VC and PE is a data problem.

At my first venture firm, the knowledge lived in people. A read on a founder traveled down the hallway; a thesis hardened over months; the rules of thumb passed like recipes. I came to love this world — and to believe that what looked like art was pattern recognition in disguise, the oldest cognitive act there is. The founder, the team, the shape of the early growth, the timing of the market, all weighed fast and below the level of words against thousands of companies seen before. The way the craft is taught gives it away: you learn venture by apprenticeship, years at someone's elbow, because there is no textbook — you load a library of cases by exposure until your eye is trained.

Crafts like this have a history, and it does not end in craft.

For most of history there was no biology, only the trained body — Shennong tasting a hundred herbs, poisoning himself again and again to learn which healed and which killed, knowledge bought one trial at a time. For centuries the instrument was a palate, an eye. Then, in our lifetime, a different one arrived: sequencing costs fell from ninety-five million dollars a genome to a few hundred, and AlphaFold cracked a fifty-year problem so fast that a structure one biologist had chased for a decade fell out of the machine in half an hour. The eye learned to see; the instrument learned to compute a pattern no eye could hold.

Not every craft makes this passage. A craft becomes a science when three things hold at once: the outcome is objective, the features you can observe carry real signal, and an instrument arrives that reads the signal at a scale no brain can. Wine tasting fails the first — "better" is only a panel of palates, and blind, the experts scatter. Roulette fails the second — the wheel is unreadable, though the casino lives on people who feel a pattern in it. Biology cleared both and waited centuries for the third.

Finance has made this passage once already. In 1980, picking stocks was learned at the elbow, reading annual reports until your eye was trained. Then the tape turned legible — disclosure law, accounting standards, the terminal on every desk by 1982 — and within a generation the quants arrived; today algorithms run roughly a third of all trading. The same act, pricing the asset and picking the winner, became something a machine could do at scale the moment the room turned legible.

The private market is the one room in that house still lit by the trained eye, and it is the biology kind of craft. The read can be taught — it travels down the hallway and out the door to found the next great firm — so it rests on a real, learnable pattern. And its outcome is objective in the way wine's never is: a deal returns or it doesn't, settled in cash. Venture writes its own answer key. The pattern is real and the outcome is real; only the third condition was ever missing, and for sixty years the only instrument was the human eye — a marvel that does not scale.

The pioneers have been building the better instrument for a decade — SignalFire's Beacon tracking eighty million companies, EQT's Motherbrain, Coatue's Mosaic. It has been slow because venture is the hardest data environment in finance: the data is sparse, a few thousand deals a year against millions of ticks; selection-biased, since you never learn what became of the deals you passed; non-stationary, the rules mutating each era, so that metrics taken as gospel a decade ago are already being rewritten; and glacial, a seed bet taking a decade to grade. The public market was the easy instance. Venture is the hard one, which is why it is last.

Two things changed together. The data is turning legible — platforms now read the hiring, the shipping, the usage, the exhaust that moves months ahead of revenue. And the instrument stopped needing the pattern named before it could learn it: where every earlier system needed a human to christen "Rule of 40" before it could track it, the new one reads the shape as it is. Non-stationarity, the very thing that made venture harder than biology, becomes the place the machine beats the brain — the brain needed the name, the machine does not.

So let me say plainly what I believe. Venture and private equity are a pattern-recognition problem, the last great craft in finance still waiting for its instrument — and the instrument has arrived. The microscope made naturalists into the first biologists; the trained eye survives the lens and inherits it. The people who spent careers learning to read founders are about to be handed one, and the nerve to commit before the evidence is in stays theirs to supply. Sixty years built this industry by hand. The thing that lets them see what their hands always knew is here, and I can't think of a more exciting place in finance to be standing.