Pragmatic techniques to get the most out of GitHub Copilot
summary
Explains that GitHub Copilot uses OpenAI large language models as supercharged autocompletes, shows ghost text completions and a chat preview inside Visual Studio Code and Visual Studio, and demonstrates concrete wins: writing an HTML-sanitizer regex, fixing TypeScript null defaults via “Fix using Copilot,” producing cron expressions for an Azure Function (verify the schedule), and generating CSS/layout tweaks. Emphasizes limits—language training data, not a compiler, limited prompt size and file context—and prescribes the prompt rule single, specific, short and the mindset trust but verify.
AI-written summary · about
chapters (12)
- 0:00Introduction to GitHub Copilot
- 2:27How Large Language Models Work
- 5:53Limitations of AI Coding Tools
- 8:34Tips for Effective Prompting
- 12:20Copilot for CSS and Layouts
- 16:55Handling Complex Logic and Tasks
- 22:13Iterating and Context Management
- 29:10Copilot Knowledge Constraints
- 30:31Utilizing Copilot Chat Features
- 36:45Refining CSS with Inline Chat
- 38:26Summary of Best Practices
- 40:00Resources and Q&A
related talks
01h 00mTips & Tricks: Get the Most Out of GitHub CoPilot
01h 00mLevel up with GitHub Copilot: using AI to learn, code, and build
40mLevel Up with Copilot
25mImproving developer happiness with AI
01h 02mLevel up with GitHub Copilot: using AI to learn, code, and build
40mUnlocking Productivity with GitHub Copilot Best Practices and Tips
53mImprove Your Code Quality and Performance using AI and GitHub Copilot
19mMicrosoft Copilot Demystified: A 20-Minute Introduction