Claude Code ABQ

Marty unveiled what he'd been quietly building for five months: a real production web app built almost entirely with Claude Code. He walked through the agentic technique behind it — running separate mentor and builder instances of Claude Code that work together, with the mentor directing and verifying the builder's work.

During the open demo floor, Tim showed off Ear Buddies, an app he built to teach ear training — learning to recognize musical tones by listening. Try it yourself at the demo link below.

Jacob manages security efforts at Bernalillo County. He took us beyond the chatbot subscription and into AI infrastructure at enterprise scale — what it takes to engineer sovereign, on-prem AI for the public sector, where data can't leave the building.

No speaker this month — the room did the coaching. Volunteers who had never used Claude Code drove real first sessions live, with experienced folks guiding from the seats. One left with the start of an inventory system for their job, plus a free week of Claude Pro to keep building. Everyone else brought what they were working on, asked questions, and shared what's working.

Tim showed how he uses git worktrees to run multiple Claude Code sessions in parallel, each on its own branch, without them stepping on each other — building a feature in one worktree while Claude fixes a bug in another. A hands-on look at a workflow you can set up the same night.

George opened up his Claude Code workflow and showed us how he actually works — a live look at the day-to-day habits, setup, and process behind his AI-assisted development.

Marty walked through his full Claude Code configuration — status line, all 28 active plugins, skills, agents, slash commands, and a multi-device tmux workflow that runs Claude Code across phone, laptop, and desktop over SSH + Tailscale. The session capped off with a live plugin build (power-pages-liquid-js-plugin) created on the spot for an attendee.

Explored how agentic AI is changing the math on what one person can build. Tim walked through how founders can validate faster, iterate continuously, and scale lean — with a real-world ed-tech platform as the case study.

Michael broke down OpenClaw — the open-source AI agent with 250K+ GitHub stars in four months. Covered what it is, why it matters, and how he uses it in his own workflow.

Explored how the biggest AI advantage isn't one developer prompting better — it's one developer building systems that make the entire team better. Covered how to design shareable AI workflow tools that give every developer on your team consistent, high-quality results from AI coding agents.

Live demo of Claude Code building tweeter-abq-2 (a Twitter-like app) in real time. Covered agentic workflows, spec-driven development with SpecSwarm, and how to guide an AI coding agent through a full project — from initial scaffold to working app.