Re­search Notes

Fast for­ward to 2013 and trans­port your­self to my first sum­mer re­search pro­gram. Every Mon­day, we give a brief board about that week’s progress. I mull ideas on pa­per over the week be­fore TeXing them up Sun­day night.

A cu­ri­ous thing hap­pened—all my progress hap­pened on Mon­day and Tues­day. I spent the rest of the week me­an­der­ing around the same ideas, check­ing spe­cial cases and writ­ing up frag­ments of ar­gu­ments. On Sun­day night I write ev­ery­thing down, and the ideas crys­tal­lize on pa­per. They lose their grip on me, and I move on to new pas­tures.

Th­ese days, I re­pro­duce the effect sim­ply by writ­ing up any par­tial re­sults as quickly as pos­si­ble. Of­ten­times, I sit down with an un­cer­tain in­spira­tion and im­me­di­ately try to TeXing it. Usu­ally, it fails catas­troph­i­cally but I gain a piece of in­sight. If suc­cess­ful, the ar­gu­ment pops out on pa­per fully-formed by nar­ra­tive force. My last pa­per—it­self only twenty pages—is a Franken­stein-es­que con­struct built out of thirty ram­bling TeX doc­u­ments.

Struc­ture is Process

Very read­able and nat­u­ral blog posts can be writ­ten re­cur­sively with min­i­mal pre-pro­cess­ing. My pro­cess is mainly de­signed to stream­line brain­storm­ing and out­lin­ing while avoid­ing two failure modes.

The first failure mode is sim­ply writ­ing too lit­tle. Thus, I em­pha­size ra­pidity over con­tent and quan­tity over qual­ity.

The sec­ond failure mode is writ­ing long, de­tailed pieces with lit­tle struc­ture, and find­ing my­self dis­agree­ing from the be­gin­ning of the ar­gu­ment by the time I reach the end. You know the loud­ness bar over the micro­phone icon in Skype set­tings? The length of my blog posts used to fly up and down just like that capri­cious lit­tle bar. I’m in­ca­pable of finish­ing two para­graphs with­out chang­ing my mind, and my pro­cess is de­signed to re­cur­sively de­con­struct into semi-in­de­pen­dent pieces that I can close and set aside.

1. Head­ers are Brainstorming

Brain­storm by break­ing your ar­gu­ment down into a hand­ful (2-5) of the core ideas. Separate in­di­vi­d­ual ar­gu­ments and anec­dotes. Write down ti­tles for each.

For this post, I started with four sec­tions: “Study Skills,” “Re­search Notes,” “AI Risk and Affor­dance Widths,” and “Struc­ture is Pro­cess.” Later on, sec­tion 3 was re­moved—I couldn’t find an an­gle to tie AI Risk with the magic of po­etry. Also, “af­for­dance width” is a won­der­ful idea, but the con­cept han­dle screeches like cheap chalk. Would some­one please syl­la­ble it?

2. In­tros are Outlines

The in­tro­duc­tion to a post defines the topic and sketches the arc of each sec­tion. It teases the reader with sneak peaks. Your post is cob­bled to­gether from un­re­lated shit but the reader won’t no­tice if you tie it to­gether with a metaphor or some­thing.

In the in­tro­duc­tion, it can be use­ful to gather links to back­ground read­ing and place the post in a larger con­text. The reader finds this thought­ful and wel­com­ing. Ac­tu­ally, it’s a con­ve­nient trick to batch-pro­cess all the link hunt­ing.

3. Recurse

Re­cur­sively ap­ply this pro­cess down to the level of sen­tences. Each sec­tion longer than a sin­gle com­plete thought should be bro­ken into sub­sec­tions and out­lined. Each para­graph longer than five sen­tences should be bro­ken into smaller para­graphs.

At the low­est level, it may be wise to treat topic sen­tences as micro-head­ers, to be writ­ten first be­fore filling out the para­graphs with con­tent. It may also be wise to re­flect on the in­te­gra­tion of the (sub-)sec­tions into the whole be­fore filling them out.

On Disclaimers

If you think your words are but pale wa­ver­ing shad­ows of the real con­tent in your heart, you are not alone. You will feel the need to pre­cede your posts with all man­ner of con­tent warn­ings, epistemic sta­tuses, and protests of hu­mil­ity—you will feel no right to cry black and white when all the world is shades of gray.

There are de­cent and gen­uine rea­sons for dis­claimers, but I’ll let you pro­duce those. I pro­pose that we use epistemic sta­tuses and dis­claimers spar­ingly.

What hap­pens to the reader when ev­ery post starts, “epistemic sta­tus: mostly true with a chance of rain?” As the reader’s eyes hit “epistemic sta­tus,” no­tice how they roll to the side and glaze over ever so slightly.


She knows it’s 10% about epistemic and 90% about sta­tus. Magic brain juice [cita­tion needed] takes over and as­so­ci­ates that in­stant “ugh” with open­ing your posts.

In­stead of con­di­tion­ing read­ers to hate us, I pro­pose we re­turn to a saner time, where the fact that your words are but a pale wa­ver­ing shadow of the grand, mys­te­ri­ous truth in your heart is the de­fault as­sump­tion about hu­man com­mu­ni­ca­tion. Where truth is a dance of suc­ces­sive ap­prox­i­ma­tions yet no step in that dance re­quires adult su­per­vi­sion. Where quib­bles over cer­ti­tude are ban­ished to the com­ments sec­tion where they be­long.