What You’ll Learn

In this live AMA, James takes questions from coaches and educators trying to figure out where they fit in the AI shift — and keeps returning to one theme: stop trying to pick the best tool, and instead build an operating system you can talk to. Drawing on nearly 30 years online (his first domain was registered in 1996), he lays out a calmer, more durable way to approach the constant churn of AI news.

By the end you’ll understand why an AI operating system is tool-agnostic, why “talking to AI like a business” beats asking it to do tasks, and why the most valuable skill now is simply knowing how to ask.

The Core Message: Rise Above the Tool Race

It’s easy to feel like you’re missing out — a new model, a new framework, a new benchmark every week. James is blunt that he can’t keep up with it either, and doesn’t try.

“Instead of thinking you know everything and asking it to do something as a tool, step up a level. Have a conversation, treat it as a business.”

💡 In Plain English: don’t shop for a better hammer every week. Hire a team, put a manager in charge, and describe what you want built.

When you have your own campus operating system, you get to work on business fundamentals instead of constantly evaluating which tool is best.

It’s Tool-Agnostic: Claude, Codex, or Gemini

A common question in the AMA: “I don’t use Claude — I’m on ChatGPT, Codex, and Gemini.” James’s answer is that it doesn’t matter. The operating system is just files on your hard drive — skills, workflows, connectors, and instructions in folders. Whatever application you use to interface with a large language model can read those files.

He proves it live: the plugin was built in Claude, but he opens Codex and shows it working on the same files. A few instruction tweaks aside, the skills and workflows don’t care which platform reads them. The operating system is the constant; the LLM is just your interface.

The Shift: From Task Prompts to Business Conversations

The old habit is opening a chat and asking “how do I do X?” James models the new one: you have a digital executive to talk to. You describe a situation — say, a cohort you’re considering — and ask what it thinks, whether it’s worth your time, whether your list can fill it, what to price it at. Together you reach a plan; then you say “put a proposal together,” approve it, and let it run.

That’s a complete change from how we were taught to use these tools — and it’s the skill worth building.

Why This Skill Matters Most

James points to Nvidia’s Jensen Huang: English is the universal programming language, and being able to talk to agents is the most important business skill you can have. The way you build that skill is by having your own operating system to practice on — asking good questions, describing your situation, and asking for proof: “if I did this, what do you think would happen?”

The Takeaway

You don’t need to track every benchmark and framework to win with AI. Get an operating system, start talking to it, and get good at the conversation — describing situations, weighing options, asking for evidence. Organize the work into departments and employees, or just talk to one orchestrator; either way it’s coordinated and it compounds. Start free by joining the community at trainingsites.io/join, and the campus operating system is $97 as a Claude Desktop plugin.

Teach more, and let the agents do the rest.

This tutorial is a recap of a live AMA session.