Recoding America

officialSite="https://www.recodingamerica.us/" />

Quick Assessment

Overview

Recoding America: Why Government Is Failing in the Digital Age and How We Can Do Better is a 2023 book by Jennifer Pahlka that examines why government technology projects frequently fail and what these failures reveal about institutional governance. Despite its title focusing on digital systems, Pahlka characterizes the work as fundamentally a cultural critique of how government operates rather than a technology book.¹

The book's central argument is that the separation between policy and implementation is "both false and self-defeating," rooted in what Pahlka describes as an elitist view that implementation is beneath the people who make policy.¹ She contends that government is trapped in "waterfall" mode — rigid, compliance-driven, and process-oriented — rather than operating in an agile, iterative, and outcome-oriented manner. According to this analysis, failing technology projects are symptoms of deeper structural problems including burdensome oversight, bureaucratic risk-aversion, insufficient technical capacity, and systems designed for internal stakeholders rather than end users.¹

The book has broader implications beyond government IT, connecting to debates about state capacity and, more recently, AI governance. Pahlka has explicitly framed the work as advocating for state capacity, stating in one review: "nowhere in the book do I use the words state capacity, but it is absolutely a book that is advocating for state capacity."² Named one of NPR's Best Books of 2023 and included in Ezra Klein's list of books that explain the current moment, the work has influenced policy discussions about government capability in the digital era.¹

Author Background

Jennifer Pahlka founded Code for America in 2010, an organization aimed at bringing modern technology practices to government services. She served as U.S. Deputy Chief Technology Officer under President Obama and helped establish the U.S. Digital Service (USDS), a government agency focused on improving federal digital services. Her government experience also includes serving on the Defense Innovation Board.²

During the COVID-19 pandemic, Pahlka co-founded U.S. Digital Response, an organization mobilizing technology volunteers to help government agencies respond to the crisis. She currently works as a Senior Fellow at the Federation of American Scientists, where her focus has shifted to AI in government.² One reviewer compared the book's potential impact to Rachel Carson's Silent Spring for the state capacity movement, suggesting it could serve as a foundational text for advocates of improved government capability.²

The Cascade of Rigidity

A central framework in Recoding America is what Pahlka terms "the cascade of rigidity" — the way multiple layers of oversight, compliance requirements, and risk-aversion compound to make government technology projects enormously expensive and slow.³ According to her analysis, this rigidity is not fundamentally about technology but about organizational culture, incentive structures, and institutional beliefs about how policy should be separated from implementation.³

The cascade manifests in several ways: procurement rules that prioritize compliance over outcomes, contracting structures that separate those who write specifications from those who build systems, and accountability mechanisms that punish visible failures while ignoring the ongoing failure of systems that don't serve user needs. Pahlka argues these dynamics create an environment where government employees are incentivized to follow process rather than achieve results, and where bringing in technical expertise to directly address problems is treated as bypassing proper procedures.¹

This framework has been applied beyond traditional IT to analyze how government might handle emerging technologies. In subsequent writing, Pahlka has explored how "AI meets the cascade of rigidity," examining how the same institutional dynamics that hamper government technology projects will likely impede both effective AI adoption and meaningful AI oversight unless actively reformed.⁴

Case Studies and Evidence

The book documents several high-profile government technology failures and rescues. The Healthcare.gov crisis in 2013 serves as a pivotal example: the initial launch of the Affordable Care Act's insurance marketplace was a catastrophic failure, with the website unable to handle basic user traffic or enrollment functions. However, the subsequent "tech surge" — which brought in skilled technologists empowered to work iteratively and focus on user needs — demonstrated that government systems can work when properly resourced and organized.³

This rescue effort spawned two lasting institutions: the U.S. Digital Service within the federal government and Nava PBC, a public benefit corporation that continues to work on government technology projects. The Healthcare.gov case became a template for how to approach government digital services differently, prioritizing user research, agile development, and empowering technical experts to make implementation decisions.³

The book also examines other failures and successes across government agencies, using these cases to illustrate broader patterns about how institutional culture affects technical outcomes. Through these examples, Pahlka argues that the problem is rarely a lack of good intentions or even a lack of funding, but rather the way organizational structures and incentives prevent good intentions and adequate funding from translating into functional systems.¹

Connection to AI Governance

In January 2024, roughly a year after the book's publication, Pahlka testified before the U.S. Senate specifically connecting her framework to AI governance. She argued that the success or failure of AI in government "comes down to how much capacity and competency we have to deploy these technologies thoughtfully."⁵ Her testimony warned against applying red tape to benign AI uses like handwriting recognition while lacking the capacity to evaluate genuinely risky applications, and called for "enablement" over "mandates and controls."⁵

Pahlka has framed AI as presenting both an opportunity and a threat in the context of state capacity: "The internet era coincided with a decline in state capacity. The AI era must see a reversal of this trend."⁶ This perspective suggests that the same institutional problems that caused government technology failures in the past two decades will be even more consequential as AI systems become more powerful and widespread.⁶

The Niskanen Center published Pahlka's subsequent analysis "AI meets the cascade of rigidity," which examined how existing government structures will likely struggle with both deploying useful AI tools and providing effective oversight of potentially risky AI systems.⁴ This work has positioned Recoding America not just as a critique of past digital failures but as a warning about future governance challenges.

Relevance to AI Safety

The book's thesis has direct implications for AI safety and governance. If governments cannot build and maintain functional IT systems, they face substantially greater challenges in overseeing frontier AI development and deployment. The capacity gap between AI developers and potential AI regulators continues to widen, raising questions about whether government institutions can effectively evaluate and regulate increasingly capable AI systems.³

Holden KarnofskyPersonHolden KarnofskyHolden Karnofsky directed \$300M+ in AI safety funding through Coefficient Giving (formerly Open Philanthropy), growing the field from ~20 to 400+ FTE researchers and developing influential framewo...Quality: 40/100, co-founder of Coefficient GivingOrganizationCoefficient GivingCoefficient Giving (formerly Open Philanthropy) has directed \$4B+ in grants since 2014, including \$336M to AI safety (~60% of external funding). The organization spent ~\$50M on AI safety in 2024...Quality: 55/100, has argued that "there's no way to get to an actual low-level risk from AI without government policy playing an important role."⁷ Similarly, Ezra Klein has suggested that government needs approximately 300 excellent AI experts from multiple domains serving in advisory, regulatory, and auditing roles to adequately address AI governance challenges.⁸ Recoding America provides a diagnosis of why this kind of technical capacity is currently lacking in government and offers a framework for understanding the institutional barriers to building it.⁷⁸

The book suggests that effective AI governance will require not just new regulations or oversight bodies, but fundamental changes to how government develops technical capacity, how it separates (or fails to separate) policy from implementation, and how it structures accountability for technical systems. Without addressing these deeper institutional issues, even well-intentioned AI policy may fail in implementation.¹

Reception and Broader Impact

Recoding America received widespread praise for bringing arguments about state capacity to a mainstream audience. Beyond being named one of NPR's Best Books of 2023, it was highlighted by Ezra Klein as one of his books that explain the contemporary moment.¹ The book has been influential in policy circles, contributing to Pahlka's subsequent Senate testimony on AI in government and broader discussions about digital government reform.⁵

Following the book's publication, the Recoding America Fund was launched to support the mission of reforming government for the digital age.⁹ The broader civic technology movement that the book documents — including Code for America, USDS, 18F, and Nava — represents a practical effort to build government capacity from within existing institutions rather than through purely external advocacy.⁹

The book has been positioned within ongoing debates about state capacity, with Pahlka's work contributing to discussions about whether and how government capability has declined over recent decades and what might be done to reverse those trends. Her emphasis on implementation over policy design has resonated with practitioners and reformers who see abstract policy discussions as disconnected from the concrete challenges of making government programs work.²

Key Uncertainties

Several questions remain about the book's framework and its applicability to AI governance:

Generalizability of Solutions: While the book documents successful interventions like the Healthcare.gov rescue and the founding of USDS, it remains unclear how widely these approaches can be replicated. The cases highlighted often involved crisis situations that created political space for unconventional approaches — it's uncertain whether similar reforms can occur without crisis conditions.

Scope of Cultural Change Required: Pahlka's diagnosis suggests deep institutional and cultural problems in how government operates. The extent to which these can be reformed through the kinds of interventions she advocates, versus requiring more fundamental restructuring, remains debated. Some critics might argue that the "waterfall" culture and the separation of policy from implementation are features rather than bugs of democratic governance structures designed for accountability.

AI-Specific Challenges: While the book provides a useful framework for understanding government technology failures generally, AI systems present novel challenges including interpretability difficulties, unexpected capability emergence, and potential existential risks. Whether the capacity-building approach Pahlka advocates is sufficient for these distinctive challenges, or whether AI governance requires different institutional innovations, remains an open question.

Trade-offs with Oversight: The book's emphasis on empowering technical experts and reducing bureaucratic oversight could create tensions with legitimate accountability mechanisms. How to balance the need for agile, expert-driven implementation with democratic oversight and accountability in the context of powerful AI systems is not fully resolved by the framework presented.

Sources