Introducing Code with Stein, a new 12-part video series we’re collaborating on with Stein (@codewithstein), a web developer with more than 15 years of experience. The series explores what happens when you connect AI models to real systems through MCP: what you can build, what can go wrong, and how to put better controls in place. Stein will walk through building MCP servers, spotting prompt injection risks, managing identity and API keys, and understanding just how much access an agent has. He’ll also cover vetting third-party servers, logging agent activity, tracking costs, and his 15 most-used MCP servers as a developer. Along the way, he’ll share his own mistakes and what he learned from them. Watch the intro for a preview of what’s coming, and subscribe to follow along as each video drops. buff.ly/jGJ7R6Z