Noah Smith delivers a devastating critique of a book that many assumed would define the decade's economic conversation. While Daron Acemoglu and Simon Johnson's Power and Progress was poised to become the standard text for AI anxiety, Smith argues it fails on its own terms, relying on historical distortions rather than rigorous evidence. For busy leaders trying to navigate the current tech boom, this review offers a necessary reality check on the prevailing narrative that innovation can simply be "steered" away from worker displacement.
The Pedigree vs. The Payoff
Smith begins by acknowledging the formidable reputation of the authors. He notes that Acemoglu is the architect of the "institutional explanation for national development," while Johnson is a leading voice on financial regulation. Together, they attempted to synthesize two major societal fears: the distrust of big tech and the panic over artificial intelligence. Smith writes, "Power and Progress weaves those two anxieties into a more-or-less coherent whole — a sum of all technological fears, if you will." Yet, despite this perfect storm of timing and celebrity, Smith observes that the book has made surprisingly little impact, noting, "in the 9 months since it came out, I'm not sure I've once heard someone reference the book or any idea in it."
The core of Smith's frustration lies in the book's central thesis: that society can and should "steer" innovation toward technologies that augment humans rather than replace them. Smith finds this argument compelling in theory but hollow in execution. He argues the authors assume their conclusions rather than proving them, stating, "The historical examples it cites simply don't support a narrative of out-of-touch technologists inventing the wrong sorts of technologies and hurting workers in the process." This disconnect between the book's ambition and its evidence is where the analysis begins to unravel.
Questionable Historiography
The most damning section of the review targets the book's use of history. Smith points out that while the general idea that technology can be misused is a truism, the specific examples chosen by Acemoglu and Johnson are factually weak. He highlights their treatment of the Haber process, a chemical breakthrough that created artificial fertilizers. The authors claim this invention "brought nothing like shared prosperity" because it was also used to create chemical weapons in World War I. Smith dismantles this claim with stark data, writing, "Nitrogen fertilizers are so important to human existence that by the most common estimates, about half of the entire population of Earth — 3.5 billion people — is only sustained thanks to this technology." To label a technology that feeds billions as a failure is, in Smith's view, a disservice to the reader.
Similarly, Smith challenges the book's interpretation of the Industrial Revolution and the Luddites. While Acemoglu and Johnson argue that early textile automation immiserated the British working class, Smith points to modern parallels where similar technologies lifted millions out of poverty. He notes that in China and Bangladesh, the adoption of garment manufacturing technologies coincided with skyrocketing wages and tripling incomes. "The same is now true of Bangladesh — the country focuses relentlessly on the garment industry, and has access to all of the old automation technologies, and yet incomes in the country have tripled since 1990," Smith writes. This suggests that the relationship between automation and prosperity is far more complex than the book's binary narrative allows.
The historical examples it cites simply don't support a narrative of out-of-touch technologists inventing the wrong sorts of technologies and hurting workers in the process.
Smith also scrutinizes the book's treatment of the Panama Canal and Henry Ford. He argues that labeling the canal a "colossal failure" ignores the fact that it elevated Panama's standard of living above its neighbors, and that praising Ford's labor relations ignores the violent suppression of union organizers. "Henry Ford was a pioneer in developing a more cooperative relationship... I'm just wondering how this 'more cooperative relationship' involved hiring thugs to gun down union organizers," Smith asks. These examples, he contends, reveal a pattern of cherry-picking facts to fit a pre-determined ideological framework.
The Case for Critical Scrutiny
Beyond the specific historical errors, Smith raises a procedural concern about the book's scholarship. He notes the absence of traditional footnotes, replaced by a "bibliographic essay" that makes it difficult to verify claims. "Instead of citing specific works in support of each specific claim... it is often difficult, and sometimes impossible, to match the sources to specific claims," he writes. This lack of transparency forces readers to either trust the authors implicitly or do the legwork themselves to debunk the assertions. Smith admits that reading the book was "very frustrating" and that he "took out that frustration by writing a very long and highly critical review."
Critics of Smith's approach might argue that he is applying a level of granular historical scrutiny that popular economics books rarely face, potentially missing the forest for the trees. The book's goal may be to provoke a broader philosophical conversation about the direction of AI, rather than to serve as a definitive history textbook. However, when a book's central policy recommendation relies on specific historical precedents, the accuracy of those precedents becomes critical. If the foundation is shaky, the policy advice—"steering" innovation—may be built on sand.
Bottom Line
Noah Smith's review serves as a vital corrective to the uncritical adoption of Power and Progress as the definitive guide to the AI era. While the book's call to align technology with human welfare is morally urgent, its historical evidence is too flawed to support its sweeping policy conclusions. The strongest part of Smith's argument is his demonstration that automation has historically coexisted with rising wages in developing economies, challenging the inevitability of worker displacement. The biggest vulnerability in the book's thesis is its reliance on cherry-picked anecdotes that ignore the broader, more complex reality of technological adoption. Readers should watch for how the "steering" concept evolves in policy debates, but they should do so with a critical eye, demanding evidence that matches the book's high ambitions.