Applied Neuroprogramming: 4D Chess for Finite Minds
This argument is also a book, rendered in the schema the post describes. Scaffold your own book with npm create sveltekitbook@latest my-book.
A far-reaching concept does not fit a finite mind in its native form. To get it from out there into in here, you have to break it into shapes the brain knows how to receive. That is not a writing problem. It is a neuroprogramming problem, and the psychology under it is plain. Working memory holds about four chunks at once. A six-digit number gets remembered by chunking — three pairs, two triples, whatever lets the brain hold the pieces while it processes the rest. Throw eight unrelated items at a person and one drops. The four-slot ceiling is the bottleneck the schema is built around.
People talk about 4D chess as if it were a joke. It is the literal architecture of the human mind. Four compartments, each focused on a different slice of the same situation. A book that uses all four at once teaches the reader to hold four streams at once. A book that wastes the slots, or fills them with the same flavor of content over and over, leaves three of the four dark.
I build the page in four dimensions. The runtime is sveltekitbook on npm. Two scaffolders ship with it — create-sveltekitbook for long-form pages, create-sveltekitslides for talks. None of them write prose. They render it inside a schema that occupies the four slots on purpose.
The four dimensions are continuum, depth, thread, and setting. The continuum is the page sequence — time, or the −5 to +5 spectrum the page sits on, or the chapter ordering. The depth is the ladder inside one page, going from the technical claim a domain reader nods through, down through the frontline manager’s read on consequences for her team, then the translation a non-specialist can carry, to the version a child could repeat back. The thread is the argument that crosses pages — the spine the reader follows from page one to the close. The setting is the cross-categorical anchor, the cultural or empirical context the claim has to sit inside before it is allowed to land.
The schema is src/lib/outline.js. A flat array of plain JavaScript objects, one per page. Every object has fields for those four dimensions, in the same order, every time. Each step down the depth ladder loses precision and gains reach. The reader who lands on page 14 with thirty seconds gets the punch on the bottom rung. The reader with five minutes climbs all four rungs and walks away with the structure, not just the slogan.
The setting dimension is the one writers most often skip and the one readers most often need. I learned this writing about scripture and trauma research, separately, before I noticed it was one mechanism. The verse in isolation is vapor. Who wrote it, who first heard it, what was happening that month in that town — that is where the meaning locks in. A child reading “blessed are the poor” without the cultural setting hears charity language. With the setting, it is a sentence said to people who knew exactly which Roman tax collector lived around the corner and how much he took yesterday. The setting is the field the verse has to sit inside before the claim is allowed to land.
Trauma research is the same mechanism in a different lab coat. I was processing dissociation, in myself, and the first research I found said I was wrong about what I was experiencing. On a longer timeline I found newer research that said, wait — maybe the people who claim total amnesia for the criminal acts they are on trial for, who have something to gain from being crazy, are not the same population as everyone else reporting dissociation. The two papers did not contradict each other. Their populations did. The finding was the smallest part of either study. The population was the bigger part. The forensic claimant and the long-arc survivor are not the same data point. Better framing, it turned out, for a boy who rebels against authority.
The processing left me with one thing the papers did not. The consequence of dissociation is loneliness. That sentence does not have a citation; I went and got it. The social feeds that suck us in are deep holes that karma and emoji reactions will not fill. The best kind of reaction is the one from your spouse or your neighbor.
When our careers force us to speak a language our spouse cannot follow, maybe we are being abusive. Not intentionally; selfishly. The same translation we owe a neighbor is the one we should be giving at home. Doing the translation is the discipline that keeps the people who live with you from going to the feed for what should be coming from you. The schema makes a writer practice that discipline. Practicing it on the page is practice for the kitchen, and practice for the company. Maybe that is why technical departments get held at arm’s length from the rest of the building: the speak does not land in rooms outside engineering, and what does not land starts to sound selfish whether you meant it that way or not. The frontline manager is the rung that has to travel — engineering’s claim plus the consequences her team will feel, in language she can repeat in a planning meeting or in the email a vice president forwards. A page that does not render that rung leaves her to invent it, and what she invents will not be what you would have written.
A book is not a feed, but most books behave like one. They scroll past. A book that gives you the courtroom finding without the population is selling you skepticism. A book that gives you the older finding without the population is selling you a diagnosis. A book that gives you both findings on a timeline with their populations spelled out leaves room for a third data point — the reader’s own experience, which the page does not pretend to know. A book like that feels less like a feed and more like the neighbor: it does not perform a reaction, it leaves room for one.
So the setting slot is a field the schema reserves and refuses to let the page render without. The schema is the harness — not a prompt, not a system message, a cage in the rendered output that forces whoever is writing the page, human or model, to put the cultural and empirical anchor down before the punchline. If the field is empty, the page looks broken on screen. The discipline is built into the layout so the writer does not have to remember to. And here is where the LLM enters cleanly, not as the protagonist. The schema does the same job for a model that it does for me. Ask the model for a chapter and you get a sermon. Ask it to fill the technical, the translation, the setting, and the cost — in that order, with no other freedom — and you get something I can edit with a pencil in twenty minutes.
The four dimensions are why the closing page works. By the last page, the reader has crossed the continuum enough times that the order is muscle memory; has climbed the depth ladder so often they jump to whichever rung they need; has followed the thread; has accumulated enough setting that the page-one claim has somewhere to stand. Now the original claim returns, and it gets to be proved in the same four dimensions. The continuum closes the loop — we are back where we started. The depth lets the proof land at whichever rung the reader is reading on. The thread snaps shut. The setting puts the proof in a place the reader has spent the whole book learning to stand. The close lands because four streams converge at once, which a finite mind can receive only because the schema spent the whole book teaching it to hold four at once.
The slide ecosystem reuses the same shape. create-sveltekitslides ships a deck where every slide is a long page. Above the fold is what the projector shows: the public claim, in technical voice for a roomful of engineers. Below the fold, scrolling down, is the rehearsal note: the same content again, in plainer voice, in fuller setting, with the things you would say only if asked. Same four dimensions, different audience for different rungs. The presenter walks down the page during rehearsal and back up during the talk. The reader at home gets all four at the same URL.
I keep the runtime small for the same reason. sveltekitbook exposes four things — gestures, an inline-markdown helper, a palette, a comments component. Routes, layout, contents, glossary — those land in the project as editable code. The four-dimensional schema has to stay stable. The surface around it should be yours to bend.
The pattern travels. Documentation wants four — claim, example, gotcha, when-not-to-use. Research wants population, method, finding, limit. Sermon notes want text, setting, tension, application. The contents change. The four slots do not.
The pattern scales past the page too. Four teams attacking one problem from four different angles, paid to compare notes for an hour at the end, would compound a problem-space faster than four companies running one experiment each and waiting for the knowledge to leak between them through attrition. People will turn up for that — expenses paid, an hour together, an award and a check. They are already running the experiments on their own time, in second jobs and hobby builds, and most of that work is wasted because nobody puts the four runs in the same room. The retrospective is the schema at team scale: four rungs, each team shows its work, the rest read across the four for what no single run could have told them.
If you write with an LLM, design the four dimensions before you ask for content. The chapter the reader can hold is the chapter that fits the four slots they were born with. The book that lands is the one that uses all four through its closing line.