Building a Second Brain That Works (Recall MCP + Claude Code)

Most "second brains" are just expensive junk drawers — you save everything you wish you knew, and it sits there forever. Here's how to actually close that gap. In this Panama AIMUG talk, Paul Phelps (Chief AI Officer, Productivity Engines) walks through the system he built connecting Claude + Recall's new MCP integration to bridge his personal knowledge base to what he's actually working on — client projects, a book, test builds — instead of just archiving content he'll never revisit. What you'll see: Why a second brain only has value once it's applied, not just stored A live demo of Paul's Claude "learn" skill — it reviews the last 7 days of saved knowledge, quizzes him, and surfaces ideas that connect directly to active work The Andrew Ng quote that reframed how Paul pitches AI to clients (cost-savings has a ceiling, growth doesn't) The two books behind the system's design: Moonwalking with Einstein and Ultra Learning Chapters 0:00 — Why most second brains fail 2:30 — What actually goes into the system (nearly 4,000 items in 18 months) 4:00 — Bridging storage to active projects 8:25 — Live demo: the Claude "learn" skill 11:40 — The Andrew Ng moment that changed his consulting pitch 19:40 — Demo continued + audience Q&A 23:55 — Close: stop collecting, start building Mentioned in this talk: Recall: https://www.getrecall.ai/?via=prodeng... Panama AIMUG: https://www.meetup.com/panama-ai-meet... Productivity Engines: https://www.productivityengines.com