GitHub releases AI programmer that helps writing code

GitHub has released a technical preview of Copilot, an AI-powered pair programmer that helps developers write code. Copilot suggests contextual code and functions, and helps solve problems by learning from the code someone writes.

Github Copilot was developed together with OpenAI and is based on the OpenAI Codex. This AI application has knowledge of how humans program and code is according to Github better than say GPT-3 at code generation because it is specifically trained on a large dataset of public source code.

As code is written, the developer gets suggestions for whole lines of code or even complete functions that Copilot suggests based on the context as you type, that is, based on the code a developer is working on. Copilot can automatically turn comments into code, complete repetitive code, and test a function as you write. Users can scroll through the suggestions Copilot suggests and choose to accept or ignore them. Suggested code can also be edited manually.

Github emphasizes that Copilot is more than an autocomplete as it understands and learns much more from context. It can tell the difference between, say, a docstring, a comment, a function name, or code itself. In addition, Copilot gets smarter as it is used more because it learns how a developer codes.

Github says Copilot works with dozens of frameworks and programming languages, but this technical preview has focused on Python, JavaScript, TypeScript, Ruby, and Go. For now, Copilot is an extension for Visual Studio Code and users must register for a waiting list to try out Copilot.