A reading list for generalists
A reading list for generalists
LessWrong post by Dylan Bowman (June 29, 2026). Bowman, identifying as a reasonably strong generalist working in AI safety, argues that the field has a shortage of generalists and that many important projects and efforts will simply not happen unless owned by capable generalists.
He assembled this reading list of 18 documents (15 blog posts, 1 book, 1 blog comment, and 1 interview transcript presented as a blog post) that have helped him most personally. The claim is practical: if you are already working hard as a generalist on a project you care about, rigorously applying the lessons in these pieces will make you improve faster.
By the numbers
- Authors cited more than once: Paul Graham (5), Ben Kuhn (4), Ethan Perez (2), Greg Brockman (2).
- Sam Altman and Eliezer Yudkowsky have fingerprints over much of the material.
- The list is deliberately tactical rather than a survey of "why generalists matter."
Dispositional
What characteristics should you try to adopt?
- Paul Graham: "What We Look for in Founders", "Relentlessly Resourceful"
- Eliezer Yudkowsky: "Shut Up and Do the Impossible!"
- Ben Kuhn: "Be impatient"
- Cate Hall: "How to be more agentic"
Strategy
How do you make good decisions with the information you have, and how can you get the additional information you need?
- Anna Salamon: "Humans are not automatically strategic". Bowman singles this out as the single most recommended item on the entire list. Reasons: it helps you see the ways otherwise-intelligent people (including yourself) fail to be strategic, and he reports having gotten "a ton of mileage" simply by becoming more strategic. He notes that the most effective coworker he has worked with is defined by being "unrelentingly strategic."
- Holden Karnofsky: "Learning By Writing"
- Ethan Perez with Mikita Balesni and Henry Sleight: "How I select alignment research projects"
- Paul Graham: "Do Things that Don’t Scale". Emphasis: do things that give you maximal information.
- Ben Kuhn: "Impact, agency, and taste"
- Richard Ngo: comment on "Against Almost Every Theory of Impact of Interpretability". The original post is domain-specific; Ngo’s response illustrates the general principle that winning at whack-a-mole does not conclusively discredit forward-chain approaches.
Project leadership
How do you get projects done when your boots are on the ground?
- Ben Kuhn: "How I've run major projects", "Categories of leadership on technical teams"
- Greg Brockman: "#define CTO", "#define CTO OpenAI". Bowman highlights these for showing how someone with a spiky skillset can find their place in an organization.
- Ethan Perez: "Tips for Empirical Alignment Research". While some advice is domain-specific to AI safety research, the document offers strong guidance on becoming an excellent individual contributor (especially a managed IC) in any domain. Notable elements include making your work legible to others and scoping project timelines appropriately.
Interpersonal/organizational
How can you understand the structure of an organization and contribute positively?
- Kevin Simler & Robin Hanson: The Elephant in the Brain (2018). Bowman calls it the best book for organizational psychology precisely because it is the best book for human psychology in general. The first section, covering the broader theory of signaling, is the key part. The lens helps you understand the (often unstated) values and motives of people around you when navigating complicated settings. Equal importance: the same lens applies to yourself — you are a monkey signaling genetic fitness through unstated means to increase status and sexual marketplace value, just like everyone else.
- Martin Sustrik: "Accountability Sinks"
- Paul Graham: "Keep Your Identity Small"
- Paul Graham: "How to Disagree"
Notes
The list is explicitly positioned as a practical operating manual for generalists already in motion, not an abstract celebration of breadth. The heavy representation of Paul Graham and Ben Kuhn pieces reflects a focus on resourcefulness, impatience, agency, information-maximizing tactics, and concrete leadership patterns.
Worth returning to for its tight curation and the author's personal "this is what actually moved the needle for me" commentary, especially the repeated emphasis on unrelenting strategy and information-seeking moves that don't scale.
This tactical list serves as a practitioner companion to the wiki's existing generalist material, including polymath, match-quality, kind-vs-wicked-learning-environments, and sampling-period.
Processed members of the list now include source and concept pages for: humans-are-not-automatically-strategic-anna-salamon, learning-by-writing-holden-karnofsky, what-we-look-for-in-founders-paul-graham, relentlessly-resourceful-paul-graham, be-impatient-ben-kuhn, how-to-be-more-agentic-cate-hall, shut-up-and-do-the-impossible-eliezer-yudkowsky, impact-agency-and-taste-ben-kuhn, how-ive-run-major-projects-ben-kuhn, categories-of-leadership-on-technical-teams-ben-kuhn, define-cto-greg-brockman, define-cto-openai-greg-brockman, tips-for-empirical-alignment-research-ethan-perez, how-i-select-alignment-research-projects-ethan-perez, do-things-that-dont-scale-paul-graham, against-almost-every-theory-of-impact-of-interpretability, accountability-sinks-martin-sustrik, keep-your-identity-small-paul-graham, how-to-disagree-paul-graham, the-elephant-in-the-brain-kevin-simler-robin-hanson (see full cluster in raw-cluster-audit.md, now full).