Github Copilot is an AI programming tool jointly launched by Microsoft and OpenAI, which automatically generates code through users’ comments, function names, etc. Never felt gimmicks outweighed reality. I tried it today and the results were amazing. In particular, some of the paragraphs in this article were also supplemented by Github Copilot.
How does Github Copilot work
Github Copilot generates code based on natural language prompts by using public source code on Github and natural language as training data. To generate code, GitHub Copilot transfers some of the files you are editing to the server as a context. The service then uses OpenAI Codex (based on GPT-3) to compose and suggest individual lines and entire functions.
Install making Copilot
GitHub Copilot is available as an extension to Neovim, JetBrains, and Visual Studio Code. For details, see github.com/github/copi…
Use making Copilot
Once the plugin is installed, you just need to write the code as usual. Github Copilot will complete the code at the right time and location. Press TAB to confirm.
The highlighted part of the image above is a reminder that the code is Github Copilot. In particular, the following sections of test code do exactly what I want.
Some questions about Github Copilot
After Github Copilot was launched, there were many questions about it. For example:
- Security of private project code
- Whether Github Copilot introduces malicious code
- Whether Copilot’s training on the common repository infringes copyright