What to Learn First — a Map of What Unlocks What

Sooner or later, every wide-minded person stands at the foot of the whole mountain of human knowledge and asks the only sensible question there is: where do I start?

You have probably typed some version of it into a search bar. What to learn first. Best order to learn things. What should I know before economics, before machine learning, before philosophy. And the internet handed you listicles — ten skills for the future of work, seven books that will rewire your brain — written by people who have never once felt what you feel. Which is not ambition, exactly. It's appetite. A pull toward more or less everything, and no idea which door to open first.

Here is the answer I wish someone had handed me twenty years ago: before you plan what to learn, map what unlocks what.

Knowledge is not a list. It is not a shelf you clear left to right. It is a web with direction in it — certain threads must be on the loom before others can cross them. Weavers have always known this, and they built it into their craft's oldest distinction. The warp comes first: the long, strong, unglamorous threads stretched tight across the frame before any pattern begins. Then the weft — the color, the figure, the part everyone admires. Reverse the order and the cloth falls apart in your hands.

Knowledge has a dependency structure

Every field leans on other fields, and the leaning is not symmetrical. You can read history without statistics, but you cannot read modern economics without it. You can love popular physics without calculus, but you cannot do physics without it. Geopolitics without history and economics underneath it is just the news with a soundtrack.

Universities have always known this too, and they encode it in the least romantic word in education: prerequisites. Open any course on MIT OpenCourseWare — the entire MIT catalog, free, no login, no catch — and the syllabus page states plainly what the course assumes you already hold. Electricity and Magnetism assumes Classical Mechanics and single-variable calculus. Algorithms assumes discrete mathematics. Nothing is hidden. The dependency graph of human knowledge already exists — it has just been scattered across ten thousand syllabus pages, where nobody thought to gather it up for people like you.

Notice, before we go further, what this reframe does to old guilt. The feeling of being "bad at" a field is very often nothing but a missing prerequisite — you were handed the weft before the warp and told the tangle was your fault. The person who bounced off economics at nineteen wasn't too slow for economics; they were reading statistics-shaped sentences without the statistics. The field wasn't closed to them. It was mis-sequenced.

So let me gather the part of the graph you actually need. Call it the Thread Map: not a curriculum, not a ranking of what matters — a map of which thread, pulled first, loosens the most knots.

The Thread Map

Six dependencies do most of the work. Hold them loosely — no honest map is complete — but these are the load-bearing ones, and they cover the territory most weavers actually want to cross. One reassurance before you read them: pulling a thread first does not mean loving it most. The warp is rarely the point of the fabric. It is what lets the point exist.

Mathematics unlocks physics, computer science, and economics. Not all of mathematics — that misunderstanding alone has stopped more learners at the gate than any real difficulty. Two threads carry nearly everything: calculus, the language of change, and linear algebra, the language of structure. With those two in hand you can walk into physics, machine learning, statistics, and the quantitative half of economics without the door slamming. If you want one concrete place to stand, Gilbert Strang's Introduction to Linear Algebra — taught alongside his 18.06 lectures, free on MIT OpenCourseWare — has gently carried a generation of self-learners across.

Statistics unlocks nearly every science — and everyday causal reasoning besides. This is, quietly, the highest-leverage thread on the map. Statistics is how you read a medical study, weigh a headline, notice that an alarming "fifty percent increase" is two cases becoming three. Psychology, medicine, economics, ecology, sociology — their journals are written in this language. And unlike most warp threads, its first payoff arrives in week one, in the morning's news. OpenIntro Statistics is a genuinely good textbook, free by design, if you want to begin tonight.

History and economics, together, unlock geopolitics. Neither alone suffices. History carries the grudges, the borders, the long memory of nations; economics carries the incentives — who needs whose oil, whose shipping lanes, whose semiconductors. Read geopolitics without them and you are memorizing other people's opinions. Read it with them and you can generate the analysis yourself — which is the difference between watching the loom and weaving on it.

Physics unlocks engineering and cosmology. Mechanics first — force, energy, momentum — because every branch of engineering is mechanics in work clothes, and every cosmology documentary that has ever kept you up past midnight is mechanics at scale. The Feynman Lectures on Physics sit online, complete and free, by arrangement with Caltech; the classroom workhorse for half a century has been Halliday and Resnick's Fundamentals of Physics. Between the poet and the workhorse, you are covered.

Psychology unlocks negotiation, design, and marketing. All three are applied psychology wearing different name tags. Every interface, every sales page, every difficult conversation is a wager about what a mind will do next — and the wager improves immensely once you have studied minds on purpose. One standard introductory text (David Myers's Psychology has anchored the course for decades) will do more for you here than a shelf of persuasion hacks.

Logic and writing unlock philosophy — and, quietly, everything else. Philosophy is unreadable until you can follow an argument's skeleton, and unwritable until you can build one. But look at what else those two skills touch: every essay, every proposal, every disagreement you will ever need to win gracefully or lose honestly. This pair is where the map folds back on itself — which brings us to the threads underneath all the others.

The meta-threads

Four threads are prerequisites of everything, which is exactly why they never appear on anyone's list of fields: deep reading, clear writing, statistical thinking, and learning how to learn.

Notice, too, what this means about where you already stand. If you have spent years reading widely, writing for work, arguing carefully, following your curiosity through one field after another — you have been laying warp all along, without credit for it. Adults who "start" a new field never start from zero. They start from every meta-thread they have spun since school, which is why the second field always comes faster than the first, and the fifth faster still.

They are not threads on the loom. They are the loom. Improve any one of them and every future field gets cheaper — a compounding return no single subject can match. This is where the learning-to-learn literature earns its keep: Scott Young's Ultralearning on aggressive self-directed projects, Barbara Oakley's A Mind for Numbers and her free Learning How to Learn course on the mechanics of memory and attention. I hold some of that genre's optimism at arm's length, but its core findings — retrieval practice, spaced repetition, doing the thing instead of rereading about the thing — are as close to settled as this territory gets.

How to use the map: walk it backwards

A map is only as good as the walking method. Here is the whole of it, in four steps.

1. Name the target. Not "get smarter" — a field. Cosmology. Game theory. Persian history. The map only works when you stand somewhere on it and face somewhere else.
2. Find the target's true prerequisites. Pull up the introductory course at two or three universities — OpenCourseWare makes this a ten-minute errand — and read what they assume. Then open the field's standard textbook and notice what chapter one takes for granted. That is the real dependency list, and it is almost always shorter than the folklore version that scared you off.
3. Learn the minimum of each prerequisite — not mastery. You are laying warp, not moving in. The question is never "have I mastered calculus?" It is "can I now read the next thing without drowning?" The moment the answer is yes, move.
4. Move forward, and circle back when you hit a wall. Walls are not failures; they are information. Each one names, with perfect precision, the exact prerequisite you actually lack — which is more than any syllabus ever did for you.

Content blocks, not whole subjects

The map becomes truly useful when you stop thinking in subjects and start thinking in blocks. "Learn math before machine learning" is paralyzing folklore; the true dependency is smaller — matrix multiplication, derivatives, the chain rule, basic probability. Weeks, not years. Microeconomics does not need real analysis; it needs derivatives and the idea of optimization. Structural engineering does not need quantum mechanics; it needs statics and materials.

Whole-subject thinking makes every journey look like a decade, and a decade is exactly the price a wide mind refuses to pay. Block thinking tells you the honest price. The honest price is almost always lower.

A worked example: walking backwards from cosmology

Say the field that keeps you up at night is cosmology — where the universe came from, where it is going. Whole-subject folklore says: physics degree first, four years, enter through the front gate or not at all. Walk it backwards instead.

The standard first textbook, Barbara Ryden's Introduction to Cosmology, assumes introductory mechanics and calculus — that is the real toll, and it is far smaller than the folklore's. Mechanics, in turn, assumes single-variable calculus: derivatives and integrals, the opening weeks of any calculus course. So the true path is three blocks long. Basic calculus, from any standard text or MIT's 18.01 on OpenCourseWare. Newtonian mechanics, from 8.01 or the opening chapters of Feynman's first volume. Then Ryden — and you are reading the universe's biography from its actual first page, with the equations meaning something under your hands.

Months, not years. That is what the map is for: replacing the imagined toll with the real one, field after field, for the rest of your life.

The honest caveat

Now let me argue against my own map, because I would rather you weave than admire cartography.

You can also just start anywhere. You can fall into Byzantine history with no Greek, no grounding, no plan, and backfill the moment you hit a wall. People do this constantly, and it works — because the wall tells you exactly what to learn next, and tuition paid at the moment of need is remembered far better than tuition paid in advance. If a fascination has you right now, follow it. A live thread pulled out of order beats a dead curriculum executed perfectly. I've written elsewhere about how to ride those fascinations in finishable sprints, and about why a mind like yours collects threads this way in the first place.

This is, after all, how self-taught minds have always moved. The autodidacts of every century — the bench chemists, the amateur astronomers, the programmers who never saw a computer-science classroom — learned in exactly this stumbling, wall-finding order, and the fields they built are doing fine. The map does not replace that path. It just turns some of the walls into doorways you can see from a distance.

The map, then, is for efficiency, not gatekeeping. It exists so that when you choose to be deliberate, your hours land where they compound — and so that the wall, when you meet it, reads as a signpost instead of a verdict. Nobody checks your warp at the border. There is no border.

Two books for the wall above your desk

Book Range — David Epstein The evidence file for the wide path — why broad sampling beats early specialization in most real domains, told through athletes, scientists, and inventors. The research backbone under everything the Thread Map assumes. Read on Amazon → Book The Pattern Weaver's Edge — Oliver Mindwell The full method this essay comes from — how a connecting mind chooses its threads, orders them, and weaves a lifetime of scattered learning into a deliberate edge. Read on Amazon →

As an Amazon Associate, we earn from qualifying purchases.

Which thread first, then?

If you pressed me — no target field, one thread, starting today — I would say statistics, with writing close behind. One rewires how you read the world's claims; the other rewires how you make your own. Every other thread on the map waits patiently downstream of those two.

But that is my answer, and the map's whole point is that yours will differ. Somewhere on it is the field you keep circling — the one you've been opening tabs about for years. Stand there. Walk backwards. Count the real blocks between you and it.

I suspect you'll find the distance is shorter than you feared. It nearly always is.