Descriptions:
Nate Herk presents a 2+ hour course on constructing a personal AI operating system (AIOS) inside Claude Code, drawing on his background scaling an AI automation agency to over $100,000 per month before selling it. The core premise is that Claude Code, configured with the right skills, memory, and integrations, can function as a persistent intelligence layer across an entire business — replacing the need to context-switch between Slack, email, Google Docs, and other apps by giving a single agent access to all of them simultaneously.
The course introduces Herk’s “Three M’s of AI” framework (mindset, method, machine) before moving into hands-on implementation. Key integrations demonstrated include a Google Workspace MCP setup that enables Claude Code to search Google Docs by date and topic, automatically score and triage unread emails based on business context, and generate Google Slides decks with iterative visual validation using Chrome DevTools screenshots and the Nano Banana 2 image model for on-brand visuals. The complete setup guide is available as a free download.
A recurring theme is tool agnosticism: structuring the AIOS so that the core skill and memory architecture survives model changes and API deprecations. Herk demonstrates this by migrating his entire AIOS to OpenAI Codex in under two minutes, then back to Claude Code. The course is aimed at both individual operators and those interested in building and selling custom AIOS setups as a service, and reflects an emerging practice of treating Claude Code’s `.claude/` folder as a durable personal infrastructure layer rather than a disposable prompt file.
📺 Source: Nate Herk | AI Automation · Published May 01, 2026
🏷️ Format: Course Lesson
