Jesse Genet, a YC-backed founder, runs 11 AI agents from home while homeschooling four children under five. Her stack spans Claude Code, Obsidian, and a Mac Mini, handling tasks from coding to personalized curriculum planning to household logistics. This is not a thought experiment. It is a working system built incrementally, starting at 5 agents and scaling to 11.
The architecture details are the reason to watch this in full. Genet explains her logging systems for tracking children's academic progress, how she structures agent handoffs, and the security considerations of running this infrastructure at home. The curriculum planning agents generate personalized lesson plans and log outcomes, creating a feedback loop that adapts to each child. The phrase 'benevolent neglect' describes both her parenting philosophy and her agent management style, and that parallel is not accidental.
The final segment, starting at 40:04, covers something most AI coverage avoids: what it means to let children interact directly with these systems, including the values questions and failure modes. Genet is a practitioner, not a theorist, and her answers are specific. If you want to understand what agentic AI looks like deployed in a real, non-corporate environment, this is the closest example currently available.
[WATCH ON YOUTUBE →]