Learnable patterns

People who want the reasoning substrate that shows up under engineering, quantitative finance, and hard-technology founding.

1. 1Single-variable then multivariable calculus, to the point of comfort rather than exam-passing.
2. 2Linear algebra — the actual working language of modern computing and statistics.
3. 3Probability and statistics, with emphasis on distributions and estimation rather than significance testing.
4. 4Mechanics and thermodynamics: the source of the habit of solving for constraints and orders of magnitude.
5. 5Practise estimating quantities you cannot look up until your answers land within an order of magnitude.

Closest routes in the dataset: Elon Musk · Jensen Huang · Bernard Arnault

The most common route in this dataset, and the one with the lowest capital requirement to start.

1. 1Programming fundamentals to the point of building something end-to-end alone.
2. 2Data structures and algorithms — enough to reason about cost, not to pass interviews.
3. 3Systems: how memory, networks, and databases actually behave under load.
4. 4Ship something real that other people use, and operate it. Operating is where the learning is.
5. 5Distributed systems or machine learning, chosen by where your industry is heading.

Closest routes in the dataset: Larry Page · Sergey Brin · Mark Zuckerberg · Jeff Bezos

People drawn to allocating capital rather than operating a product company.

1. 1Accounting first — read a balance sheet, income statement, and cash-flow statement without help.
2. 2Valuation: discounted cash flow, and why its inputs are usually the argument.
3. 3Read primary documents. Annual reports and filings, not commentary about them.
4. 4Write your investment thesis down before you act, and review it against what happened. This is the entire discipline.
5. 5Study position sizing and drawdowns before returns. Surviving is the prerequisite for compounding.

Closest routes in the dataset: Warren Buffett · Alice Walton · Jim Walton

People who build the machine that delivers the thing, rather than the thing itself.

1. 1Learn to talk to customers without leading them, and to separate what they say from what they do.
2. 2Instrument something and read the data — funnels, cohorts, retention.
3. 3Study supply chain and inventory as a cash-conversion problem, not a logistics one.
4. 4Own a process end-to-end and reduce its cycle time by half. Nothing teaches operations like a deadline you own.
5. 5Learn how pricing actually works; it is the highest-leverage operational decision most companies never revisit.

Closest routes in the dataset: Michael Dell · Jeff Bezos · Amancio Ortega

People building where perceived value, not technical capability, is the moat.

1. 1Study how scarcity and price signal quality, and where that mechanism breaks.
2. 2Learn distribution and retail control — in this dataset, brand fortunes are built by owning where the product is sold.
3. 3Understand creative direction as a management problem: hiring, backing, and replacing taste.
4. 4Study a house that was revived and one that was diluted. The difference is almost always distribution discipline.
5. 5Learn the operating mechanics — lead times, sell-through, markdown — that determine whether the brand story survives contact with inventory.

Closest routes in the dataset: Bernard Arnault · Amancio Ortega