Words, code, truth, kindness, other arcana
Hi! In case you want to get to know me or my work, this is a cheat sheet.
-
@ojoshe on Twitter/X (this is best for DMs)
-
github.com/jlevy (for open source)
-
linkedin.com/in/jlevy (slow on messages here)
Please reach out if you think you should!
Generally, my interests lie at the intersection of AI, writing, software engineering, design, and knowledge tools:
-
✨NEW✨ Lately, I’m focusing on workflows for agent work and agent coding:
-
Inspiration: I believe the purpose of technology is to improve our collective knowledge and intelligence and to solve human problems. The technologist I most admire is Doug Engelbart. His work is as relevant today as it was 50 years ago. By sharing a vision of what was possible, he shaped how we use software for decades. I feel privileged to have met him a few times.
-
AI and software: I have a broad interest in the theory, the practice, and the human elements of software engineering, and the history of software and AI. I’ve worked on conversational AI systems since early days with the original Siri team at SRI. I’ve built search engines with prominent ex-Google engineers.
-
Writing and open source: I’ve edited (and published) about a dozen books and I’ve written a few popular guides (2M+ readers), including The Holloway Guide to Equity Compensation (the most widely read guide to equity compensation in startups), The Art of Command Line (150K+ stars on GitHub), and a guide to AWS. And I’ve written a number of open source libraries in Python. Popularity isn’t the best measure of value, but I’m in the top 20 most-starred accounts on GitHub.
-
Publishing: I started and built most of Holloway, a small startup that has rebuilt the digital book publishing process with dozens of in-depth, web-based guides in entrepreneurship, creativity, and personal growth. We also publish print books.
-
Startups and entrepreneurship: I’ve been a founding engineer a few times and a founder. I like working with founders and sometimes consult for or advise startups and investors.
-
Engineering: I’ve loved programming since I was a kid. I have been a founding engineer and engineering leader at several startups in AI, SaaS, and web search, some failed, some successful (backend/distributed systems, full stack web/JS/Node/TypeScript/React, Python, Java, C/C++, and older esoteric things like Lisp and Prolog). Also devops—AWS has given me the honor of being an “AWS Hero.”
If there is one intellectual theme in my life, it is understanding the essential and beautiful complexity of the world as best I can. I aim to know a few things about many topics and many things about a few topics. I majored in physics, math, and computer science and my graduate degree is from Berkeley.
But for me what ultimately matters is simply kindness and human connection.
If you have any oddly specific topic you care a lot about, I’m curious. For me, this includes exceptional books of all kinds, moral philosophy, media theory, human relationships, psychology, mental health, psychedelic medicine, fitness and nutrition, politics and media, startups and entrepreneurship, AI (history and future), product design, typography, software engineering, logic and foundations of mathematics, running or hiking the mountains and deserts of California, fingerstyle guitar, 15th-18th century lute music, early American blues, Brazilian Portuguese, truth, kindness, and some other arcana.
A few selected projects:
-
Agent coding
-
✨NEW✨ rust-porting-playbook: An extensive agent playbook for automated porting of Python to Rust
-
✨NEW✨ tbd: Agent coding power tool with better beads, knowledge injection 👈 check it out, what I use every day now for agent coding
-
✨NEW✨ tryscript: Golden testing for CLI apps (very useful for helping agents write CLIs for agents!)
-
✨NEW✨ Markform: Structured Markdown documents for agents and humans
-
-
On writing
-
✨NEW✨ Practical Prose: Extensive guidelines and tools for high-quality practical writing with agents
-
on-books: Readings and notes on the past, present, and future of books
-
-
Document, text, and LLM libraries ✨NEW✨
-
repren: Agent-friendly bulk replace/rename tool
-
flowmark (Python) and flowmark-rs (auto-port to Rust): The most advanced (and now fastest) auto-formatter for Markdown
-
chopdiff: Parsing, diffing, diff filtering, chunking, and windowed transforms of text to support LLM applications
-
frontmatter-format and sidematter-format: Simple conventions and formats for metadata and assets on any file or alongside any file
-
-
Modern Python infra
-
simple-modern-uv: A minimal, modern Python project template, good if you’re starting a new project, with uv/ruff/Pyright/GitHub actions/PyPI publishing (used by uvtemplate)
-
strif: Tiny, useful Python lib for strings and files
-
funlog: Minimalist decorators for logging, timing, and tallying function calls
-
prettyfmt: A tiny library for more beautiful Python outputs
-
py-app-standalone: Standalone, relocatable Python app builds using uv
-
-
Knowledge and content experimentation
-
Handy hacks wrapping useful projects
A few new writings (these are rough and I’d appreciate feedback) ✨NEW✨
-
The Philosophy of Command Line: Some philosophical reflections on software
-
Linters for Thought: Reflections on Software Engineering and Collective Intelligence (draft!)
Some older writings and guides:
-
the-art-of-command-line: Master the command line, in one page (150K+ stars, available in ~20 languages and one of the top 40 most-starred GitHub repos of all time, but in need of an update, which I hope to do soon!)
-
og-aws: A practical guide to Amazon Web Services (35K+ stars, for a while one of the most popular docs ever for AWS)
-
og-equity-compensation: Original open source version of a guide to equity compensation for startup employees and founders (9K+ stars)






