GitHub Copilot - Notes: Part 1
- Radek Stolarczyk
- 7 hours ago
- 3 min read
Mitigating AI Risks
AI provides benefits (innovation, efficiency) but also introduces risks
Main Risks:
Lack of transparency → AI decisions are difficult to understand
Bias and harmful outcomes → unfair decisions, privacy issues
Risk Mitigation:
Governance frameworks → establish rules and policies for AI use
Transparency → ensure AI systems are explainable
Human oversight → humans monitor and guide AI decisions
Responsible AI:
Approach to building AI that is safe, ethical, and trustworthy
Focus on:
Fairness
Reliability
Transparency
Keep people at the center of design and decision-making
Microsoft and GitHub’s Six Principles of Responsible AI
Six principles guide responsible AI development and use
Principles:
Fairness
AI should treat all people equally
Avoid bias and unfair outcomes
Use balanced data and test across different groups
Reliability and Safety
AI must work consistently and as expected
Handle unexpected situations safely
Minimize physical, emotional, and financial harm
Privacy and Security
Protect user data and ensure confidentiality
Collect only necessary data with user consent
Use anonymization and encryption
Restrict and monitor access to sensitive data
Inclusiveness
AI should be accessible and usable by everyone
Work across diverse populations and regions
Support people with disabilities and different languages
Ensure no group is excluded
Transparency
AI systems should be understandable and explainable
Clearly communicate how they work and their limitations
Enable auditing, logging, and monitoring
Accountability
Humans are responsible for AI systems
Monitor performance and manage risks
Organizations must take ownership of AI outcomes
GitHub Copilot
GitHub Copilot is an AI pair programmer that helps developers write code faster and more efficiently
Developed by Microsoft and GitHub in collaboration with OpenAI
Powered by OpenAI Codex (trained on large amounts of public code)
Core Benefits:
46% of new code generated by AI
55% faster developer productivity
74% of developers feel more focused
Supported Environments:
Works with major IDEs: VS Code, Visual Studio, JetBrains, Vim/Neovim
Key Features:
Code Autocompletion
Suggests and completes code and comments in real time
Copilot Chat
Ask questions about code
Get explanations, generate tests, and documentation
Context-aware (understands your project)
Pull Request Summaries
Automatically summarizes code changes
Improves collaboration and clarity
Code Review Assistance
Suggests improvements and identifies issues
Helps speed up code reviews
Copilot CLI
Suggests commands and scripts
Explains terminal output and errors
Copilot Spaces
AI collaboration workspace
Helps with planning, design, and project understanding
Copilot Coding Agent
Executes multi-step coding tasks
Can generate files, features, and project scaffolding
Subscription Plans:
Free:
Limited usage (code completions + chat requests)
Access to advanced models
Pro:
Unlimited completions and chat
Better performance and features
Pro+:
Includes Pro features + higher limits and priority access
Business:
For organizations
Adds security, policy control, and compliance features
Enterprise:
Advanced organizational features
Personalized suggestions using internal codebase
Deep integration with GitHub
Interact with GitHub Copilot
Copilot provides multiple ways to interact and assist during development
Works in real time based on your code context
Key Interaction Methods:
Inline Suggestions
Real-time code completions as you type
Accept: Tab or →
Reject: Esc or continue typing
Useful for repetitive code and boilerplate
Command Palette
Access Copilot features quickly
Shortcut: Ctrl+Shift+P (Cmd+Shift+P on Mac)
Examples: Explain code, generate tests
Copilot Chat
Ask questions in natural language
Get code, explanations, and help
Context-aware (understands your project)
Inline Chat
Context-specific help inside code
Shortcut: Ctrl+I (Cmd+I on Mac)
Used for edits, explanations, improvements
Slash Commands (used in chat):
/explain → explains selected code
/suggest → suggests code
/tests → generates unit tests
/comment → converts comments into code
Other Features:
Comments to Code
Write comments → Copilot generates code
Multiple Suggestions
Provides alternative solutions
Cycle through options (Alt+] / Option+])
Code Explanations
Explain selected code blocks
Helps understand existing code
Automated Test Generation
Generates unit tests for functions/classes
Improves code quality and reliability
GitHub Copilot Setup, Configuration, and Troubleshooting
Sign Up for GitHub Copilot:
Go to GitHub Settings → Copilot section
Start free trial or choose a subscription
Install Copilot extension for your environment
Supported: VS Code, Visual Studio, JetBrains, Neovim, GitHub.com
Configure in VS Code:
Install Extension:
Go to Visual Studio Marketplace → GitHub Copilot
Install and open in VS Code
Sign in to GitHub when prompted
Enable/Disable Copilot:
Use status icon in bottom bar
Options:
Disable globally
Disable for specific language
Inline Suggestions Settings:
Go to File → Preferences → Settings
Extensions → GitHub Copilot
Enable/disable auto completions
Can configure per language
Troubleshooting:
Logs:
Open Command Palette (Ctrl+Shift+P)
Use:
Developer: Open Log File
Developer: Open Extensions Logs Folder
Advanced Logs:
Help → Toggle Developer Tools (Electron logs)
Diagnostics:
Command Palette → “GitHub Copilot: Collect Diagnostics”
Common Issues:
Network restrictions (firewall/proxy) may block Copilot
Use diagnostics to inspect or share with support