Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News GitHub Copilot Adopts Paid Model, Still Free for Some Open-Source Maintainers and Students

GitHub Copilot Adopts Paid Model, Still Free for Some Open-Source Maintainers and Students


After almost one year in technical preview, GitHub Copilot is now prime time-ready for students and individual developers, says GitHub, while companies and larger organizations could get access to it before the end of the year.

GitHub Copilot is an AI-powered service that aims to help developers write new code by analyzing already existing code as well as comments. This could be described as a sort of code completion "on steroids", but it can go as far as to suggest the implementation of whole functions based on their signatures.

When you type code or comments, GitHub Copilot suggests the next line of code. But it’s not only a single word or line of code. GitHub Copilot can suggest complete methods, boilerplate code, whole unit tests, and even complex algorithms.

According to GitHub, over 1.2 million developers used Copilot in the last twelve months, with a shocking 40% figure of code written by Copilot in files where it is enabled.

The transition to general availability mostly means that Copilot ceases to be available for free. Interested developers will have to pay 10 USD/month or $100 USD/year to use the service, with a 60-day free trial. Yet, GitHub will offer Copilot for free to verified students and to maintainers of popular open source projects.

Since its launch, the open source community has been extensively discussing the implications of Copilot and similar services and sharing hands-on experience with it.

While the service has raised concerns about the security of the suggestions it provides or code correctness, a number of senior developers have also highlighted its benefits when it comes to generating boilerplate code. More recently the general consensus seems to be that the tool can come up with great suggestions in a few cases, while providing acceptable ones or completely missing the mark in others. Some users suggested the possibility of enabling or disabling it depending on the kind of code that one is going to write, e.g. enabling it before starting to write unit tests. As Microsoft's Cassie Breviu stresses, it is important the programmer knows how to correctly ask things and understand what they are doing:

When you think about things like that and you think about AI as your pair programmer, you still need to understand what you're doing. You still need to understand functionally how to build your program and you still need to understand the code that's coming out. It doesn't replace us. It makes us more productive and it changes your workflow, which is really cool.

Whatever its merits may currently be on technical or usefulness ground, AI-based code generation is still in its infancy. This means it is understandable that it is not perfect yet, but it will likely improve over time and help increase developer productivity. However, as it often happens with AI-powered tech, there is also a number of concerns of legal and moral nature that have been topic of discussion since the very beginning. It is a matter of speculation whether the delay in offering Copilot to companies may be related to legal concerns, such as what would be the status of a codebase including a Copilot-suggested code snippet identical to some GPL or copyleft code. According to GitHub, while not frequent, there is definitely a possibility that Copilot outputs code snippets that match those in the training set.

GitHub Copilot is not the only AI-based developer assistant currently available. Most notably, Tabnine and Kite are two older alternatives. And it is likely that new players will join this arena, as the introduction of AWS CodeWhisperer preview seems to point out.

About the Author

Rate this Article