Writing maketh the 10x Developer. More so the 10x development team.
Writing is thinking. Software is peoples' thoughts on repeat. Developers who can pen their thoughts clearly multiply their impact. This matters even more in group work. Common sense rules; no literature major necessary.

TL;DR: I think the proverbial 10x developer is real and I think they are that good because they habitually "write the thing down". Which makes them better thinkers. Which clarity of mind betters not only their odds of solving known problems better and faster, but also their odds of unearthing the right unknown problems better and faster. Further, they influence people around them to become reliably better, be it by direct mentoring, osmosis, or bar setting. Small reliable improvements compound. That is how they 10x. I, for one, am on it and I am game to help.

Det. Lester Freamon:
Seems that Stringer Bell is worse than a drug dealer . . .

Det. Roland 'Prez' Pryzbylewski:
He's a developer.

— The Wire.

LR: Grab a coffee…

As a software developer, it bothers me when developers disdain writing.

Somehow, our tribe manages to simultaneously deify banging out reams of code and scorn deliberate, long form reasoning. Perhaps we believe that fetching more coin for one's work than another confers natural superiority over them. And since all of one's writer friends must bartend or babysit or ghostwrite or self loathe while feeding SEO machines just to feed themselves, "writing words" must be low value and low status. The Efficient Market cannot possibly have made such a gobsmacking boondoggle, right? … Right? 1

Bloody hell. Writing powers our entire modern civilization. 2

You and I are here because of it. We are software authors. Writing is what we do. Great writing we admire is stunningly difficult because it requires at least the same intensity of mental effort, abstract creativity, and intellectual discipline sustained over long periods of time; months and years 3.

Do we fail to see it because it is blindingly obvious?

Anyway, we don't have to be so high falutin'. We can just look around and think for ourselves.

Det. Roland 'Prez' Pryzbylewski:
"Failure to properly identify myself as a police officer." Sounds like what I was guilty of most of my career, actually.

— The Wire.

On the solo extreme of the spectrum, consider legendary individuals like, say, Dennis Ritchie and Brian Kernighan 4. Consider the legions of C programmers that learned at their feet, using The C Programming Language book (first edition, 1978). Consider the great C programmers who came after them and authored great C programs — severally by themselves — cURL, SSH, ffmpeg, Linux, SQLite to name just a few. You see, C became great not only because it was revolutionary, or benefited from great timing. It became great because Archmages taught how to wield it, by writing with clarity. Even if their book lightly influenced just one of the other great software authors, K&R would each be One Million x developers, by way of the value unlocked by that developer advocacy, to use today's in term. Even at that colossal magnitude of writing-powered leverage, they are not alone by a long shot. If you can think up other writerly developers with like impact, you may grudgingly agree that a bit of hyperbole isn't unwarranted.

Somewhere across the spectrum, toward somewhat larger groups of people, consider popular acquisitions like Whatsapp, Instagram, YouTube etc. Kings' ransoms paid to acquire things built, scaled, and operated by absurdly small engineering teams. Things that have gone on to repay their purchase prices many times over 5. Software is extreme leverage.

Now, a team of developers is a finickier beast than a lone one. The larger the headcount, the easier it is to play Chinese Whispers, and the faster it is to get worse-er. Which is to say, communicating very well with each other becomes mission-critical (if you want to perform at an elite level, of course).

I will bet good money that those unicorn teams that exited for tens of millions of dollars of valuation per developer weren't A-teams merely because they were full of A-players. They were so highly leveraged by dint of being A-grade written communicators.

Well, actually I don't need to bet. Someone already bet big for us.

At the behemoth end of the spectrum, if my six-page narrative memo helps convince my CEO to fund the multi-billion dollar cloud computing project, the memo alone is worth my weight in gold, to my career. I'm not saying "Copy Bezos.". I'm saying notice that Bezos understood the incredible value of hard-coding common sense writing practices into the organisational fabric. Similarly, lore of the "BillG Review" is semi-known 6. And I've heard of "Jerry Yang Reviews" that some veteran Yahoo! s may recount.

I contend these are not random anomalies, merely ones that got popular press coverage. Trust me not… root about the woodwork and you'll see for yourself; the ones and twos, the squads, the giants.

Omar Little:
Look man, I do what I can do to help y'all. But the game is out there, and it's either play or get played.

— The Wire.

That said, mere mortal developers, like Yours Truly, may aspire to such towering international brilliance only at great peril. Not gonna happen, yo. That said, I believe we can aspire to be 10x better than we are today.

"How?" is an open secret… Learning to write — and therefore think — with clarity is how we get there.

However, there is good news and bad news. Luckily, also some relieving news.

The good news?

We have it easy compared to actual writers.

For us, is not about becoming literature majors, and holding down a software job to feed ourselves while we try to get published. For us, the writing is about putting common sense stuff down in a centrally searchable database (a wiki). We can forgo grammar and lyrical prose 7 as long as we at least write structured outlines, using simple phrases and headlines and bullet points and tables and suchlike. It need not be compelling, merely useful. Also writing like a writer is a nice cherry on top. By the way, it is easy to notice that very good developers are frequently very good writers too. Vice-versa, I believe good writers can train to be good developers. The mental wiring is very similar.

The bad news?

There are no silver bullets.

Writing is unnatural, especially for teams. Remember Bezos? Well, even someone with his smarts, charisma, and sweeping authority over his company had to work to make it work…

  • It is a conscious choice. We have to culture ourselves into pervasive, thoughtful, effective engineering writing not just individually, but as teams and whole org charts. LLMs may make writing life easier, but only we can do the work to make it work.
  • It is not a one off activity. Our kind of writing remains useful only through repeat use and progressive revision throughout the life of a software.
  • It requires widespread buy-in. One can't force it. Doing so will reliably cause more damage than good, by violently convincing people that it sucks, because the experience of it will in fact suck for everyone involved. If you find yourself in a leadership position in a writing-averse culture, boy do you have your work cut out. How will you save your people from the septic floodwaters of Meeting Overflow?
  • It is not a template. For example, if you try to copy Bezos and some imagined "Amazon Way", you will at best create a poor facsimile, which will only degrade over time. Just like those who tried and failed and still do, to recreate the Toyota Way. Many are seduced by the allure of their Zen-like philosophy, lofty principles, and relentless success. Few notice how deep their writing practice goes, and how central it is to the ongoing success of their Way. So draw inspiration by all means, but work intelligently with your own context.
  • It will reveal -your- nature and values. If you fear that you might create a nightmare bureaucracy of soul-sucking process documentation and inter-personnel file redirection, you may need to stop right now and do some heavy soul-searching.

Maybe some more great points that elude my mind. But you catch the drift, yes? Ain't no silver bullet.

The relieving news?

It is not an all-or-nothing deal. We can be gradual. We can start small. Even one person is enough to begin with.

We can learn and apply "developerly writing" 8 in a series of wee progressions, apt for the local context. Each progression can be made to systematically compose with the ones that came before, such that composition compounds value. Sustained compounding delivers integer multiplication.

With some perseverance, smart tooling, and some luck (fortune favours the prepared), we can become the 10x version of ourselves.

Mr. Little, how does a man rob drug dealers for eight or nine years, and live to tell about it?

Omar Little:
Day at a time, I suppose.

— The Wire.

  1. So far, The Market, though arguably more humane than monarchs or central planners, appears to have become better at giving society what some people desire, rather than what all people need. The distribution of access to clean air, water, shelter, and a full belly come to mind. Yes, hundreds of millions of us enjoy services and conveniences available only to nobility of the past. Yes, the wretched of society are marginally better off today than up to the early 1900s, staying abreast of shifting poverty lines. Yes, the literate vastly outnumber the illiterate, a sea change. Yet, society at large appears to be worse for the wear and heading for a Really Bad Time. Even so, I remain optimistic that human ingenuity and compassion will prevail, in my own lifetime. We will not yet fail our solitary pale blue dot.

    Malik 'Poot' Carr:
    World going one way, people another.

    — The Wire.

  2. Much to the chagrin of Plato and Aristotle. They weren't wrong, I think. Notice how we've jumped the shark and made the Internet our memory? And, like, if you've been around the Internet lately, you know those dudes were on to something. College degrees correlate weakly with wisdom. Yet, they did not anticipate — could not have anticipated — the fire-like revolution the written word would wend through our social, political, technological, economic, private, and secret lives. One that we have come to take for granted, like the fire before it.↩︎

  3. Gently prod any master programmer who has written one book, and they will confess immediately. That they will gladly choose to cry while banging out yet another soul-crushing Enterprise Web App or cookie cutter ETL job, over writing a second book, and it's not just because a full stomach beats an empty one.↩︎

  4. I'm not invoking Ken Thompson's name because as Brian Kernighan remarks in this Changelog interview, Ken Thompson is a singularity, in a universe of his own.↩︎

  5. Alas, they also performed ritual Enshittification. Corporate karma consumes all.↩︎

  6. Steven Sinofsky's account 010. Our BillG Review on his wonderful blog, Hardcore Software. By the way, Steven is staunchly in the writerly camp. Says he, '“Writing is thinking” is my favorite expression for how to work in a company.'.

    It is really incredible the amount of pushback I see from companies, startups to big, about writing. In particular around the notion that writing is the antithesis of agile. Writing ossifies and cements decision or plans that should change, it is said. My view is that agility comes from planning. Without plans, activities are just brownian motion. And you can’t have plans, especially shared plans, without writing.

    Read more here.↩︎

  7. Like, I can totally publish this disastrous ditty because I don't have a writing career to worry about.

    Code is code.
    Prose is prose.

    Code is prose.
    Prose is code.

    For both are written.
    And both must compose.

    — A terrible poem by Yours Truly
    (No LLMs were harmed in the production of this verse.)

    (I added this line to make the text rag more obviously like a 10-xing hockey stick curve.)

  8. Over the years, all of these types of writing have found they way into my developer toolbox. I use them judiciously, even as a soloist. In no particular order:

    • Commit messages
    • Function names, API names, domain models and contracts
    • Doc-strings and in-line comments
    • Concept notes and Rationales
    • Feature designs
    • Project plans and updates
    • Software architecture descriptions
    • Architecture decision records
    • Checklists and runbooks
    • Usage guides
    • Tutorials and teaching material
    • Bug reports
    • Post-mortems
    • Reviews and critiques of code and of writing
    • Release notes
    • Research notes
    • Think pieces
    • Code and prose riffs for the fun of it