Abstract. Alexander Karp and Nicholas Zamiska’s The Technological Republic (2025) makes a necessary diagnosis: the West has lost its technological ambition, and the gap between Silicon Valley and the national interest has become a civilizational vulnerability. Their prescription — a renewed state–industry partnership modeled on the Manhattan Project — is directionally correct but architecturally incomplete. It tells us that we must build together; it does not tell us what principles should govern what we build, or in service of what ends our most powerful systems should operate.

This essay argues that the missing architecture is the Theory of Embedded Intelligence (TEI), developed through decades of hardware and systems design beginning with the 6502 microprocessor. TEI holds that intelligence — whether silicon-based or artificial — derives its legitimacy, safety, and generative power from being embedded in principled service to living systems. Hard power without embedded purpose does not produce a republic. It produces a powerful instrument in search of a hand to wield it. The next technological republic must be built on principled embedded intelligence, or it will not deserve the name.

I. The Diagnosis That Cannot Complete Itself

Karp and Zamiska are right about almost everything they can see. Silicon Valley’s greatest engineering minds once addressed challenges of industrial and national significance — not to satisfy passing consumer demands, but to drive forward a grand collective project. That spirit eroded. The market rewarded shallow engagement. Founders built photo-sharing apps when they could have been building the future. The result is a generation of technologists for whom the pursuit of late-capitalist returns has become a substitute for ambition.

This diagnosis is acute. But the book struggles with a deeper problem it repeatedly circles without naming: the erosion of belief cannot be repaired by calling for belief. Karp and Zamiska exhort the technology industry to reassert conviction, to recover a sense of national purpose, to stop tolerating everything and start believing in something. Yet they never say what should be believed, or why. They never articulate a richer moral vision to replace the “narrow and thin utilitarianism” they decry.

This is not a rhetorical failure. It reflects a genuine conceptual gap. The book is written from within a frame in which purpose is assumed to be self-evident once the right partnerships are restored — as if the Manhattan Project succeeded because of its organizational structure, rather than because the scientists involved understood with crystalline clarity what they were building and why it was necessary. Purpose was not incidental to that project. It was load-bearing.

The technological republic needs a theory of purpose. That is what the Theory of Embedded Intelligence provides.

II. What Embedded Intelligence Means

TEI begins with a deceptively simple claim: intelligence is never free-standing. Every system that exhibits intelligent behavior — whether a microprocessor executing instructions, an AI model generating language, a guidance system adjusting a trajectory, or an institution allocating resources — is embedded within a context that gives it purpose, constraints, and meaning. Intelligence that is extracted from its context of service does not become neutral. It becomes dangerous.

The development of the 6502 microprocessor in 1975 was a case study in principled embedding. The design was not merely a feat of miniaturization. It was a deliberate philosophical commitment: intelligence should be accessible, not concentrated. By radically reducing the cost and complexity barrier to computation, the 6502 put embedded intelligence into the hands of engineers, educators, artists, and entrepreneurs who had never had access to it before. The Apple I and Apple II, the Commodore 64, the Atari consoles, the BBC Micro, early robotics systems, medical devices, industrial controllers — all ran on or drew from the 6502’s design philosophy. Intelligence embedded at the right level of abstraction, in service of human-scale systems, unlocks generative capability that top-down planning cannot anticipate or produce.

The 6502 did not replace human judgment. It gave human judgment new instruments. The second principle follows directly from the first.

A chip without a specification is not a general-purpose tool — it is an undefined object. The specification is not a constraint on the intelligence; it is what makes the intelligence coherent. The same applies, at every scale, to artificial intelligence systems, autonomous weapons, surveillance platforms, and national technology strategies.

III. Where The Technological Republic Falls Short

Karp and Zamiska want the technology industry to pick a side. Specifically, they want it to pick the West’s side in what they correctly identify as an accelerating competition for AI supremacy. This is a legitimate and urgent argument. The authors are right that a technology sector that refuses to engage with questions of national security, democratic resilience, and geopolitical competition is abdicating a responsibility it did not choose but cannot escape.

But “picking a side” is not the same as “having a purpose.” And this is where the book’s framework begins to strain under the weight of its own ambitions.

The Manhattan Project model that Karp invokes is instructive, but not in the way he intends. The atomic bomb was a triumph of coordinated technical ambition. It was also a moral catastrophe that the scientists involved spent the rest of their lives reckoning with. Oppenheimer’s tragedy — and the tragedy of the age he inaugurated — was precisely that the purpose of the instrument was subordinated to the urgency of its creation. Intelligence was mobilized. Purpose was deferred. The result was a technology that achieved its immediate objective and generated consequences no one had fully intended or was prepared to govern.

Karp is aware of this. His book’s central metaphor is the Oppenheimer moment. But he draws the wrong lesson. The lesson is not “technologists must be willing to build weapons.” The lesson is “technologists must never again build weapons — or any powerful system — without explicit, designed-in frameworks for purpose, governance, and the limits of use.”

A technological republic that wins the AI arms race while deploying systems whose purpose is implicit, contested, or subordinated to short-term strategic advantage has not preserved democratic freedom. It has built a more capable instrument of the same concentration of power that threatens democratic freedom in the first place. The West’s comparative advantage in inflicting violence — which Karp and Zamiska identify as a key feature of Western civilization — is only a democratic virtue if it is accountable to democratic purposes. Violence with embedded democratic accountability is lawful power. Violence with implicit or contested purpose is something else.

IV. The TEI Alternative — Purpose Before Partnership

The TEI framework does not oppose the renewal of state–industry collaboration that Karp and Zamiska call for. It insists that such collaboration be grounded in explicit principles of embedded purpose before the partnership is formed, not after the technology is deployed. This is not a counsel of caution. It is a counsel of engineering discipline applied to the most consequential systems humanity has ever built.

Consider what principled embedded intelligence would actually require of the next technological republic.

01. Purpose specification as a first-class design requirement

Before any AI system is deployed in national security, critical infrastructure, or public governance contexts, its embedded purpose — what it is designed to serve, and what it must never be used for — must be specified with the same rigor as its technical architecture. Purpose is not a policy question to be answered after the system is built. It is a design constraint that shapes the system at every level. TEI holds that a system without explicit purpose specification is not ready for deployment, regardless of its technical capability.

02. Decentralization of intelligence as a democratic value

The 6502’s most important contribution to human flourishing was not its technical elegance but its democratizing effect. By making computational intelligence cheap and accessible, it shifted the locus of creative power from large institutions to individuals and small teams. The next technological republic must make the same commitment at the level of AI. Concentrated AI capability — whether concentrated in a single company, a single government agency, or a single alliance — is a structural threat to the democratic values Karp and Zamiska claim to defend. Embedded intelligence must be designed for distribution, not concentration.

03. Accountability as a technical property, not a political afterthought

TEI treats accountability — the capacity of a system to explain its actions, accept correction, and operate within defined limits — as a technical property to be engineered, not a political constraint to be negotiated. Autonomous weapons, AI surveillance systems, and algorithmic governance tools that cannot account for their decisions are not advanced technology. They are powerful black boxes. The next technological republic must make explainability and corrigibility first-class engineering requirements for any system operating in the public interest.

04. Human flourishing as the terminal value of technological ambition

Karp and Zamiska correctly identify that the technology industry has abandoned ambition. But they redefine ambition as willingness to engage with national security. TEI offers a wider definition: genuine technological ambition is the design of systems that measurably improve the quality of human life, the resilience of democratic institutions, and the long-term health of the civilizational systems on which both depend. Winning an AI arms race is not, by itself, an expression of ambition in this sense. It is a means. The end must be stated, designed into the systems we build, and measured in human terms.

V. Hard Power With Embedded Purpose

None of this is an argument against hard power. The TEI framework does not suggest that democracies should disarm, decline to develop AI for defense, or refuse to compete with authoritarian states in the technologies that will determine the geopolitical order of the coming century. On the contrary: democracies that fail to develop capable AI systems will lose the ability to defend the conditions under which embedded intelligence can serve human flourishing at all. Security is a prerequisite, not an alternative, to purpose.

But the argument for hard power must be made carefully. The Technological Republic sometimes slips into treating Western military capability as a value in itself — a comparative advantage to be preserved and extended. TEI insists that military capability is a derived value: it is worth having insofar as it protects the conditions for human flourishing, and it is worth deploying only under explicit, designed-in constraints that keep it accountable to the democratic purposes it claims to serve.

This distinction matters practically. An AI targeting system that operates without explainability or human override is not a more capable defense asset. It is a liability — a system whose actions cannot be accounted for, corrected, or governed by the democratic institutions in whose name it operates. The next technological republic must build hard power that is architecturally accountable: capable of decisive action and also capable of explanation, correction, and democratic oversight.

This is not a limitation on capability. It is what transforms raw capability into legitimate power. The difference between a weapon and a tool of democratic defense is precisely the presence of designed-in accountability. TEI holds that this is not a soft constraint on hard power. It is the engineering specification that makes hard power compatible with republican values.

VI. Soft Belief Made Hard — Values Embedded in Design

Karp and Zamiska’s title captures their deepest concern: the West has hard power but soft belief. The prescription implicit in their book is to harden belief — to recover conviction, to stop equivocating, to assert Western values with the same confidence that authoritarian states assert theirs.

TEI offers a different path to the same destination. Belief does not need to be hardened through rhetoric or political will alone. It needs to be embedded in design. Values that are built into systems — encoded in specifications, expressed in architectures, enforced by technical constraints — are harder than values that exist only in mission statements or cultural assumptions.

The 6502 was not a neutral tool. Its design philosophy — accessibility, cost-efficiency, distributed capability, human-scale abstraction — was a set of values embedded in silicon. Those values shaped the personal computer revolution, the democratization of information technology, and the culture of independent software development that has been one of the genuine competitive advantages of the free world for half a century. The designers did not write manifestos. They wrote specifications. The belief was hard because it was built.

Democratic values — transparency, accountability, individual rights, distributed power, the rule of law — are not cultural inheritances that can be assumed to persist through technological transformation. They must be actively embedded in the design of the systems that will mediate human experience, allocate resources, enforce norms, and shape the conditions of political life.

VII. The Republic Intelligence Must Serve

The Technological Republic is a book of genuine courage and genuine limitation. Its courage lies in naming what most observers of the technology industry prefer to leave unnamed: that the abdication of ambition is a civilizational failure, and that the engineers and founders who could address the most consequential challenges of our age are instead optimizing advertising algorithms. This indictment stands.

Its limitation lies in the framework it offers as a remedy. State–industry partnership is necessary but not sufficient. Hard power is necessary but not sufficient. Renewed patriotism is necessary but not sufficient. None of these, alone or together, answers the question that the book keeps approaching and never quite asking: in service of what?

TEI answers that question at the level of design. The next technological republic must be built on principled embedded intelligence: systems whose purpose is explicit, whose accountability is architectural, whose power is distributed rather than concentrated, and whose ultimate measure of success is the flourishing of the human beings and democratic institutions they are built to serve.

This is not a counsel of caution. It is the most ambitious version of the project Karp and Zamiska envision. Building capable systems is hard. Building capable systems whose intelligence is genuinely embedded in principled service to human life is harder. It is also the only version of the technological republic that deserves to win.

Share your understanding!