A new AI extension makes coding easier by automatically suggesting code as you write.

GitHub unveiled its code writing AI tool Copilot to the public this week and you can try it with a free 60-day trial.

Copilot is built on OpenAI’s code-specific language model Codex which was trained using publicly available code including GitHub repositories.

Thomas Dohmke, CEO of Microsoft-owned GitHub, said Copilot is the first time an AI system will be easily accessible for writing code.

“Just like the rise of compilers and open source, we believe AI-assisted coding will fundamentally change the nature of software development, giving developers a new tool to write code easier and faster so they can be happier in their lives,” he said.

“GitHub Copilot distills the collective knowledge of the world’s developers into an editor extension that suggests code in real time, to help you stay focused on what matters most: building great software.”

It works through an extension installed into your editor or development environment.

As you start writing, Copilot automatically fills in your code based on the context. Sometimes these suggestions will be a single line, variable, or comment, other times it will fill out whole functions for you.

Copilot has native support for Visual Studio, Visual Studio Code, Neovim, and Jetbrains IDEs.

Github is charging around US$10 per month – the website only lists prices in USD – or US$100 per year to use Copilot beyond the introductory 60-day trial.

Verified students and maintainers of popular open-source projects can use Copilot for free.

Copilot hasn’t been entirely welcomed with open arms from the community.

People like freelance developer Bert Regeer have taken exception to the way GitHub is selling Copilot right back to the people who, by contributing to GitHub repositories, helped train the AI tool.

In a Hacker News thread, Regeer suggested he should have been compensated for his contribution to certain Python libraries that were used to train Copilot.

“It feels very icky that Github now gets to sell what it learned from my code base, when it has already been shown to replicate code with a 100 per cent match,” he said.

During last year’s beta tests for Copilot, Austrian developer Armin Ronacher demonstrated how it produced, character-for-character, the implementation of a fast inverse square root from video game Quake III.

Notably, Copilot’s code also copies the code’s comments verbatim including “what the f***?” on one line. When prompted, Copilot filled out its own version of a license for the code.

GitHub reportedly filtered part of the Quake III code, along with a large list of words the AI won’t interact with, to limit any overt copyright violations.

Information Age’s tests of Copilot found it could still be induced to reproduce the code, albeit with one comment censored to just say “what the”.

Copilot now provides a link to the fast inverse square root Wikipedia page when prompted to fill out a license for the code. It has a setting that lets you decide if you want to allow or block pasting code snippets directly from public repositories.