CMSC 398Z - Effective use of AI coding assistants and agents

Planned assigned readings

There will be assigned readings each week. The following are some of the readings that will be assigned in the first part of the semester.

• week 1: (due Sept 12th)

  • NYTimes How Do You Teach Computer Science in the A.I. Era?
  • The Fly Blog - My AI Skeptic Friends Are All Nuts
  • Harper Reed - An LLM Codegen Hero's Journey
  • optional: How AI Coding Agents Will Change Your Job

• week 2: How people are coding with AI tools (due Sept 19th)

  • Harper Reed - My LLM codegen workflow atm
  • Simon Willison - Here's how I use LLMs to help me write code
  • Decoder - Why tech is racing to adopt AI coding

• Teaching coding wih AI tools - Leo Porter @ UCSD

  • In This Era of AI, Will Everyone Be a Programmer?
  • CS1-LLM: Integrating LLMs into CS1 Instruction

• How Google is using AI coding tools

  • AI in software engineering at Google: Progress and the path ahead
  • What do professional software developers need to know to succeed in an age of Artificial Intelligence?
    • Read sections 1 and 5, skim rest of paper
    • or listen to NotebookLLM summary (coming)
  • How much does AI impact development speed? An enterprise-based randomized controlled trial
    • Read Sections 1, 3, 7 and 8, skim rest of paper
    • or listen to NotebookLLM summary (coming)

• Impact of AI on the broader undergraduate curriculum

  • Asking AI for a website of the Catalog of Ships by Ethan Mollick
  • Will the Humanities Survive Artificial Intelligence? | The New Yorker
  • What Happens After A.I. Destroys College Writing? | The New Yorker

Confirmed topics for first part of semester

We are continuing to evolve the course and the field is rapidly changing. We have fairly high confidence in this part of the schedule:

• Intro & Purpose
• Setting the stage and configuring your computer
  • Python
    • lists, dictionaries, tuples
    • library management, version compatibility, using uv
    • doctest and unittest
  • VSCode
  • csv files, dataframes and databases
  • json
  • git basics
  • markdown
  • Additional useful background reading: https://missing.csail.mit.edu/
• Using GitHub Copilot in VSCode
  • GitHub slash commands: /explain, /fix, /doc
  • Autocomplete
  • Chat
  • Prompt and context engineering for software development
  • Planning and documenting design decisions
  • Concerns about correctness and security of generated code
• Building software that uses LLMs
  • Simon Willison - Building software on top of Large Language Models, (offered at PyCon)
  • Structured output & constrained generation
  • Run Your Own AI – AnthonyLewis.com
  • Security and prompt injection
    • The lethal trifecta for AI agents
• When AI generated code passes all test cases, what else matters?
  • Maintainability
  • Software architecture, API design
  • Efficiency and scalability
  • Security
  • Solving the right problem

Likely additional topics

We will likely cover most of the following topics, but the order of topics and the precise set of tools for each topic are subject to change. As the schedule firms up, topics will be moved to the above list.

• Other coding models
  • Cursor ai
  • Claude code
  • Google announces Gemini CLI: your open-source AI agent
• Other ways to access LLMs
  • Google AI studio
• Using AI tools with continuous integration, commit hooks, code reviews
  • git branching, forking and pull requests
  • Continuous integration, commit hooks
  • code reviews
  • GitHub Copilot agent
  • Asynchronous coding agents and AI code reviews
    • Gemini Jules
    • OpenAI codex
    • LangChain
    • Code Rabbit
• Vibe coding, and environments that allow you to use AI to create and deploy web applications.
  • Claude artifacts
  • Google AI studio
  • GitHub Sparks

Additional topics that might be covered

We may cover the following topics:

• AI assistants for UI/UX
• MCP
• agents tool use