Catching what your AI missed
Why your AI-built app still breaks in production, and how to use AuditBuffet to grade it across the dimensions humans care about (security, accessibility, performance, content) before customers do.
Conference sessions and the Vibe Code Club monthly series: practical, build-it-yourself walkthroughs of using AI to ship real internal business tools. Best on a laptop, sound on, drive with arrow keys.
My session at the 10th Annual VBRC. A 40-minute walk through five small internal tools (inquiry form, mini CRM, field schedule, invoice generator, CSV dashboard) built for a fictional landscaping company. Watch the deck, then take the source for any of them and feed it to your AI of choice (Claude, Cursor, Copilot, Gemini, ChatGPT) as a jumping-off point for your own.
A monthly meetup for veteran-owned business owners learning to build their own software with AI. Each session is a single self-contained deck. Drive it with the arrow keys.
Why your AI-built app still breaks in production, and how to use AuditBuffet to grade it across the dimensions humans care about (security, accessibility, performance, content) before customers do.
Move past chat-only AI. Wire Claude into your actual tools through the Claude CLI and MCP servers so it can read files, run commands, hit APIs, and finish tasks instead of describing them.
The single biggest lever for AI output quality isn't the model. It's the context. How to assemble the right files, rules, references, and examples so the AI you already have produces work you can actually ship.