Over a decade ago, a mysterious package arrived at my door. Inside: a track jacket and a nearly indecipherable book on improv. I didn’t know it at the time, but I had just been inducted into the Palantir worldview. 

There was no guide, no handbook—only deep-end immersion (or, as Shyam would say, “gamma rays”). The ethos was recognizable—bias to action, deep engagement with users, and a commitment to “Save the Shire” at all costs—but never explicitly articulated. 

Despite shaping a generation of technologists, Palantir never codified its philosophy. Now, Alex Karp and Nicholas Zamiska have done it for them. 

At its core, The Technological Republic is a call for a more patriotic technology sector—one that stands for something beyond the marginal satisfaction of needs. Companies, Karp and Zamiska argue, should realize their existence is contingent on the West’s dominance, and should apply their intellectual firepower towards supporting its liberal-democratic governing substrate.  

This book arrives at an inflection point: Palantir alumni hold key government positions, China’s AI ambitions grow bolder, and tech’s counterculture has gained real power in Washington. Karp’s argument is more than a philosophy—it’s a battle plan. 

The book is, as Tyler Cowen would say, self-recommending. Their case is lucid, their writing crisp, their vision compelling. They are erudite without being academic, opinionated without being bombastic, novel without being glitzy. It is a read that will stand for decades. 

But good arguments can still have blind spots. And this one has at least three: the danger of entanglement, the seduction of utopianism, and the risks of unchecked dynamism. 

I. The Danger of Entanglement 

Karp and Zamiska want a Manhattan Project for AI. The first one won a war—but also birthed the military-industrial complex. 

When tech and government get too close, that might well drive innovation. They create a ruling class. Call it the “power elite” (Mills) or a “caste” (Baltzell)—the result is the same: governance by access, not law. 

Karp and Zamiska assume the right partnerships will yield the right outcomes. But institutions run on rules, not trust. The Founders understood this: they designed a government that pits ambition against ambition, ensuring stability through process, not faith in good actors. 

That lesson holds today. The answer isn’t deeper entanglement but real competition. Arnold Kling is right: we need independent oversight—COOs, auditors, and strict interoperability rules so the best software wins. Light rules, not relationships, should determine who shapes public technology. 

Yet entanglement isn’t the only risk. The greater danger is conviction unmoored from realism. 

II. The Seduction of Utopianism 

Karp and Zamiska correctly critique “technological agnostics”: technologists whose only firm commitment is a bland, superficially unobjectionable thin utilitarianism. But they underestimate the greater, opposite threat: utopians with too much conviction. 

Utopians exist across the ideological spectrum. There are left-wing utopians, obsessed with engineering society from above. There are crypto-utopians, convinced that a few well-placed algorithms will render institutions obsolete. There are techno-utopians, who see innovation as an unqualified good, as if the only thing standing between humanity and perfection is another software update. 

The Founders knew better. They weren’t French or Russian revolutionaries; they didn’t believe in remaking the world. They channeled ambition instead of erasing it. 

Sowell saw history as a battle between constrained and unconstrained visions of human nature. Karp and Zamiska lean in the right direction but don’t go far enough. A technological republic needs the same realism that guided the first republic: institutions designed for humans as they are, not as we wish them to be. 

III. The Risks of Unchecked Dynamism 

“Dynamism” sounds good. So do “agility” and “disruption.” But technology shouldn’t just move fast—it should move well. 

Karp and Zamiska assume dynamism is inherently good. But the Founders knew some things should not move fast. That’s why they built vetogates—deliberate obstacles to slow bad ideas. The Founders were optimizing for maximin, not maximax.  

Perhaps one distinction is between legislative and executive.  Legislative decisions—the “what”—need stability, predictability, and deliberation. Executive decisions—the “how”—should be efficient, effective, and dynamic. 

Technology should improve execution, not dictate governance. However sophisticated, software cannot replace law. The rule of law depends (among other factors) on predictability, and predictability depends on limits. 

IV. Values Without Compromise 

Two decades ago, Palantir made a counterintuitive argument: privacy and security weren’t opposites. It was controversial then—but correct. 

Today, I argue the same for the rule of law and efficiency: they are not opposites but complements. That’s the core principle guiding the AI we build for judges at Learned Hand. Our challenge mirrors Palantir’s in 2004: creating technology that respects the system it serves. 

I still proudly wear my Palantir track jacket. My time there shaped me in ways I’m still uncovering. Now, I finally have a book that explains why. I don’t buy every argument in it—but I buy the premise: technology should stand for something beyond itself. 

In an age that prefers technology without principle, that argument matters.