The Measurable Gets Eaten.

On the credential ladder, the live edge, and what stayed scarce.

00:00 / 11:24
Caravaggio, The Calling of Saint Matthew, 1599–1600
Caravaggio, The Calling of Saint Matthew, 1599–1600. Contarelli Chapel, San Luigi dei Francesi, Rome. Public domain via Wikimedia Commons.

I. The Next Thirty Seconds

“if you value intelligence above all other human qualities, you’re gonna have a bad time.”

Ilya Sutskever, October 6, 2023

“That which speech does not reveal, by which speech itself is revealed: that alone is Brahman.”

Kena Upanishad

There is something I have started noticing in the last year and I am only now finding the language for it.

A friend, a doctor, sits with a frightened patient for a beat longer than would be polite. The silence does something the prescription cannot.

A founder I know reads a hostile board across a table and changes the subject by half a degree. The room turns.

A teacher I watched in Bengaluru takes a question she has never been asked before. Instead of answering, she opens her hands.

I notice these moments more often now because in each of them I find myself thinking the same thing.

A real, capable, contemporary model could have produced everything written or drawn or argued in the room, except the next thirty seconds.

The next thirty seconds is the live edge.

And the premium is on it.

II. The Credential Ladder

“πάντα ῥεῖ” — Everything flows.

Heraclitus

For most of the last thirty years, the world organized its ambition around the work that happens elsewhere.

The proof at the desk. The brief before the meeting. The deck polished over the weekend. The diagnosis arrived at by checklist. The memo circulated to fourteen people who never met. The model trained over a quarter. The first draft of anything.

This was the work that paid. The work that selected for the entrance exam, the engineering seat, the consulting interview, the corner office in whichever city. JEE and Gaokao and the SAT and the A-Levels selected for it. The universities ranked for it. The professions credentialed for it.

And here is the detail that stops me every time. AI did not solve Metallurgy first. It solved Computer Science. The most credentialed, highest-paid, most prestigious information profession on the planet. The engineer working with molten metal in an actual factory is still working with molten metal. The engineer at a desk writing code is the one the model replaced. (One study across more than a hundred thousand developers found coding agents lifted how much code got written by 180%, and how much actually shipped by 30%. Writing got cheap. The rest still runs through a person.)

The pattern, once you see it, is hard to unsee.

The credential was never really about knowledge. It was about access to knowledge. The gatekeeping of language, of complexity, of arcane systems that only the trained could enter. The investment banker’s authority lived in the Excel model built at 2am. The doctor’s authority rested partly on the Latin names she could attach to what she saw. The chartered accountant’s value lived inside tax codes most people could not read. The profession was the lock, and the training was the key.

For fifty years this arrangement held.

I was one of them. The math table at noon.

III. When It Became Cheap

“All that is solid melts into air.”

Marx, The Communist Manifesto (1848)

In the last eighteen months, almost all of it became cheap.

This, I think, is what Ilya was telling us, quietly, in the autumn of 2023. The thing you spent thirty years sharpening. Intelligence as legibility, intelligence as document, intelligence as deliverable. It just stopped being scarce.

The mechanism is not complicated. A thing you can measure is a thing you can train against. The credential ladder didn’t just happen to select for work that got cheap — it selected for measurable work, and measurable work is the first thing that gets eaten.

What stayed scarce is not the opposite of intelligence. Intelligence was never the bottleneck. Permission is. Accountability is. Someone still has to put their name on what the model does.

What stayed scarce is the live edge.

IV. The Real Room

“Status is something one does.”

Keith Johnstone, Impro (1979)

“Yato hasta tato drishti / Yato drishti tato manaha” Where the hand goes, the eye follows. Where the eye goes, the mind follows.

— Nandikeshvara, Abhinaya Darpana

In 1979, a theatre director in London published a book called Impro. It was about teaching adults, dulled by school, how to be present on a stage.

I have been carrying my copy for almost a decade. I am only now reading it correctly.

Its four pillars are status, spontaneity, narrative, and masks.

Status is how you read the hidden weight of a room before you act. Spontaneity is the willingness to accept what the room offers instead of blocking it. Narrative is reincorporating what the room actually gave you back into the story you are telling. Masks is the role flexibility that lets one person be the expert, the listener, the witness, and the calm authority inside the same forty-five minutes.

Johnstone was not writing for the AI lab. He was writing about what the body and the mind have to do to survive a real room.

A real room is the part that is improvised.

V. Two Kinds of Work

There are two kinds of work.

The kind that happens elsewhere, and the kind that happens in a real room with bodies, with stakes, with no possibility of going back and editing the take.

The credential ladder spent fifty years selecting for the first kind and quietly condescending toward the second. The investment banker was serious. The hotelier was in hospitality. The software engineer was technical. The bartender was just serving drinks. The doctor-as-diagnostician was a scientist. The swimming instructor was just a swimming instructor.

The model replaced the investment banker’s model. It did not replace the bartender reading a stranger’s mood at closing time. Or the pilot deciding whether to land in crosswind. Or the wedding planner holding together a ceremony when the caterer is late and the uncle is drunk and the bride is crying. Or the swimming instructor in the water with a child who is afraid. It did not replace the ER nurse at 3am, or the teacher with the question she has never been asked, or the founder who must read a hostile room and find the half-degree shift that turns it.

These people were always already in real rooms. The credential ladder just was not paying much attention.

Every profession with a live edge trains for it. Pilots use flight simulators for when an engine fails at altitude. Surgeons use cadaver labs. Firefighters drill in burn houses. Each of them has a named discipline for the moment the script breaks.

The people deploying AI evaluated the model for infinite hours and rehearsed the room for zero.

The model is not in a room.

VI. The Honest Question

“Pothi padhi padhi jag mua, pandit bhayo na koi Dhai aakhar prem ka, padhe so pandit hoy.”

(Reading book after book, the whole world died, and not one became wise. He who reads even two and a half letters of love: only he is wise.)

Kabir

Whatever field you happen to be in, there is a quiet question I have started asking myself, and I will offer it to you.

What portion of your work happens at the live edge? In a real room, with bodies, with stakes, with no possibility of going back.

Be honest.

Most of us spent twenty years getting very good at the elsewhere. We were rewarded for it. The credential ladder selected for it. We were quietly proud, in a way we did not say out loud, of how much of our work could happen without a body in a room.

The elsewhere is the part that got cheap.

The live edge is the part that stayed.

VII. Two Tables

“ಬಂತಣ್ಣ ಸಣ್ಣ ಸೋಮವಾರ, ಕಾಣಬೇಕಣ್ಣ ಸೋಮೇಶ್ವರ.”

(The festival of Sanna Somwara has come, and we must now see Someshwara.)

Da Ra Bendre

I sat at two tables at lunch in eighth grade.

The math table at noon. The auditorium at four.

I had been doing theatre since I was seven, and math for longer than I can remember. For most of the last twenty years I assumed the math table was the one with the answer. I had a few quiet theories about what the theatre was for. I was wrong about most of them.

The math is done by the machine now.

The live edge is mine, and yours, and the bartender’s, and the pilot’s, and the nurse’s, and the teacher’s, and the wedding planner’s, and the swimming instructor’s, and the founder’s, and anyone willing to walk into a real room with a real body and stand there when the script breaks.

The premium is on the live edge.

Show up.


Further Reading

The books that go deepest on what it actually takes to survive a real room.

Impro: Improvisation and the Theatre — Keith Johnstone (1979). The foundational text. The chapter on status alone rewires how you read every meeting.

Impro for Storytellers — Keith Johnstone (1999). Johnstone on narrative and recovery. What to do when a scene collapses.

Truth in Comedy — Halpern, Close, Johnson (1994). Where “yes, and” was codified as a discipline. More technical than Johnstone. The mechanics of building on what the room gives you.

Never Split the Difference — Chris Voss (2016). Tactical empathy and labeling under pressure. The same principles, from the FBI hostage negotiation side.

Pitch Anything — Oren Klaff (2011). Status and frame control for high-stakes rooms. Required reading on how status works in the first sixty seconds.

← all writing