Podcast Summary: How I AI

Episode: The Beginner's Guide to Coding with Cursor | Lee Robinson (Head of AI Education)
Date: September 22, 2025
Host: Claire Vo
Guest: Lee Robinson (Head of AI Education at Cursor)

Main Theme

This episode is a practical, hands-on guide to using Cursor, an AI-powered code editor, for both beginners and advanced users interested in coding. Host Claire Vo and guest Lee Robinson break down how Cursor helps users demystify code, automate software engineering tasks, and accelerate learning—even if they've never written a line of code before. The discussion includes tips for setting up projects, using AI agents, and leveraging rules and commands for an efficient workflow. The episode also covers effective prompting styles, code review automation, and tips for learning foundational coding concepts.

Key Discussion Points and Insights

1. What Is Cursor and Who Is It For?

• Cursor’s Position in the Ecosystem:
  • Lee describes Cursor as "an AI code editor" that integrates models from OpenAI, Anthropic, Google, and custom models from Cursor (02:55).
  • Aims to serve both total beginners and power users.
• Why Use Cursor?
  • The code-centric IDE interface makes code visible and accessible—important for learning.
    • Claire: “Looking at code like this is how I learned to code...playing with something like cursor that puts the code structure and syntax front and center might be one of the best ways to actually learn.” (03:30)
    • Lee: "It was very overwhelming. I was like, what do all of these things do?" (04:49)
  • Empowers non-traditional users (e.g., marketers, learners) to automate coding tasks.

2. Practical Coding Tips for Beginners (and Non-Coders)

• Getting Started:
  • You can use web-based coding tools to generate prototype code and then migrate into Cursor to take advantage of its AI functionality (06:27).
  • Claire offers a natural language breakdown of key tools (typed languages, linters, formatters, tests) for total beginners (09:43).
• AI Agent Autonomy:
  • Lee demonstrates how the AI agent handles instructions such as "fix the lint errors," running terminal commands autonomously to identify and repair code issues (11:45, referenced earlier at 00:07).
  • "Rather than giving you step by step instructions, you're just putting the thing into the GPS and it just figures it out along the way." (00:41, 11:45)

3. Key Tools Every Codebase Should Use

• Typed Languages:
  • Forces code consistency, makes error-checking easier for both humans and AI (10:52).
• Linters:
  • Automatically check for code mistakes and formatting; makes AI outputs more reliable.
• Formatters:
  • Automatically format code, removing style and formatting errors as a concern.
• Tests:
  • Provide safety nets; AI can be prompted to fix failing tests automatically (07:44).

4. Productive AI Workflows in Cursor

• Parallel Development:
  • Use the code editor for main tasks, and run agent operations on the side—for example, generating routes or features in the background (15:50).
• Human-in-the-Loop Process:
  • See changes visually with diffs, allowing users to learn from AI-suggested changes (15:08).
• Custom Commands and Rules:
  • Codify repeatable advice or corrections as rules (e.g., code review checklists, banned phrases).
    • Lee: “Rules are just a way for you to codify the places where the models went wrong.” (19:12)
• Natural Language Prompts:
  • Trigger AI actions using everyday language; e.g., “run code review,” “add new route” (21:19).

5. Advanced Configuration and Agent Management

• Model Selection:
  • “Auto” model selector chooses the best AI model for the task, but advanced users can manually select for complex tasks (23:56).
• Context Management:
  • Awareness of chat context limits helps prevent degraded results; start new chats for big features to avoid losing important context (25:33).
• Claire's Tip:
  • “Agents have to context switch...sometimes for those one-off questions...I try to just kick off a one-off chat to answer them...” (26:41)

6. Writing with AI — Customizing Tone and Avoiding Generic Output

• Curation of Writing Rules:
  • Lee uses a “mega prompt” with a list of banned phrases or crutches to make marketing or documentation less generic (29:17).
  • Example: “Instead of saying ‘this is game changing’, you can just say here’s the specific benefits…” (29:17)
• AI as Editor, Not Writer:
  • Lee: Writes his own first drafts, then uses the AI as a reviewer to help clean up language and catch repetitive phrases (34:26).
    • “As with all writing rules, they’re meant to be broken…” (30:53)

7. Entry Points and Learning Pathways for All Skill Levels

• Vibe Coding vs. Traditional Coding:
  • “The term vibe coding is so catchy…The democratization of anybody being able to build software is an amazing thing...But if you want to become a professional developer...you will have to learn a little bit about the code.” (36:42)
• Recommended Languages for Beginners:
  • JavaScript (visual feedback in browser) and Python (readable, easy syntax) (40:28).
  • Lee shares a story of parents teaching their kids to code at a Cursor meetup.
• Concrete Learning Tips:
  • Follow rebuild tutorials for apps you know; read code, decompose it, research unfamiliar terms (42:07).
    • “Read your code. That’s going to be the sub headline of this episode.” (42:46)

Notable Quotes & Memorable Moments

• On AI Agent Autonomy:

  "Rather than giving you step by step instructions, you're just putting the thing into the GPS and it just figures it out along the way."
  — Lee Robinson, (00:41/11:45)

• For Beginners Feeling Intimidated:

  "It was very overwhelming...what do all of these things do?...But over time there’s been better and better abstractions built."
  — Lee Robinson, (04:49)

• On AI as a Patient Teacher:

  "It’s like you’ve got this expert programmer just sitting right beside you who is very, very happy to explain how everything works and it has all the time in the day to explain."
  — Lee Robinson, (28:04)

• On Writing with AI:

  "As with all writing rules, they're meant to be broken. Sometimes I'm like, yeah, actually I'm happy saying 'pretty much.' I know it's a little bit more informal, but it's fine. I like it."
  — Lee Robinson, (30:53)

• The Value of Reading Code:

  "Read your code. That’s going to be the sub headline of this episode."
  — Claire Vo, (42:46)

• On Prompting and Frustration:

  "I am not afraid to admit…I do get a little upset…I have to always remember…when something goes wrong and I’m not very explicit about what actually wasn’t good…it doesn't know exactly what I wanted. So I have to…explain calmly what I actually meant."
  — Lee Robinson, (43:12/44:19)

Timestamps for Key Segments

• What is Cursor and why use it? — 02:55–04:49
• Beginner workflow: Setting up, code editor orientation — 04:49–06:27
• Tasks Cursor can automate (Linting, Typing, Formatting, Testing explained) — 07:44–10:52
• Demo: Agent fixes code autonomously — 11:45–15:17
• Parallelizing coding tasks, human-in-the-loop learning — 15:50–17:26
• Non-engineer applications for code editors — 17:26–18:49
• Advanced setup: Rules and custom commands for agents — 19:12–21:56
• Model selection and context management tips — 23:56–26:41
• Recap of concepts learned — 26:41–28:04
• Learning from the AI agent (asking for explanations) — 28:04
• Customized writing prompts for non-generic AI output — 29:17–34:20
• Best entry points and languages for new coders — 36:42–41:26
• Tips for learning to code (read, decompose, ask for help) — 42:07
• Prompting advice when AI isn't doing what you want — 43:12–44:32

Closing

Lee can be found at leerob.com and @leerob on social media. He’s building a free AI education course on YouTube as part of the Cursor project. The episode inspires listeners to embrace code—no matter their background—by using tools like Cursor, learning through direct experimentation, and leveraging AI not just for automation, but also for growth and education.

TL;DR:
Cursor is a user-friendly AI code editor built to empower both beginners and experts. It's not just an automation tool but also a rich educational environment—whether you're learning code structure, automating reviews, or customizing AI-driven writing. The episode is filled with real-world advice for starting with code, configuring your workflow, and using AI as a tireless, patient coach.