How to "Teach" AI to Teenagers (christinaasquith.substack.com)

Hack Club is pitching a radical alternative to classroom-driven AI education: hands-on, volunteer-driven project building. This year its teenagers produced nearly 17,000 open-source technical projects (most using AI), more than the number of AP Computer Science students in 30 U.S. states, and the organization aims to have 100,000 teens building AI-enabled projects by the end of next year. Hack Club funds prizes (about $1M distributed this year), runs large online competitions and in-person events, and gives students free access to modern AI tooling to accelerate development and learning. For the AI/ML community this matters because project-based learning produces practical “AI fluency” faster than lecture-driven courses. Teens use tools like GitHub Copilot, ChatGPT, Claude, Cursor and Windsurf to code, while also experimenting with GPUs, algorithms, neural networks, infrastructure and hardware (from mobile apps to custom OSes, flight simulators and circuit boards). They confront real-world issues—data, platforms, statistics, servers, code and ethics around originality—while developing a founder mindset. The model implies a shift in workforce dynamics (many more founder-like creators empowered by AI) and signals that scalable, exploratory, tool-centric education could be a more effective pipeline for AI talent than traditional curricula.