You sit down to write the opening. Two hours in, you've produced a paragraph that reads fine but doesn't hold. You move on, write the rest of the piece, come back to the opening, and the paragraph you slaved over has to go — it was building a setup for a piece you hadn't written yet. Karlsson's move at Step 5: don't write the opening or the ending first. Write them last. Once the piece exists, the opening is almost always obvious — it's already in the piece somewhere, usually a throwaway line in the middle that, moved to the top, makes the whole thing click into place.1
Karlsson is direct about what goes wrong when he tries to write the opening first:
"If I start writing an opening and I know this is going to be an opening, I get really pretentious. I get really tense. It's overworked. It takes a lot of work. I'll sit for two, three hours working on that opening and then finally I get through it and I can start writing through the piece and then I reread it and it doesn't hold. It doesn't hold."1
The mechanism is recognizable. An opening is supposed to set up everything that follows. When you don't yet know what follows, you over-design the opening to be flexible enough for many possible pieces. The result is generic. It's optimized for a piece that doesn't yet exist, and once the actual piece exists, the opening's flexibility looks like vagueness.
The other failure mode is labor signal. A reader can usually feel an overworked opening — the prose is tighter, more polished, more pretentious than the rest of the piece. The contrast tells the reader the writer wasn't sure yet, and that telegraphs uncertainty to the rest of the piece.
Once the piece is whole, the opening reveals itself. Karlsson's example from Looking for Alice:
"In the middle of the piece there was just a throwaway line where I said that someone once asked Gertrude Stein if she was a lesbian and Stein answered 'no, I just like Alice.' And I was like, wait — if I move that to the first line, that's just going to make the whole essay kind of come together. So it just arrived in the writing. I didn't have to think of it because once the essay was there and I could see clearly on paper what it was about, then it was obvious what the opening should be."1
Three things have to happen for this to work. First, the piece has to be drafted (Step 4 done). Second, the writer has to reread with fresh eyes — looking for what's already there that could function as opening, not trying to invent something new. Third, the writer has to be willing to extract a line from its current home and move it to the top, which usually requires light surgery on the surrounding text.
The reason this works: a throwaway line that earns the opening position is one that compresses the piece's actual insight in a single beat. The writer couldn't have written that line as an opening because they didn't yet know what the piece was about. The line is the piece's truth speaking through the writer before the writer consciously recognized it. Promoting it to the top is recognizing what was already there.
Karlsson applies the same logic to endings, though he discusses them less explicitly. The ending is also obvious once the piece is whole, because the ending is the thing that resolves the animating question — and you can't know what resolves the animating question until you've written through the resolution. Trying to write the ending first produces a destination without a journey.
The full Step-5 logic is:
Karlsson notes that this pattern recurs in his work: "Often it's like that — I can just take something and move it to the top once I've understood what the true insight of the piece is."1 The pattern is not occasional; it's the standard method.
A common writing-school move is to write the opening as the thesis — the sentence the piece will defend. The thesis-opening assumes the writer knows the conclusion before drafting. For analytical essays, this can work. For Karlsson's discovery-driven pieces, it can't — the conclusion is what emerges from the writing, and the opening that previews the conclusion can only be written after the conclusion exists.
This is also why the Karlsson method is hard to teach as "write thesis first." His pieces are not theses with evidence; they are journeys with a recognition at the end. The opening's job is to invite the reader on the journey, not announce the destination. You can only write the invitation once you know what the journey turned out to be.
The principle is named directly at lines 102-110 of the transcript. The Looking for Alice Gertrude Stein example is the canonical case Karlsson uses to illustrate. He also describes the failure mode of writing-the-opening-first in detail — the pretentiousness, the two-to-three-hour labor for a paragraph that doesn't hold, the rereading and recognition that it has to go.
The method assumes the piece can be redrafted. Editors and writing-school contexts often require the writer to submit an opening before they've drafted the piece (the lede, the pitch, the proposal). Karlsson's method is incompatible with this workflow — the opening can't be known until the piece is drafted. A writer trapped in editor-required-opening-first workflows has to either: (a) draft the whole piece secretly before pitching, (b) pitch a generic opening they know will be replaced, or (c) write a different kind of piece (thesis-driven rather than discovery-driven) that supports opening-first composition.
The method also assumes the writer has the detachment to find the right opening line in the existing draft. This is harder than it sounds. Writers in close-to-the-piece state cannot easily see which line carries the compression. The temporal-gap discipline (write Step 4 on Tuesday, find the opening on Friday) often produces better results than same-day extraction.
Orlean's the-wait-what-hook-juxtaposition and Karlsson's beginning-written-last converge on what an opening should do — produce a "wait, what" reaction in the reader — but diverge on how to produce it. Orlean engineers the hook by selecting an opening detail that is jarring against the rest of the piece (the Larroche lead). Karlsson lets the opening surface from the body of the draft (the Gertrude Stein line). Both produce the wait-what effect. The split tells you the wait-what is not a stylistic device but a structural property — it can be engineered or discovered, but it's the same thing either way. The implication: a writer who can do both (engineer-when-needed, discover-when-possible) has a wider operating range than a writer locked into one mode.
Pink and Deen both teach opening-as-hook-as-craft — the hook is something to construct deliberately, with technique (Pink's reader-obligation framing, Deen's 10-second-test compression). Karlsson treats hook-construction as downstream of piece-construction — the hook can't be made; it can only be recognized once the piece exists. The contrast reveals that Pink/Deen are teaching how to do something when the piece-coherence is already established, whereas Karlsson is describing the move that creates the piece-coherence to begin with. The two views aren't competing; they're operating at different stages of the same process.
Finding-the-opening-from-the-body is not specific to essay-writing. It shows up wherever a maker has to introduce a finished work that emerged from a process they couldn't predict in advance.
Behavioral-mechanics: The Frame and Anchor (Influence) — influence-practice teaches that the opening anchor sets the field for everything that follows; whoever controls the first reference point controls the negotiation. Karlsson's beginning-written-last move says the same: the opening line controls the field for the piece. Watch a skilled negotiator at the start of a deal. They don't open if they can help it — they let the other party open so they can see what field the other party is operating on before committing to one. Writers face the same trap. You can't pick the opening before you know what field you want, and you can't know what field you want until the piece has drifted into it. Karlsson's opening has to come from the field that emerged in the draft, not from a guess about the field before drafting. Same discipline as the negotiator who waits.
Eastern-spirituality: Sankalpa and Completed Intention — a sankalpa is an intention stated as if already accomplished; the practice teaches that effective intention requires the practitioner to know the completed form before stating it. Karlsson's opening is a sankalpa-shaped sentence: it can only be true once the piece is complete. Try this on yourself. Write down a sankalpa-style intention for something you've never accomplished. Read it back. It sounds generic, hopeful, weightless. Now write the same intention in retrospect of an experienced completion — recall a specific moment when the completed state was real for you, and write from inside that memory. The sentence has charge now. The same logic governs essay-openings. Statements made before the completion produce generic intentions; statements made after produce specific ones with real charge.
Narrative-craft: Cold Open Then Credits Hook Architecture (Gelb) — Gelb's cold-open works the same way: the hook stands outside the main arc, often filmed last after the documentary's truth has emerged from the edit. The opening is engineered in reverse, once the work knows what it is. Open-from-the-body crosses formats — essay, documentary, novel, screenplay. Anywhere a long work emerges through process rather than pre-planning, the opening is downstream. Try writing a novel's first sentence on day one. Reread it on day 365 after you've finished the draft. You'll probably move it to chapter three or cut it entirely. The cold-open principle is just acknowledging this in advance and refusing to commit early.
Psychology: Retrospective Narrative Coherence — autobiographical-memory research shows humans construct life-narrative coherence retrospectively, not prospectively. The person at 50 narrates their twenties as if it were heading somewhere; the person in their twenties had no such coherence — they were drifting through unconnected events. The narrative was added in retrospect once the destination was known. Karlsson's beginning-written-last is the writing-craft instance of the same cognitive operation. The piece, in process, drifts through clusters without obvious connection; the coherent opening can only be written once the destination has been reached and the path back can be retraced. What looks like a structural-craft question (when to write the opening) is actually a psychological-tolerance question — can you sit with incoherence long enough for retrospective coherence to emerge? Some technically skilled writers can never adopt the method because they can't tolerate the in-process drift. Less polished writers often find it natural, because they're less invested in the appearance of order during the drafting phase.
The Sharpest Implication
The most common failure mode in writing is defending the opening you committed to. You wrote a great opening on day one. It's polished, it's clever, you're proud of it. Then the piece drifts somewhere else over the next two weeks, and your opening no longer fits — but you don't want to lose it because you worked so hard on it. So you bend the piece to fit the opening rather than the opening to fit the piece. The result is an essay whose first paragraph is fine and whose body is misshapen. Karlsson's method makes this failure structurally impossible — there is no committed opening to defend. The piece is allowed to drift toward whatever it actually wants to be, and the opening is found afterward, from inside the drift. Writers locked into outline-first habits routinely sabotage their best work this way without noticing. The fix is not better outlines; it's no outlines.
Generative Questions