GitHub introduces an AI pal to help you code better

GitHub is already one of the most popular platforms to build and host coding projects. Last year, it launched a built-in Integrated Development Environment (IDE) to let you code directly into the browser. Now, it’s introducing an AI-powered automcompletion tool called GitHub Copilot.

The aim of this feature is to learn how you write code, and to suggest lines and functions automatically, learning how to. What’s more, the Copilot tool helps you find alternative ways to solve a problem and use, new APIs easily.

GitHub has developed this feature in partnership with OpenAI, but it’s not based on GPT-3. Here’s what the company said about its technical capability on the blog:

OpenAI Codex has broad knowledge of how people use code and is significantly more capable than GPT-3 in code generation, in part, because it was trained on a data set that includes a much larger concentration of public source code.

The feature is also capable of autofilling code that repeats itself, turning comments or descriptions into functioning code, and helping you build tests for your project., and help you build tests for your project.

Currently, this function is available only as a technical preview, so you’ll have to apply to try it via this page. The technical preview supports many languages but the firm said it “works especially well with” Python, JavaScript, TypeScript, Ruby, and Go.

AI-powered autocompleters have been around for a while. Last year, when we talked to Codata, its founders said that the tool helped you write 25% less code — and more than a million developers were already using it. At that time, another such tool called Kite — focused mostly on Python-based programming — had more than 250,000 users.

GitHub entering this market will give developers more options to use AI for completing their projects faster.