READING LIST — The AI-Native Operator's Reading List

Thematic & Conceptual · 12 resources · multi-format (books, essays, talks, podcasts) · ~build over a month

Framing — who this is for, and why

This is for the person who has to make decisions in the age of capable AI — a founder, an operator, a manager — not for the researcher and not for the doomscroller. The goal isn't to make you an AI expert; it's to give you the working models to decide what to build, what to automate, what to defend, and how to keep your own judgment intact when the model sounds certain. It deliberately balances capability optimism (what's now possible), strategic realism (where the durable advantage actually is), and critical counterweight (which claims are hype). Read it and you'll stop asking "is AI good or bad?" and start asking the better questions: good at what, for my work, and defensible how?

Each entry says what it is, why it earns a slot, and where it connects into the rest of this library.

Start here — the working relationship

1. Co-Intelligence — Ethan Mollick (book). The most practical single starting point: AI as a collaborator, the jagged frontier of what it can and can't do, and four rules you can apply within the hour. Read it first because it installs the stance everything else refines. → Full summary in this library; pairs with the Jagged Frontier and Bitter Lesson mental models.

2. "One Useful Thing" — Ethan Mollick (newsletter, ongoing). The living companion to the book. Because any AI book dates fast, a well-run newsletter is how you keep the frontier map current. Subscribe and skim; it's the antidote to reading one 2024 book and thinking you're done.

The strategic core — where the advantage really is

3. "The Bitter Lesson" — Rich Sutton (essay, ~5 pages). The most important five pages here. Scale and general methods beat clever hand-crafted structure over time — which tells you not to build your moat out of cleverness the base model will subsume. → Full mental-model page in this library; the spine of the two entries below.

4. 7 Powers — Hamilton Helmer (book). The strategy classic that answers the question the Bitter Lesson raises: if the core capability commoditizes, where's the moat? Answer: the seven durable powers — data-network-effects, switching costs, counter-positioning, cornered resources. → Anchors the entire Moats category.

5. Competing in the Age of AI — Iansiti & Lakhani (book). What changes when the firm's core becomes an "AI factory": the old ceilings on scale, scope, and learning lift, and the operating model — not the product — becomes the battleground. The best bridge from "AI as a tool" to "AI as the center of the company." → Pairs with the Vertical AI Agents framework.

6. The Cold Start Problem — Andrew Chen (book). Because most defensible AI businesses still win or die on network effects and distribution, not model quality. How to get a network past critical mass, one atomic market at a time. → Pairs with Network Economies (moat) and the Do-Things-That-Don't-Scale mental model.

The frontier — what's coming, argued from both sides

7. Situational Awareness — Leopold Aschenbrenner (essay series, 2024). The most-discussed bull case for rapid capability scaling and its geopolitical stakes. Read it for the strongest version of "this goes much further, fast" — then read the next two as ballast.

8. The Coming Wave — Mustafa Suleyman (book). From a builder-turned-insider: why AI (and synthetic biology) form a wave that's extraordinarily hard to contain, and what the containment problem means for anyone deploying it. The serious version of the risk conversation, minus the theatrics.

9. AI Snake Oil — Narayanan & Kapoor (book). The essential skeptic. A field guide to telling real AI capability from hype and broken predictive claims. Read it against entries 7–8 so you leave with calibration, not a vibe. → Pairs with the mental models on Survivorship Bias, Base Rates, and the Expert Problem.

Craft, judgment, and the human layer

10. "Machines of Loving Grace" — Dario Amodei (essay, 2024). A deliberately concrete, optimistic account of what powerful AI could do well — a useful counter to reflexive doom, and a prompt for "what would we actually build if it worked?"

11. A great podcast in the rotation — e.g., a builder-focused AI show (Latent Space / No Priors / Lenny's Podcast episodes on AI). Format matters: podcasts are how you hear operators reason through the frontier in real time, before it's in any book. Pick one, make it a habit, drop episodes that are hype. → The multi-format discipline of this whole library.

12. AI Snake Oil's opposite number on taste — writing on craft and curation (e.g., Rick Rubin's The Creative Act, or essays on taste-as-moat). When generation is free, judgment about which output is good becomes the scarce, defensible skill. Close the list here on purpose: the human layer — taste, trust, editorial judgment — is where operators keep their edge. → Pairs with the Branding moat and the Rick Rubin playbook.

Related lists

Building in the Age of Agents · The Compute & Semiconductors Reading List · AI Safety & Governance for Operators · Best Strategy Books · Systems Thinking.