Collaborative coding and version control service GitHub has announced a new feature it hopes will make light-night hacking a little less lonely: Copilot, an artificially intelligent pair programmer.
"GitHub Copilot draws context from the code you’re working on, suggesting whole lines or entire functions," explains GitHub's Nat Friedman of the company's launch. "It helps you quickly discover alternative ways to solve problems, write tests, and explore new APIs without having to tediously tailor a search for answers on the internet. As you type, it adapts to the way you write code — to help you complete your work faster."
The Copilot is a partnership between GitHub and OpenAI, driven by the latter's OpenAI Codex system. Unlike general-purpose text generation models like GPT-3, OpenAI Codex is tuned specifically for generating source code — "in part," Friedman explains, "because it was trained on a data set that includes a much larger concentration of public source code."
"With GitHub Copilot, you’re always in charge," the company claims in the feature's documentation. "You can cycle through alternative suggestions, choose which to accept or reject, and manually edit suggested code. GitHub Copilot adapts to the edits you make, matching your coding style."
The technical preview is available now for those programming in Visual Studio Code, though spaces on the program are limited. It's free to use, but may not remain that way: "If the technical preview is successful," GitHub explains, "our plan is to build a commercial version of GitHub Copilot in the future. We want to use the preview to learn how people use GitHub Copilot and what it takes to operate it at scale."
Interested parties can sign up on the dedicated site.