research.ruby.im

Applied research from ruby.im: a small open-source intelligence toolkit built in Ruby, and a connected series on how the best builders, teams and labs extract high leverage — all from public data, with the method shown and the limits stated.

Start here

New to this? Read the leverage map for the big picture, then see it in specific people and specific teams. The tradecraft is how to think about what you find, and the research system is how the data was gathered.

The leverage series

• Every leverage pattern, tier-ranked The map: 23 distinct ways to multiply output, ranked S to C by how much they multiply, how permissionless they are, and how long they last. Start here.
• How high-leverage builders actually work The patterns in people: Simon Willison, antirez, Mitchell Hashimoto, Linear, 37signals and Anthropic's Claude Code team — their real tools and setups, compared and contrasted.
• The high-leverage teams The patterns in organisations: about a hundred teams grouped by how they extract leverage, plus the refined lessons from running the toolkit to assemble it.

Analysis & tradecraft

• Intelligence tradecraft, amplified by AI The deep dive: the highest-leverage analytic techniques intelligence agencies use to understand a complicated world — competing hypotheses, key-assumptions checks, premortems, calibrated forecasting — and what AI does to each.

The research system

• Building an intelligence research stack The whole build: a token-cheap collection toolkit over free structured sources, Katana and GPT Researcher, an honest access policy, and three worked case studies. The method behind the series.
• Retrieval-augmented generation The foundation, on ruby.im: a full how-to for RAG that cites page numbers, with real chunking math, retrieval metrics and an honest benchmark against naive RAG.

Meta

• Changelog Every change to this subsite, newest first.

Provided as is, for information only, with no warranty and no liability. Not legal or professional advice. Security and radio material is for authorized, lawful use only. Full disclaimer.

Visits are logged on our own server: IP address, browser and device details, and pages viewed. We use them to understand traffic and spot attacks, and never share or sell them.