Code the Dream AI Course Development

A workspace to develop hands-on, accessible AI teaching materials for Code the Dream. This informal space is to experiment with lesson plans, identify the most important skills, and create learning paths for the AI bits in Python 200. Student facing-material will ultimately be part of the full Python 200 course at Code the Dream.

Background and motivation

AI has become a huge field that could easily cover a year of material. In Python 200, we have three weeks, and will focus in on a few core topics:

• Week 1: Intro to LLMs: From large-language-models to chatbots.
• Week 2: Retrieval-augmented generation (RAG): Supplementing AI with external knowledge bases.
• Week 3: Agents: Giving AI systems the ability to plan and use external tools.

For more details about the plan for a given week, please click on the link for that week.

Additional considerations

For Week 1: AI literacy beyond just technical considerations is essential. Students should not just understand the basics of how LLMs work and what their limitations are, but also the significant ethical dimensions of this work. We won’t have time to dive deeply into all these issues, but they should be acknowledged and incorporated into discussions week. Some key points to cover:

• LLMs are trained on a vast corpus of internet data — this introduces bias, misinformation, and uneven quality.
• Ethical concerns include fairness, access, environmental impact, and explainability (there is a great overview of ethical considerations here).
• Many public discussions of AI are full of hype. Students should be equipped to think critically when people make outlandish claims (e.g., about AGI). This requires that they understand how the models work -- at least at a coarse level.
• Students should have a general grasp of concepts like transformers, attention mechanisms, and the strengths/limits of LLMs (e.g., text vs. visual reasoning).