GitHub's Copilot X, designed to assist developers by providing code suggestions, has now expanded its capabilities with the introduction of Copilot Chat.

Thomas Dohmke, CEO of GitHub, announced the launch of GitHub Copilot X, an AI-powered developer experience, in March 2023. The tool was touted as a revolutionary step towards integrating AI into the coding process, promising to make coding more efficient and accessible. Earlier this month, Mario Rodriguez, VP of Product at GitHub, announced the public beta availability of GitHub Copilot Chat for every organization. The chat feature was designed to provide real-time assistance to developers, answering their queries and helping troubleshoot issues. GitHub Copilot is available starting at $10 per month after a 30-day trial.

GitHub's official documentation provides a comprehensive guide on using Copilot Chat, from setting it up to troubleshooting authentication issues. The tool is designed to be a comprehensive assistant for developers, providing real-time code suggestions and helping to troubleshoot issues.

One of the key features of Copilot Chat is its ability to provide context-aware code completions. This means that the tool can understand the context of the code that the developer is working on and provide relevant suggestions. For example, if a developer is working on a function to calculate the factorial of a number, Copilot Chat can provide a suggestion for the implementation of the function based on the context.

Another feature is the ability to answer queries about code. Developers can ask Copilot Chat questions about their code, and the tool will provide answers based on its understanding of the code. This can be particularly useful for junior developers who are still learning and need guidance. The tool also helps with troubleshooting issues. For example, if a developer is facing authentication issues with their code, they can ask Copilot Chat for help, and the tool will provide suggestions on how to resolve the issue.

However, the real-world application of the tool has been met with mixed reviews. Some users have found the tool to be extremely helpful, particularly for generating unit tests. Others have found the tool to be lacking, particularly when it comes to identifying and correcting simple errors.

For example, one user noted that the tool failed to identify a spelling mistake in a JavaScript function. This suggests that while the tool can provide complex code suggestions, it may still struggle with simpler tasks. "The tool has limitations including difficulty with complex structures or rare languages, potential biases from training data, possible security risks requiring code review, occasional reproduction of public code, generation of inaccurate code, and poor responses to non-coding questions," says Aruna Pattam, Head - Generative AI Practice at Capgemini in Sydney. On the other hand, a user recommended signing up for the waitlist for Copilot Chat, indicating a positive experience with the tool. "As an experienced developer, the addition of GitHub Copilot Chat has been the most impressive technological advance I have encountered, so far," says Sayvai Tosu, a Front-end Engineer with SigTech in London. This suggests that the tool can be extremely useful for some developers, depending on their specific needs and use cases.

The mixed reviews highlight a key issue with AI tools in development - while they hold great potential, their effectiveness is often dependent on the specific use case and the user's expectations.

For junior developers or those working on simple tasks, Copilot Chat could prove to be a valuable tool. However, for more complex tasks or bespoke code, the tool's current capabilities may fall short. Moreover, the tool's effectiveness as a learning resource for junior developers is debatable. While it could provide valuable insights and suggestions, there's also the risk of the tool providing incorrect or suboptimal solutions, leading to potential misunderstandings or misconceptions.

GitHub's Copilot Chat represents an exciting step forward in the integration of AI in the coding process. However, it's clear that the tool still has room for improvement. For now, GitHub Copilot Chat is a coder's companion tool, not a replacement. As Pattam says, "for safety, ensure to review and test all generated code". As AI continues to evolve and improve, we can expect tools like Copilot Chat to become more refined and effective, eventually becoming an indispensable part of the developer's toolkit.

Have questions or comments about this article? Reach out to us here.

Banner Image Credits: Attendees at Great International Developer Summit