Ai
73 posts
Evergreen 1.7 Lets the Phone Do the Work
Evergreen 1.7 turns the iPhone from a place you look things up into a place you actually get things done, plus a handful of fixes worth having.
PyGeoHash v3.3.2: Reliability at the Edges
PyGeoHash v3.3.2 continues the reliability run with antimeridian fixes and stricter validation, plus native ARM64 wheels and smoke-tested artifacts.
What a Month of Portfolio-Wide Engineering Looks Like
651 merged pull requests across 75 repositories in a month, two human gates, and what changes when the unit of maintenance stops being the repo.
MCP Goes Stateless, and Sampling Goes With It
The 2026-07-28 MCP spec turns the protocol stateless, deprecates sampling, and moves tasks into an extension. Why the stateless change is the one that matters.
Distilling Skills From a Knowledge Base You Build While You Work
How to combine Swamp and the Open Knowledge Format: grow an OKF knowledge base across your projects as agents work, then distill reusable skills from it.
PyGeoHash v3.3.1: A Round of Sharp-Edge Fixes
PyGeoHash v3.3.1 is a bugfix release that closes a buffer overrun, rejects bad coordinates, fixes distance and bounding box edge cases, and ships py.typed.
Swamp Club: Deterministic Automation for AI Agents
Swamp is an automation framework built to be operated by AI agents rather than by people writing pipelines. What it is, and how I run my autopilot on it.
Setting Up Lexicon with Claude Code and Claude Desktop
A walkthrough of connecting Lexicon's MCP server to Claude Code and Claude Desktop so your prompt library becomes available as slash commands.
Google's Open Knowledge Format
Google Cloud's Open Knowledge Format is a tiny markdown spec for feeding curated context to AI agents. What it is, and how I've been using it.
Building Reusable Prompt Templates in Lexicon
How to use Lexicon's template system to build reusable, parameterized prompts with variables, validation, and multi-message chains.