AI First Learning, by Stephen Chin, Cassandra Chin

A keynote talk by Cassandra Chin and Stephen Chin at JFocus conference exploring how to effectively use AI and LLMs as learning tools rather than just code generators. They introduce the GRAPES methodology (Grounded, Reproducible, Academic, Personalized, Episodic) to address LLM shortcomings like hallucinations, non-determinism, and lack of personalization. The talk demonstrates LLM failures with concrete examples, offers prompting strategies to improve learning outcomes, and introduces an open-source VS Code extension providing a curated AI-first programming curriculum. They also discuss declining CS graduate employment rates since ChatGPT's introduction and the growing demand for Gen AI skills in the job market.