Book Recommendations from 2025

(Read in 2025 — not necessarily published in 2025)

What connects these books is not ideology or genre. It’s that each takes a system we tend to romanticize—markets, technology, states, intelligence, liberation—and shows how it actually functions once you factor in constraint, power, and path dependence.

Probably Approximately Correct — Leslie Valiant

Valiant is asking a deceptively simple question: how is learning possible at all in a world that is too complex, noisy, and adversarial to ever be fully understood? Drawing on computational learning theory, evolutionary biology, and cognitive science, he argues that intelligence does not come from optimizing against a perfect model of reality. It comes from rules and heuristics that are good enough to survive repeated interaction with the environment.

The book matters because it quietly demolishes the fantasy that better data or more compute will give us mastery—whether in AI or public policy.

Counter-intuitive fact: Valiant shows that biological evolution itself can be formalized as a learning algorithm operating under severe information constraints. Intelligence emerges from survivability, not optimization.
Author note: A Turing Award winner, Valiant wrote this book late in his career to unify AI, biology, and learning—largely as a corrective to AI overconfidence.

The Golden Road — William Dalrymple

This is not a story of empires or conquest. It’s a story of ideas, techniques, and administrative knowledge moving across Asia long before Europe became central. Dalrymple traces how Indian mathematics, astronomy, medicine, literature, and statecraft were transmitted through merchants, monks, scholars, and bureaucrats—shaping Islamic and Asian civilizations in lasting ways.

What makes the book useful today is its implicit argument about power: influence travels most effectively through standards and mental models, not armies.

Counter-intuitive fact: Indian numerals were embedded in Asian and Middle Eastern trade systems centuries before Europe adopted them, via North Africa.
Author note: Dalrymple works directly with Persian, Sanskrit, and Urdu sources, allowing him to reconstruct intellectual networks most Western histories ignore.

Small, Medium, Large — Colleen Dunlavy

Dunlavy dismantles the founding myth of American capitalism. The U.S. did not industrialize through small firms scaling up in free markets. It industrialized through a particular kind of state—one that shaped firm size, standardized production, underwrote risk, and created demand through procurement and tariffs.

The book reframes industrial policy as something states do constantly, whether they admit it or not.

Counter-intuitive fact: Early American manufacturing firms were often larger and more vertically integrated than European ones because state institutions made scale viable.
Author note: Dunlavy has spent her career treating firms as political institutions, not neutral market actors—decades before industrial policy became fashionable again.

House of Huawei — Eva Dou

This is the clearest anatomy I’ve read of how a modern technology firm operates under geopolitical pressure. Huawei is neither a simple arm of the Chinese state nor an independent market actor. It is an organization built to survive ambiguity—legal, political, and strategic.

Dou shows how governance structure, internal competition, R&D discipline, and long time horizons make Huawei unusually resilient.

Counter-intuitive fact: Huawei reinvests a higher share of revenue into R&D than most Silicon Valley firms—despite sanctions.
Author note: Dou built deep sourcing inside both Chinese corporate systems and U.S. policy circles, which is why the book avoids caricature.

The Shadow of the Sun — Ryszard Kapuściński

This is Africa seen before “development” hardened into frameworks and metrics. Kapuściński reports from revolutions, coups, droughts, and everyday life when post-colonial states were still improvising authority and legitimacy.

It’s not a policy book, but it explains something most policy books miss: how fear, rumor, fatigue, and dignity shape institutional outcomes.

Counter-intuitive fact: Several newly independent African states began with fewer trained civil servants than colonial administrations decades earlier.
Author note: Kapuściński often worked alone, speaking local languages, which is why his reporting captures institutional fragility from the inside.

Milton Friedman: The Last Conservative — Jennifer Burns

Burns treats Friedman not as an economist, but as a political entrepreneur. The book shows how his ideas succeeded because they were simple, portable, and institutionally embeddable: into think tanks, universities, media, and policy machinery.

It’s a study in how paradigms win.

Counter-intuitive fact: Friedman rose to dominance after Keynesianism had already begun collapsing on its own terms—he filled a vacuum more than he defeated an orthodoxy.
Author note: Burns is an intellectual historian who treats ideas as technologies of power, not just arguments.

A Grain of Wheat — Ngũgĩ wa Thiong’o

Set on the eve of Kenyan independence, this novel strips liberation of romance. It shows how secrecy, betrayal, and moral compromise become survival strategies—and how they poison the post-independence state.

The book is essential for understanding why post-conflict governance is so fragile.

Counter-intuitive fact: Ngũgĩ suggests colonial violence often survives independence, reproduced by new elites using the same tools.
Author note: Ngũgĩ later abandoned English for Gikuyu as an act of intellectual decolonization, at major personal cost.

The Paper Menagerie — Ken Liu

These stories are not about technology so much as about translation—between languages, generations, and value systems. Progress always creates loss, and Liu insists we account for it.

It’s science fiction as moral calibration.

Counter-intuitive fact: The title story is the only work to win the Hugo, Nebula, and World Fantasy awards simultaneously.
Author note: Liu is both a software engineer and a major translator of Chinese science fiction, bridging technical and cultural worlds.

Other Minds — Peter Godfrey-Smith

A book about octopuses that ends up dismantling human exceptionalism. Godfrey-Smith shows that intelligence evolved multiple times, in radically different forms.

It’s a useful corrective for anyone who assumes there is one model of rationality.

Counter-intuitive fact: Octopus intelligence evolved entirely independently from mammals—complex cognition is not rare, just contingent.
Author note: Godfrey-Smith is a philosopher who became a diver to study octopuses in their natural habitats.

An African History of Africa — Zeinab Badawi

This is Africa told forward—through archaeology, trade, climate, language, and political organization—rather than as a prelude to colonization.

The book resets the baseline for thinking about African institutions.

Counter-intuitive fact: Africa will drive the majority of global labor-force growth this century.
Author note: Badawi has interviewed nearly every major African leader of the last four decades.

Empire of AI — Karen Hao

Hao shows that AI power concentrates not by accident, but by design—through capital requirements, data access, and regulatory gaps.

It’s less about algorithms than about oligopoly.

Counter-intuitive fact: Training frontier AI models now costs on the scale of national infrastructure projects.
Author note: Hao broke major AI labor and governance stories at MIT Technology Review before they became mainstream.

Messy Cities — edited by Dylan Reid

An argument against the fantasy of total control in urban planning. Cities work because of informality, adaptation, and layered use—not despite them.

Counter-intuitive fact: Informal settlements often outperform formal housing in adaptability and social resilience.
Author note: Reid bridges planning theory and journalism, forcing abstraction to confront lived reality.

The Land Trap — Mike Bird

Bird shows how land quietly distorts growth, inequality, and politics by channeling capital into speculation rather than productivity.

It’s a political economy book disguised as housing analysis.

Counter-intuitive fact: In the UK, over half of household wealth growth since the 1990s comes from land appreciation.
Author note: Bird turned to land economics after noticing how housing crises masked political choices.

Capitalism: A Global History — Sven Beckert

Beckert strips capitalism of its myths, showing how markets were built through state power, coercion, and empire.

It’s a reset for debates on markets versus states.

Counter-intuitive fact: Global cotton markets relied on coerced labor long after slavery’s formal abolition.
Author note: Beckert helped found the modern “history of capitalism” field.

Chokepoints — Edward Fishman

A manual for modern power. Sanctions work not because they are broad, but because they target narrow dependencies.

Counter-intuitive fact: Control over insurance, clearing, and standards often matters more than control over production.
Author note: Fishman worked directly on U.S. sanctions policy.

Beyond Banks — Dan Awrey

A grounded book on fintech and money that refuses ideology. It’s about trust, finality, and systemic risk.

Counter-intuitive fact: The most important payments innovation is not speed—but reversibility and final settlement.
Author note: Awrey is a former derivatives lawyer obsessed with failure modes.

A Fine Balance — Rohinton Mistry

A novel that understands second-order effects better than most policy papers. Set during India’s Emergency, it shows how everyday bureaucratic cruelty corrodes trust and dignity.

Counter-intuitive fact: The Emergency did more damage through routine administrative abuse than overt repression.
Author note: Mistry trained as a mathematician; his structural precision shows.

---

How this changed how I work

These books trained me to stop asking whether policies are well-intended and start asking whether they are structurally survivable.

I now assume systems will be stressed, misused, and politicized—and design backward from that reality. It has made my work less elegant, and far more honest.