You need to sign in or sign up before continuing.
Reflection — An Honest Take 8 min

Honest Take — Before You Begin

Honest Take — Module 13: The Long Arc — Public Practice & the 5-Year Capstone #

The compounding is real and most engineers quit at month 8. That sentence is half this module; the rest is a slower way of saying it, plus the document that makes it survivable. Audience growth in a teaching practice does not follow a linear function — it is a long flat period, twelve to eighteen months of publishing into what feels like silence, then non-linear afterward if you sustained the cadence. Forty subscribers, then fifty, then forty-eight because someone unsubscribed. A post you are proud of earning three hundred views and two comments, both typo corrections. This is not evidence the practice isn't working; it is the shape of the practice working. The reason engineers quit at month 8 is not lack of ability — it is the gap between effort invested and visible signal returned, hitting people trained on the deploy-and-it-runs feedback loop. Teaching's loop is years long. The post you ship today finds its real audience in a search result two years from now. The cost of stopping at month 8 is not the eight months; it is what those months would have become at month 30.

The infrastructure decisions are unglamorous and they are the module. Own your domain — every rented platform changes its rules, its algorithm, its monetization, roughly every five years, and the writers who treated rented as home watch the change reprice their distribution overnight; use rented platforms as CDNs, never as the origin server. Invest in the newsletter even though its growth is slower than the blog's, because the list is the one audience asset you actually own and can export: 800 engaged readers who chose to hear from you are worth more, professionally and economically, than 80,000 algorithmic impressions.

Pick your audience model deliberately — stadium (broad weak-tie reach), club (narrow deep relationship), or federation (participatory community) — because the three have different economics, different time costs, and different psychic costs, and the engineer who tries to run all three burns out faster than the one who picks one for twelve months. And share the work-in-progress, per Kleon, without falling into the build-in-public theater trap: sharing drafts, problems, and process compounds trust; live-tweeting your psychology and revenue screenshots is a marketing genre, optional at best. The audience is built from the trail, not the finished pieces.

Then the capstone, and I want to be direct about it because I have watched many curricula die exactly here: everything before this is reading unless you write the document. The trap is treating it as goal-setting — the genre of vision boards and aspirational timelines that everyone produces and nobody executes against. The capstone is an architecture decision record, a form you already know how to write: decision, alternatives considered, rationale, consequences, revisit date.

The sections I would not let you elide: the single subject (one thing you intend to be known for in five years — not three, not "engineering broadly"; the narrowing feels like exclusion because it is, and that is the mechanism), the audience model, the minimum sustainable cadence rather than the aspirational one, and the section engineers skip because it feels like planning to fail: kill criteria. "If after 18 months I average fewer than X readers and don't enjoy the work, I rotate mediums or stop." Practices without kill criteria become zombie practices maintained out of obligation years past usefulness. The kill criteria are a gift to your future self — permission to stop without guilt, decided against pre-committed thresholds instead of against a bad week. Writing "what I will not teach" will be the hardest section and the most useful: each explicit no makes the yes sharper.

Two operational notes and one emotional one. Share the document with exactly one trusted reader — five readers produce contradictory advice and paralysis; one produces the single note that matters, usually about the thing you knew was missing but couldn't see until it was named. Store it somewhere durable that you will actually re-encounter — a git repo, a page on your own domain, anywhere except the document graveyard — and calendar the six-month re-read. And expect to feel slightly underwhelmed when you finish it. The decisions on the page will be smaller and more concrete than the ambitions swirling in your head. That is correct. The point of writing it down is to surface the small concrete decisions and let the swirl go. The disappointment is the practice acquiring a shape — which is the only way a practice survives Year 1 to collect what Year 3 pays.

Conclusion #

Cadence beats brilliance; the flat period is long and normal; survive month 8. Own the domain, grow the list, pick one audience model on purpose. Then write the capstone as an ADR — single subject, real cadence, what you won't teach, kill criteria — share it with one reader, store it durably, and re-read it every six months. The document will feel smaller than your ambitions. That is the point, and it is also the mechanism.

Predictions #

  • You will quit, or seriously consider quitting, somewhere between months 6 and 10. The most useful thing this file can do is inoculate you in advance: re-read it when the moment arrives, because the moment will not feel like a moment — it will feel like a reasonable reallocation of time.
  • One post from your first year will, by year three, account for 30-50% of your incoming search traffic, and you will not have predicted which one. Every post matters more than its launch metrics suggest.
  • You will be tempted at least twice to migrate platforms because a newer tool looks better. Resist without a concrete technical reason; migration always costs more than it appears to.
  • The newsletter will lag the blog for the first year, then quietly become the part of the practice you check first.
  • The "what I will not teach" section will be the hardest to write and the one you revisit most; your kill criteria will feel pessimistic when written and read as wise at month 14.
  • Your one trusted reader will catch a specific elision you could not see — most likely in the audience model or the cadence. Take the note.
  • The 5-year vision will read, at year five, as either obviously right or beautifully wrong, with very little in between. Either outcome means the document did its job: it was specific enough to be falsified.