Skip to content

The Archive · Complete back-catalogue

Every edition,
since the first one.

A browsable record of all 1,847 daily editions published since April 2019 — one resource per weekday, hand-picked by six full-time curators. Search by topic, role, or reading time. No algorithms; no infinite scroll.

1,847 editions 0 missed days 6 curators Avg. read 4 min 12 sec
An editor's desk with an open notebook, fountain pen, and coffee mug in warm morning light.

The Record

All editions, indexed.

The live editorial record — every weekday since launch, in reverse chronological order. Filter the catalogue by who you are, what you're working on, or how much time you have.

  1. No. 1,847 Wed · Mar 26, 2025
    The Quiet Tyranny of the Generic Dashboard

    sylver.dev · Curated by Maya Okafor · 5 min read · Design & Craft

    A working designer's autopsy of the eleven boxes that ship in every analytics product — and the three you'd actually use. Filed under: stop rebuilding Mixpanel.

  2. No. 1,846 Tue · Mar 25, 2025
    What Postgres Can Do That Your ORM Hides From You

    pganalyze.com · Curated by David Lin · 8 min read · Engineering

    A senior engineer's letter to the next person who reaches for a NoSQL store to solve a join problem. Includes the four queries every backend should know cold.

  3. No. 1,845 Mon · Mar 24, 2025
    The Post-ZIRP Operating Playbook, Five Years In

    stripe-press.com · Curated by Maya Okafor · 11 min read · Operations

    What the surviving 2021 cohort actually changed about how they hire, price, and forecast. Not a manifesto — a working note from five CFOs who made it through.

  4. No. 1,844 Fri · Mar 21, 2025
    Writing for Engineers Who Hate Writing

    aeon.co · Curated by Helena Ruiz · 7 min read · Writing

    A short essay on treating prose the way you treat a code review: one idea, one verb, one paragraph at a time. Read it before your next design doc.

By the numbers

Six years, no missed days,
one link per morning.

  • 1,847 Editions published since April 2019
  • 0 Missed publishing days — ever
  • 6 Full-time curators on staff
  • 4:12 Average read time, per Substack

Searchable across topic, role, and reading time. Filtered by a 6-person editorial team — no algorithms.

Four long arcs

What the archive, read in order, actually traces.

I.

The rise of AI product UX.

Beginning in the spring of 2022 — about eighteen months before the rest of the industry noticed — the editors started linking the first working essays on how to design a chat interface that didn't feel like a 2003 help-desk widget. The arc now runs to 140+ editions and includes a working critique of the empty-state pattern, four teardowns of consumer copilots, and a quiet, definitive piece on prompt-input ergonomics that has been cited inside the design orgs at Figma and Linear.

II.

The post-ZIRP operating playbook.

From the rate hikes of mid-2022 forward, the editors tracked how the surviving 2021 cohort actually changed — pricing pages rewritten in three weeks, growth teams cut by a third and rebuilt around sales-assist, board decks rebuilt around contribution margin instead of top-line. Forty-eight editions form a working playbook no consulting firm has matched. The arc is the most-requested body of work in subscriber mail.

III.

The craft of writing for engineers.

A quieter, longer arc — ninety editions over three years — on the specific writing problems engineers face in design docs, incident postmortems, RFCs, and the dreaded all-hands slide deck. Treats prose the way the rest of the catalogue treats code: one idea per paragraph, one verb per sentence, kill your adverbs. The running thread has its own page in the archive and is downloaded as a 38-page PDF by teams at Stripe and Vercel.

IV.

Slow tools for thinking.

The oldest arc, beginning with the very first edition in April 2019: links to notebooks, paper systems, plain-text archives, outliners, and the occasional essay on why you should write fewer notes. Sixty-one editions and counting, and the one the curators quietly consider the spine of the whole publication.

From the curators

How the catalogue is built.

  1. Every edition is filed by hand. The editor who curates it chooses the three tags — a topic, a role, and a reading-time bucket — and writes the one-paragraph Why it matters that ships in the email. We don't auto-tag, we don't auto-summarise, and we don't trust algorithms to count words. A human reads the piece, then files it. That human process is the entire product.

    — Maya Okafor, co-founder & editor on filing & tagging
  2. Reading time comes from Substack's own analytics — the median seconds a reader actually spends on the edition, not a word-count guess. That's why the average is 4 minutes 12 seconds, not the inflated five we'd get from an algorithm. We publish the real number because the archive is more useful when you know what you're signing up for.

    — David Lin, co-founder & editor on reading time
Read the full editorial method →