Why Palantir Makes New Hires Read a Theater Improvisation Book

Keith Johnstone’s Impro is on Palantir’s onboarding reading list. It has nothing to do with software, national security, or AI. That’s the point.

The book is about improvisational theater. Its author, Keith Johnstone, spent decades teaching actors how to stop blocking and start creating. The lessons he extracted along the way — about status, spontaneity, and what education does to raw talent — are among the most precise diagnostics of organizational behavior I’ve read. $PLTR clearly agrees.

Reading it through the lens of company building, the message is uncomfortable: most of what makes a “good candidate” on paper is exactly what makes someone difficult to work with in practice.

The talent problem isn’t what you think

Johnstone’s most counterintuitive observation is that talent is not normally distributed the way intelligence is. “Intelligence is proportional to population, but talent appears not to be related to population numbers.” You can’t recruit for it at scale using standard proxies.

Worse, formal education actively suppresses it. “The dullness was not an inevitable consequence of age, but of education.” The students most damaged by this process, the ones hardest to unlock, are not the failures. They are “the star performers at bad high schools. Instead of learning how to be warm and spontaneous and giving, they’ve become armoured and superficial, calculating and self-obsessed.”

This is a direct challenge to every credential-driven recruiting process in Silicon Valley. The polished candidate who knows how to perform competence is often the one least capable of actual creative output. The armor is the problem. “To actually create something means going against one’s education.”

What Palantir actually screens for is whether someone has preserved the imaginative responsiveness that education systematically trains out. “I began to think of children not as immature adults, but of adults as atrophied children.”

Yes-sayers vs. blockers

The most operationally useful framework in the book is the binary between acceptors and blockers.

“Those who say ‘Yes’ are rewarded by the adventures they have, and those who say ‘No’ are rewarded by the safety they attain.”

In improvisation, blockers kill scenes. In organizations, they kill momentum. Johnstone’s observation is that blocking is a skill. “Bad improvisers block action, often with a high degree of skill. Good improvisers develop action.” High-status players block because they’re afraid of being humiliated. They’d rather kill the idea than be seen to fail with it.

This maps cleanly onto organizational dysfunction. The person who says “that won’t work” three seconds after you propose something is not being rigorous, they are most likely just protecting their status. The person who says “yes, and—” and develops the action is the one you want in the room.

Palantir’s deployment model — putting engineers directly inside customer organizations, often in chaotic, ambiguous situations — selects hard for acceptors. You cannot be a blocker in that environment. There’s no script for FDEs.

Status is the operating system

The chapter on status dynamics is the sharpest management reading I’ve encountered outside of explicit leadership literature. The core insight: “Things said are not as important as the status played.”

Every interaction in an organization is a status transaction. Most people are running these transactions unconsciously and inconsistently. A manager who can read them — and who understands that high-status security manifests as relaxation, not aggression — has a meaningful edge.

“When the highest-status person feels most secure he will be the most relaxed person.” The corollary is that leaders who micro-manage, dominate airtime, or punish failure publicly are broadcasting insecurity, not authority. The status play undermines itself.

The book’s “ten golden rules for people who are Number 1” are worth reading twice. Rule six stands out: “You must reward your immediate subordinates by permitting them to enjoy the benefits of their ranks.” Retention is a status problem before it’s a compensation problem.

What good leadership looks like

Johnstone’s model of the teacher (which translates directly to the manager) is radically non-hierarchical in one specific way: he accepts blame for failure as a tool for liberating performance. “I play low status, and I will explain that if the students fail they are to blame me.” By removing the threat of personal failure from the learner, you remove the defense mechanism that suppresses their best output.

The parallel in team management is creating an environment where people’s first instinct is execution rather than self-protection. That requires the leader to absorb downside risk publicly — to be the one who falls on the sword when an experiment doesn’t work. Most leaders do the opposite.

The goal, as Johnstone frames it, is helping people “abandon control while at the same time they exercise control.” That is the highest-leverage thing a manager can do: build enough safety that the team stops blocking themselves.

The real reason it’s on the reading list

Palantir is not a normal software company and it has never pretended to be. The things it needs from its people — comfort with ambiguity, the ability to walk into a hostile room and read what’s actually happening, the willingness to develop action in real time without a playbook — are not teachable via technical onboarding.

Impro is on the list because it names the underlying capabilities more precisely than any management book does, and because it forces a reckoning with a hard truth: the selection and development of genuinely creative, high-agency people requires dismantling the exact habits that made them look attractive on paper in the first place.

“The real avant-garde are not imitating what other people are doing, or what they did forty years ago, they’re solving the problems that need solving.”

That’s the job description. The book explains why it’s hard to hire for it.