GitHub have just launched an interesting new AI driven project, GitHub Copilot. GitHub Copilot is developed in collaboration between GitHub and OpenAI and is using artificial intelligence to support the developers. Instead of just being a better Intellisense, Copilot actually creates code dynamically using machine learning, based off comments or function names provided by the programmer. Currently it supports several languages, but Python, JavaScript, Ruby and Go are currently recommended. Copilot is implemented as a Visual Studio Code extension and there is a waiting list to gain access.
Details of GitHub Copilot from the FAQ:
What is GitHub Copilot?
GitHub Copilot is an AI pair programmer that helps you write code faster and with less work. GitHub Copilot draws context from comments and code, and suggests individual lines and whole functions instantly. GitHub Copilot is powered by OpenAI Codex, a new AI system created by OpenAI. The GitHub Copilot technical preview is available as a Visual Studio Code extension.
How does GitHub Copilot work?
OpenAI Codex was trained on publicly available source code and natural language, so it understands both programming and human languages. The GitHub Copilot editor extension sends your comments and code to the GitHub Copilot service, which then uses OpenAI Codex to synthesize and suggest individual lines and whole functions.
GitHub CoPilot opens up a world of opportunity, but also is full of potential pitfalls. One such example has already been documented, when Copilot copied GPL Quake code verbatim, an action that would cause many people to be out of compliance with that particular source license. That said, Copilot is in the early days and it will be interesting to see what this develops into. If you are interested in a AI assisted Intellisense, be sure to check out Microsoft’s other project, IntelliCode (previously discussed here). If you are interested on getting on the Copilot waiting list, you can apply here.
You can learn more about GitHub Copilot in the video below.