A Forever Gift
Campus
Preview — Pre-publication draft, not yet board-endorsed. See something to fix? Tell us →
America's School Trust Library
Architectural plan view of the Writing Room — desks with manuscripts in progress, a central conference table for coauthor meetings, flanking bookcases, and side reading nooks. Hand-drafted in oxidized navy ink on parchment, in the visual register of the Library's Spatial Discovery Blueprint.

The Eighth Anchor: Interlude — Nine Months in About a Minute

535 words · Substrate Sacred_Compact_v5.0 · Last synced May 8, 2026

← Back to Stewards of the Republic
Stewards of the Republic — all sections
  1. Prologue — A Forever Gift, Now What?
  2. I. The Question
  3. II. The Sacred Compact — How America Built a Forever Promise
  4. Interlude — The Institution That Held the Future
  5. III. The Drift — A Typology of How Forever Promises Come Apart
  6. IV. The Pattern — Why Trusts Drift, and Why They Are Sometimes Seized
  7. Interlude — Nine Months in About a Minute (you are here)
  8. V. The Counter-Architecture — Seven Anchors Plus a Watchful Crew
  9. V.5. The Knowledge Stack as Demonstration
  10. Interlude — Two Kinds of Mind
  11. VI. The Coming Trusts — Designing for the AI Age
  12. VII. A Civic Practice for the AI Age
  13. VIII. A Letter to the Architects

Interlude — Nine Months in About a Minute

Nine months of my life copied in about a minute.Dave Sullivan

A few years after the Tektronix episode I left finance for the College of Business at Oregon State, and not long after that I cofounded, with my best friend Mike Magee, the Corvallis PC Club. The club was the local instance of a pattern then repeating itself, in slightly different forms, in college towns and technology corridors across the country: people who had bought the new machines were trying, mostly without help, to figure out what to do with them. I became the club’s librarian. The job was less grand than it sounds. We had boxes of floppy diskettes containing the basic programs people needed — text editors, simple spreadsheets, drawing utilities, communications software — and we sold them, roughly at cost, to whatever member needed a copy. The mechanics were primitive. The role was the role I am, forty years later, filling again at a different scale.

In the early 1980s I started writing what became Computing Today, the first college textbook to treat personal computing as a discipline worthy of serious study rather than as a lab manual. I worked on it eight months at twelve-hour days, seven days a week, the same sprint feeling I am working under now, only a younger man’s body to do it in. I wrote in WordStar, embedding formatting codes by hand in the text so that each chapter would print in an understandable way on a dot-matrix printer. My editor was a chain-smoking woman named Kathy Fields at Houghton Mifflin in Boston. Partway through the sprint, I bought a pair of five-megabyte hard drives — the second one I bought specifically so I could back up the manuscript I was producing on the first.

The day I finished writing the book, I asked the operating system to copy the entire manuscript — over a megabyte of work, nine months of my life — from one drive to the other. The drives clicked away for almost a full minute. They were done. The whole book, copied across, in about a minute. It felt amazing. Nine months of my life copied in about a minute. That sentence is the one my hand still wants to write whenever a new cognitive technology arrives and shifts the cost of some operation that used to be expensive. It is the sentence behind the cognitive-technology argument the next chapter is going to make.

In the second edition I gave a name and a chapter to the failure mode that worried me most about the new machines: garbage-in, gospel-out — a variant of the older mainframe-era saying GIGO, applied to the personal-computer printout. The seductive credibility of nicely-formatted-but-wrong computer output. The phrase did not originate with me; the concept goes back to the 1950s, when mainframe operators learned that an authoritative printout could be confidently and entirely wrong. What was new in 1996 was that the audience for the wrong printout had expanded from a few hundred specialists in glass-walled computer rooms to tens of millions of people sitting at desks, and the failure mode mattered correspondingly more. Spreadsheets in 1996; AI hallucinations in 2026. The same warning, written by the same person, thirty years apart. The cognitive technology has improved enormously. The failure modes the technology invites have not. Some of what I learned in 1984 about what computers ask of the people who use them turns out to apply, with adjusted vocabulary, to what AI is now asking of the people who use it. I will let the book make that argument; the autobiography belongs here, in the interlude.

Preview publication — cite the linked primary sources; interpretive framing is provisional.