A reading list for generalists

AI safety researcher and generalist published a curated reading list of 18 essays and blog posts aimed at helping generalists improve their effectiveness. The list, which includes works by Paul Graham, Ben Kuhn, and others, covers dispositional traits, strategic decision-making, and project leadership. The author hopes to address a perceived shortage of generalists in the AI safety community.

I, along with many others in AI safety, believe there is a shortage of generalists in the community and that there exist many projects and efforts that by default will not happen unless they are owned by a strong generalist. As someone who is a reasonably good generalist, I decided to assemble a reading list of the essays and blog posts that have personally helped me the most. I would love others to comment with pieces they think should be on this list. The crux of this reading list is the idea that if you’re working hard as a generalist on a project you care a lot about, then by rigorously applying the lessons from these documents you will improve more quickly than you otherwise would. By the numbers: - I’ve attached 18 documents to start this reading list. - The authors cited more than once are Paul Graham 5 , Ben Kuhn 4 , Ethan Perez 2 , and Greg Brockman 2 . Sam Altman and Eliezer Yudkowsky also have their fingerprints over a lot of the content. - The items are 15 blog posts, 1 blog comment, 1 interview transcript in blog post form, and 1 book. Dispositional What characteristics should you try to adopt? - Paul Graham: "What We Look for in Founders" link https://paulgraham.com/founders.html , "Relentlessly Resourceful" link https://www.paulgraham.com/relres.html - Eliezer Yudkowsky: "Shut Up and Do the Impossible " link https://www.lesswrong.com/posts/nCvvhFBaayaXyuBiD/shut-up-and-do-the-impossible - Ben Kuhn: "Be impatient" link https://www.benkuhn.net/impatient/ - Cate Hall: "How to be more agentic" link https://usefulfictions.substack.com/p/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" link https://www.lesswrong.com/posts/PBRWb2Em5SNeWYwwB/humans-are-not-automatically-strategic - If I were to recommend one single item from this list it would be this one because 1 it’s good to understand the ways in which otherwise-intelligent people are unstrategic, and 2 it’s good to understand the ways in which you are not automatically strategic. I have gotten a ton of mileage in my short career thus far simply by being more strategic. The defining trait of the most effective coworker I’ve ever had is that he is unrelentingly strategic. - Holden Karnofsky: "Learning By Writing" link https://www.cold-takes.com/learning-by-writing/ - Ethan Perez with Mikita Balesni and Henry Sleight: "How I select alignment research projects" link https://www.lesswrong.com/posts/7GmDs4BqrFW3kk4nP/how-i-select-alignment-research-projects - Paul Graham: "Do Things that Don’t Scale" link https://paulgraham.com/ds.html - In particular, do things that give you maximal information. - Ben Kuhn: "Impact, agency, and taste" link https://www.benkuhn.net/impact/ - Richard Ngo: comment on "Against Almost Every Theory of Impact of Interpretability" link https://www.lesswrong.com/posts/LNA8mubrByG7SFacm/against-almost-every-theory-of-impact-of-interpretability-1?commentId=PazvJmFcv4DBXepzi - The original post is great but domain-specific. Ngo’s response shows the more general principle that winning whack-a-mole doesn’t conclusively discredit forward-chainy approaches. Project leadership How do you get projects done when your boots are on the ground? - Ben Kuhn: "How I've run major projects" link https://www.benkuhn.net/pjm/ - Ben Kuhn: "Categories of leadership on technical teams" link https://www.benkuhn.net/leadcats/ - Greg Brockman: " define CTO" link https://blog.gregbrockman.com/figuring-out-the-cto-role-at-stripe , " define CTO OpenAI" link https://blog.gregbrockman.com/define-cto-openai - I think these posts are especially useful in showing the process a person with a spiky skillset can take to find their place in an organization. - Ethan Perez: "Tips for Empirical Alignment Research" link https://www.lesswrong.com/posts/dZFpEdKyb9Bf4xYn7/tips-for-empirical-alignment-research - Some of the advice is domain-specific to AI safety research, but the document as a whole provides great guidance on becoming an excellent IC, in particular an excellent managed IC, across any domain. Some good bits are the ones about making your work legible to others and scoping your 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 link https://en.wikipedia.org/wiki/The Elephant in the Brain - The Elephant in the Brain ended up being the best book for organizational psychology by virtue of being the best book for human psychology in general. I think that the first section which goes over the broader theory of signaling is the important part to read. Understanding the values and motives of the people around you can help greatly in navigating complicated settings, but of equal importance is understanding that the core lens also applies to you: you’re a silly monkey trying to signal its genetic fitness via unstated means in order to increase status and sexual marketplace value like the rest of us. - Martin Sustrik: "Accountability Sinks" link https://www.lesswrong.com/posts/nYJaDnGNQGiaCBSB5/accountability-sinks - Paul Graham: "Keep Your Identity Small" link https://www.paulgraham.com/identity.html - Paul Graham: "How to Disagree" link https://www.paulgraham.com/disagree.html Discuss https://www.lesswrong.com/posts/sH4cFDDjRdGrn3p2o/a-reading-list-for-generalists comments