Flow-Pattern / Who holds the pen?
From monk chants to AI-assisted prose. Does the source even matter?
One simple Flow ritual
22:00 hits.
I put on my headphones and open myNoise, choosing the “Tibetan Choir” track. The steady chant blocks out the room. After ten seconds my shoulders relax; after thirty, the words start to flow. That mix—set time, headphones, ambient sound—is my flow pattern.
But the moment the first lines appear, a louder question shows up: Who’s really holding the pen—me, or the machine beside me?
myNoise – ambient sound generator
Why all the fuss?
A brand-new ScholarshipOwl survey (May 2025, 12 k+ students) shows how fast AI writing has spread:
97 % of Gen Z students have tried at least one AI tool.
1 in 5 already used AI to write a college- or scholarship-essay draft.
That jump has set off the usual alarms about “fake voices” and “cheating.”
Yet Shopify CEO Tobi Lütke flips the story. In a memo he shared on X he wrote:
“Reflexive AI usage is now a baseline expectation at Shopify…
If you’re not using AI reflexively, you’re in trouble.”
His point: good writers (and workers) will treat AI like spell-check; automatic, boring, everywhere.
So the real debate isn’t if we use AI. It’s:
Skill vs. Tool – Does letting a model help erase craft, or just cut busy work?
Credit & Trust – When should writers tell readers a silicon co-author was in the room?
Next, we’ll look back at quills & ink, typewriters, and ghostwriters to see how every new aid sparked the same fight; and what history can teach us now.
Ghostwriters, Typewriters, ChatGPT
Same argument, new gadget. Every time a fresh tool shows up, writers split into doom-criers and cheerleaders.
1883 | The “devilish” typewriter
Mark Twain’s Life on the Mississippi was the first book submitted in typed form. He loved the speed, then raged that the machine was “full of defects—devilish ones,” begging Remington to stop using his name as an ad. The Washington Post
1980s | The word-processor “dishwasher”
When PCs landed on newsroom desks, many veterans said writing would turn “mechanical.” Teacher-editor William Zinsser spent the decade convincing colleagues the new screens were “just a dishwasher for words,” not the end of craft. The Week
2020s | ChatGPT & friends
Today’s models can spit out 1,000 coherent words in seconds, so the cycle restarts. Some see creative ruin; others, a time-saver like every tool before. The pattern is clear: each upgrade shifts the work from toil to choice—and forces us to redraw the honesty line. MIT
2024 | The invisible ghostwriter
Fast-forward: the UK’s Society of Authors is now demanding that celebrity memoirs list the ghostwriters on the cover—proof that authorship credit still sparks heat long after quills and ribbons have gone. The Guardian
Skill vs Output
Half the time, readers can’t even tell.
A 2024 UK–US study found just 50 % of consumers could spot the AI article, and 56 % actually preferred it over the human version. Bynder
Another US survey ran six head-to-head copy battles. AI won every round. Search Engine Land
So, does craft still matter?
Yes, because sameness shows up under the hood:
Style fingerprints: Carnegie Mellon linguists found LLMs lean on the same dense nouns and pet words (“palpable”, “intricate”) far more than humans do. Tech Xplore
Choice of story: Models remix; they rarely set the agenda.
Lived detail: They don’t smell the market stall, misplace the train ticket, or catch your roommate’s raised eyebrow.
Think of AI as the fast-draft intern. It can fill a page, but voice still comes from the edits you keep; and the cuts you make.
The skill behind all skills
“Learning to learn is what experts call the ultimate survival tool, one of the most important talents of the modern era—the skill that precedes all others.” — Ulrich Boser
Today even more than ever, the world won’t let us coast on a four-year degree and 30 years of rinse-and-repeat. We have to keep acquiring fresh skills, then retire and replace them just as quickly. That starts with metalearning—Scott Young’s first ultralearning principle: “First, draw a map.” Before diving into a new field, study how others learned it, spot the best resources, and connect it to strengths you already have. Scott Young
Traditional schooling trains us what to think, then moves on. Today, skills expire faster than software versions, so the real edge is knowing how to pick up new ones; quickly and on your own. The Ness Labs guide breaks that meta-skill into three simple practices:
Experimentation – set up tiny, low-risk tests (“I’ll try X for 15 minutes”).
Metacognition – watch your own thinking: What worked? What dragged?
Iteration – keep the good, ditch the bad, run the loop again.
Quick experiment for this week
Pick one micro-skill: you’ve been postponing; maybe a keyboard shortcut, a prompt style, or a design trick.
Draw your map (10 min): Google or GPT “how to learn X fast,” skim two guides, and list the key steps.
Run a 15-minute first rep.: Then note Plus / Minus / Next to shape round two.
Master the process of learning and every other skill: writing, prompting, editing. It compounds on top.
Your turn:
Where could a draw-the-map first approach speed up something you’re learning right now?
Does AI help or hinder, your own learning loop?
Drop a comment and let’s compare maps. See you next Sunday.